[WIP] Simplifying refactor by hunhoffe · Pull Request #74 · amd/IRON

hunhoffe · 2026-02-05T22:55:46Z

The draftiest of draft PRs.

Added

Changed

Removed

PR Merge Checklist

The PR is rebased on the latest devel commit and pointing to devel.
Your PR has been reviewed and approved.
All checks are passing.

…utput quality

… score calculation in llama

… now)

… block fused -- 2.5 TPS

… -- 4.4 TPS

…straction with new iron/ structure

github-actions · 2026-02-06T13:53:53Z

📊 Test Results for Small Benchmark/Test Suite

0685306 (2026_02_06_13_53_22)

IRONCLAD

Tested on 2026_02_06_13_53_22 at commit 0685306.

Test	Checks	Latency (mean)	Bandwidth (mean)	GFLOP/s (mean)
axpy_1_cols_2_channels_2048_tile_2048_3.0_0	✅ 5/5	147.50	0.08	n/a
axpy_2_cols_2_channels_2048_tile_1024_3.0_0	✅ 5/5	175.52	0.07	n/a
axpy_4_cols_2_channels_2048_tile_512_3.0_0	✅ 5/5	190.22	0.07	n/a
axpy_8_cols_2_channels_2048_tile_256_3.0_0	✅ 5/5	222.50	0.06	n/a
dequant_1_cols_1_channels_2048_tile_2048_0	✅ 5/5	175.22	0.03	n/a
dequant_1_cols_2_channels_2048_tile_1024_0	✅ 5/5	157.66	0.03	n/a
dequant_2_cols_1_channels_2048_tile_1024_0	✅ 5/5	172.32	0.03	n/a
dequant_2_cols_2_channels_2048_tile_512_0	✅ 5/5	149.76	0.04	n/a
dequant_4_cols_1_channels_2048_tile_512_0	✅ 5/5	157.38	0.03	n/a
dequant_4_cols_2_channels_2048_tile_256_0	✅ 5/5	176.18	0.03	n/a
dequant_8_cols_1_channels_2048_tile_256_0	✅ 5/5	182.34	0.03	n/a
dequant_8_cols_2_channels_2048_tile_128_0	✅ 5/5	215.82	0.02	n/a
eltwise_add_1_cols_2_channels_2048_tile_2048	✅ 5/5	176.44	0.08	n/a
eltwise_add_2_cols_2_channels_2048_tile_1024	✅ 5/5	152.98	0.08	n/a
eltwise_add_4_cols_2_channels_2048_tile_512	✅ 5/5	193.54	0.07	n/a
eltwise_add_8_cols_2_channels_2048_tile_256	✅ 5/5	212.26	0.06	n/a
eltwise_mul_1_cols_2_channels_2048_tile_2048	✅ 5/5	118.54	0.10	n/a
eltwise_mul_2_cols_2_channels_2048_tile_1024	✅ 5/5	147.76	0.09	n/a
eltwise_mul_4_cols_2_channels_2048_tile_512	✅ 5/5	136.58	0.09	n/a
eltwise_mul_8_cols_2_channels_2048_tile_256	✅ 5/5	173.24	0.07	n/a
gelu_1_cols_1_channels_2048_tile_2048	✅ 5/5	172.32	0.05	n/a
gelu_1_cols_2_channels_2048_tile_1024	✅ 5/5	157.00	0.05	n/a
gelu_2_cols_1_channels_2048_tile_1024	✅ 5/5	202.28	0.04	n/a
gelu_2_cols_2_channels_2048_tile_512	✅ 5/5	168.64	0.05	n/a
gelu_4_cols_1_channels_2048_tile_512	✅ 5/5	151.74	0.06	n/a
gelu_4_cols_2_channels_2048_tile_256	✅ 5/5	173.08	0.05	n/a
gelu_8_cols_1_channels_2048_tile_256	✅ 5/5	214.44	0.04	n/a
gelu_8_cols_2_channels_2048_tile_128	✅ 5/5	265.54	0.03	n/a
gemm_1792x896x1152_64x32x48_8cols_ccolmaj	❌ 0/5	2068.96	4.57	n/a
gemm_192x384x64_48x96x16_4cols	❌ 0/5	320.66	0.74	n/a
gemm_192x384x64_48x96x16_4cols_bcolmaj_ccolmaj	❌ 0/5	214.48	1.08	n/a
gemm_2048x2048x2048_64x64x64_1cols	❌ 0/5	46284.84	0.54	n/a
gemm_2048x2048x2048_64x64x64_2cols_bcolmaj	❌ 0/5	26656.06	0.94	n/a
gemm_2048x2048x2048_64x64x64_8cols_bcolmaj_ccolmaj	❌ 0/5	7492.40	3.36	n/a
gemm_384x1536x1792_32x48x64_4cols_bcolmaj	❌ 0/5	1967.92	4.11	n/a
gemm_896x1792x640_32x64x80_8cols_ccolmaj	❌ 0/5	1389.78	4.90	n/a
layer_norm_1_cols_1_channels_2048_tile_2048	✅ 5/5	147.40	0.06	n/a
layer_norm_1_cols_2_channels_2048_tile_1024	✅ 5/5	170.02	0.05	n/a
layer_norm_2_cols_1_channels_2048_tile_1024	✅ 5/5	153.40	0.06	n/a
layer_norm_2_cols_2_channels_2048_tile_512	✅ 5/5	190.64	0.04	n/a
layer_norm_4_cols_1_channels_2048_tile_512	✅ 5/5	174.84	0.05	n/a
layer_norm_4_cols_2_channels_2048_tile_256	✅ 5/5	206.72	0.04	n/a
layer_norm_8_cols_1_channels_2048_tile_256	✅ 5/5	176.96	0.05	n/a
layer_norm_8_cols_2_channels_2048_tile_128	✅ 5/5	248.28	0.04	n/a
matrix_vector_mul_128x128_32tsi_128tso_1col0	✅ 5/5	n/a	0.24	n/a
matrix_vector_mul_2048x8192_1tsi_1024tso_2col0	✅ 5/5	n/a	24.23	n/a
matrix_vector_mul_2048x8192_1tsi_2048tso_1col0	✅ 5/5	n/a	12.83	n/a
matrix_vector_mul_2048x8192_1tsi_256tso_8col0	✅ 5/5	n/a	41.99	n/a
matrix_vector_mul_2048x8192_1tsi_512tso_4col0	✅ 5/5	n/a	38.47	n/a
matrix_vector_mul_8192x2048_4tsi_1024tso_1col0	✅ 5/5	n/a	13.11	n/a
matrix_vector_mul_8192x2048_4tsi_1024tso_2col0	✅ 5/5	n/a	23.79	n/a
matrix_vector_mul_8192x2048_4tsi_1024tso_4col0	✅ 5/5	n/a	39.70	n/a
matrix_vector_mul_8192x2048_4tsi_1024tso_8col0	✅ 5/5	n/a	40.93	n/a
mem_copy_16_cores_2_chans_2048_tile_128_False0	✅ 5/5	176.58	0.05	n/a
mem_copy_1_cores_1_chans_2048_tile_2048_False0	✅ 5/5	191.50	0.04	n/a
mem_copy_2_cores_1_chans_2048_tile_1024_False0	✅ 5/5	129.94	0.06	n/a
mem_copy_2_cores_2_chans_2048_tile_1024_False0	✅ 5/5	178.38	0.05	n/a
mem_copy_4_cores_1_chans_2048_tile_512_False0	✅ 5/5	166.84	0.05	n/a
mem_copy_4_cores_2_chans_2048_tile_512_False0	✅ 5/5	180.82	0.05	n/a
mem_copy_8_cores_1_chans_2048_tile_256_False0	✅ 5/5	154.50	0.06	n/a
mem_copy_8_cores_2_chans_2048_tile_256_False0	✅ 5/5	183.82	0.05	n/a
mha0	✅ 5/5	47565.12	0.18	n/a
relu_1_cols_1_channels_2048_tile_2048	✅ 5/5	174.54	0.05	n/a
relu_2_cols_1_channels_2048_tile_1024	✅ 5/5	161.22	0.05	n/a
relu_4_cols_1_channels_2048_tile_512	✅ 5/5	163.32	0.05	n/a
relu_8_cols_1_channels_2048_tile_256	✅ 5/5	164.26	0.05	n/a
rms_norm_1_cols_1_channels_2048_tile_2048	✅ 5/5	195.98	0.05	n/a
rms_norm_1_cols_2_channels_2048_tile_1024	✅ 5/5	154.02	0.05	n/a
rms_norm_2_cols_1_channels_2048_tile_1024	✅ 5/5	143.18	0.06	n/a
rms_norm_2_cols_2_channels_2048_tile_512	✅ 5/5	147.98	0.06	n/a
rms_norm_4_cols_1_channels_2048_tile_512	✅ 5/5	161.72	0.05	n/a
rms_norm_4_cols_2_channels_2048_tile_256	✅ 5/5	153.88	0.05	n/a
rms_norm_8_cols_1_channels_2048_tile_256	✅ 5/5	156.02	0.05	n/a
rms_norm_8_cols_2_channels_2048_tile_128	✅ 5/5	181.08	0.05	n/a
rope_1c_32rows_512cols_32arows_0m	✅ 5/5	190.64	0.55	n/a
rope_1c_32rows_512cols_8arows_0m	✅ 5/5	160.80	0.50	n/a
rope_2c_32rows_512cols_32arows_0m	✅ 5/5	136.04	0.75	n/a
rope_2c_32rows_512cols_8arows_0m	✅ 5/5	146.78	0.52	n/a
rope_8c_32rows_512cols_32arows_0m	✅ 5/5	163.12	0.63	n/a
rope_8c_32rows_512cols_8arows_0m	✅ 5/5	173.20	0.44	n/a
sigmoid_1_cols_1_channels_2048_tile_2048	✅ 5/5	138.88	0.06	n/a
sigmoid_2_cols_1_channels_2048_tile_1024	✅ 5/5	141.92	0.06	n/a
sigmoid_4_cols_1_channels_2048_tile_512	✅ 5/5	142.10	0.06	n/a
sigmoid_8_cols_1_channels_2048_tile_256	✅ 5/5	164.20	0.05	n/a
silu_1_cols_1_channels_2048_tile_2048	✅ 5/5	178.58	0.05	n/a
silu_2_cols_1_channels_2048_tile_1024	✅ 5/5	193.08	0.04	n/a
silu_4_cols_1_channels_2048_tile_512	✅ 5/5	188.62	0.04	n/a
silu_8_cols_1_channels_2048_tile_256	✅ 5/5	204.18	0.04	n/a
softmax_2_cols_2_channels_32768_tile_1024	✅ 5/5	172.88	0.76	n/a
softmax_2_cols_2_channels_32768_tile_512	✅ 5/5	188.58	0.72	n/a
softmax_4_cols_4_channels_32768_tile_2048	✅ 5/5	216.70	0.62	n/a
swiglu_decode_1x2048x2048_0	❌ 0/5	n/a	n/a	n/a
swiglu_prefill_256x2048x2048_0	❌ 0/5	n/a	n/a	n/a
tanh_1_cols_1_channels_2048_tile_2048	✅ 5/5	138.52	0.06	n/a
tanh_2_cols_1_channels_2048_tile_1024	✅ 5/5	165.56	0.05	n/a
tanh_4_cols_1_channels_2048_tile_512	✅ 5/5	164.62	0.06	n/a
tanh_8_cols_1_channels_2048_tile_256	✅ 5/5	192.60	0.04	n/a
transpose_2048_M_64_N_1_cols_1_channels_64_m_64_n_8_s0	✅ 5/5	183.46	2.97	n/a
transpose_2048_M_64_N_1_cols_2_channels_64_m_64_n_8_s0	✅ 5/5	158.74	3.35	n/a
weighted_rms_norm_1_cols_2_channels_2048_weights_2048	❌ 0/5	n/a	n/a	n/a
weighted_rms_norm_2_cols_2_channels_2048_weights_1024	❌ 0/5	n/a	n/a	n/a
weighted_rms_norm_4_cols_2_channels_2048_weights_512	❌ 0/5	n/a	n/a	n/a
weighted_rms_norm_8_cols_2_channels_2048_weights_256	❌ 0/5	n/a	n/a	n/a

📈 Trends (vs main branch) for Small Benchmark/Test Suite

0685306 (2026_02_06_13_53_22)

IRONCLAD Trends

axpy_1_cols_2_channels_2048_tile_2048_3.0

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`130b6ea` — 2025-12-05 21:15:49	0.13 (-1.96%)	0.11 (+2.52%)	0.10 (+1.86%)	0.08 (-1.91%)	0.02 (-1.72%)	162.40 (+1.95%)	120.96 (-2.36%)	117.00 (-1.85%)	94.40 (+2.05%)	23.11 (+4.09%)
`0a6c11c` — 2025-12-04 00:39:10	0.13 (n/a)	0.10 (n/a)	0.10 (n/a)	0.08 (n/a)	0.02 (n/a)	159.30 (n/a)	123.88 (n/a)	119.20 (n/a)	92.50 (n/a)	22.21 (n/a)

axpy_1_cols_2_channels_2048_tile_2048_3.0_0

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`0685306` — 2026-02-06 13:45:43	0.09 (-40.41%)	0.08 (-33.85%)	0.08 (-24.80%)	0.07 (-30.71%)	0.01 (-69.42%)	171.40 (+44.28%)	147.50 (+47.06%)	146.90 (+33.06%)	130.90 (+67.82%)	15.56 (-24.05%)
`1531781` — 2026-01-23 18:49:24	0.16 (n/a)	0.13 (n/a)	0.11 (n/a)	0.10 (n/a)	0.03 (n/a)	118.80 (n/a)	100.30 (n/a)	110.40 (n/a)	78.00 (n/a)	20.49 (n/a)

axpy_2_cols_2_channels_2048_tile_1024_3.0

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`130b6ea` — 2025-12-05 21:15:49	0.11 (-26.77%)	0.09 (-9.09%)	0.09 (-10.81%)	0.08 (+8.25%)	0.01 (-55.42%)	152.00 (-7.60%)	135.46 (+6.28%)	142.80 (+12.18%)	115.00 (+36.58%)	15.03 (-41.36%)
`0a6c11c` — 2025-12-04 00:39:10	0.15 (n/a)	0.10 (n/a)	0.10 (n/a)	0.07 (n/a)	0.02 (n/a)	164.50 (n/a)	127.46 (n/a)	127.30 (n/a)	84.20 (n/a)	25.64 (n/a)

axpy_2_cols_2_channels_2048_tile_1024_3.0_0

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`0685306` — 2026-02-06 13:45:43	0.09 (-33.78%)	0.07 (-41.71%)	0.07 (-48.28%)	0.06 (-28.55%)	0.01 (-47.28%)	219.60 (+39.96%)	175.52 (+68.03%)	174.50 (+93.24%)	132.80 (+51.08%)	31.86 (+7.93%)
`1531781` — 2026-01-23 18:49:24	0.14 (n/a)	0.12 (n/a)	0.14 (n/a)	0.08 (n/a)	0.03 (n/a)	156.90 (n/a)	104.46 (n/a)	90.30 (n/a)	87.90 (n/a)	29.52 (n/a)

axpy_4_cols_2_channels_2048_tile_512_3.0

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`130b6ea` — 2025-12-05 21:15:49	0.12 (-10.21%)	0.10 (-8.41%)	0.10 (-6.20%)	0.07 (+3.52%)	0.02 (-23.48%)	166.70 (-3.36%)	128.78 (+7.33%)	121.80 (+6.65%)	103.90 (+11.36%)	22.95 (-18.82%)
`0a6c11c` — 2025-12-04 00:39:10	0.13 (n/a)	0.11 (n/a)	0.11 (n/a)	0.07 (n/a)	0.02 (n/a)	172.50 (n/a)	119.98 (n/a)	114.20 (n/a)	93.30 (n/a)	28.27 (n/a)

axpy_4_cols_2_channels_2048_tile_512_3.0_0

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`0685306` — 2026-02-06 13:45:43	0.09 (-45.04%)	0.07 (-52.19%)	0.06 (-57.91%)	0.06 (-39.27%)	0.01 (-54.69%)	220.80 (+64.65%)	190.22 (+105.78%)	198.10 (+137.53%)	143.10 (+82.06%)	30.05 (+28.51%)
`1531781` — 2026-01-23 18:49:24	0.16 (n/a)	0.14 (n/a)	0.15 (n/a)	0.09 (n/a)	0.03 (n/a)	134.10 (n/a)	92.44 (n/a)	83.40 (n/a)	78.60 (n/a)	23.38 (n/a)

axpy_8_cols_2_channels_2048_tile_256_3.0

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`130b6ea` — 2025-12-05 21:15:49	0.10 (-16.19%)	0.08 (-10.40%)	0.09 (-2.09%)	0.07 (-17.38%)	0.01 (-10.61%)	184.40 (+21.00%)	153.18 (+11.99%)	144.10 (+2.20%)	127.10 (+19.34%)	21.26 (+34.76%)
`0a6c11c` — 2025-12-04 00:39:10	0.12 (n/a)	0.09 (n/a)	0.09 (n/a)	0.08 (n/a)	0.01 (n/a)	152.40 (n/a)	136.78 (n/a)	141.00 (n/a)	106.50 (n/a)	15.78 (n/a)

axpy_8_cols_2_channels_2048_tile_256_3.0_0

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`0685306` — 2026-02-06 13:45:43	0.08 (-51.21%)	0.06 (-58.83%)	0.06 (-61.58%)	0.04 (-53.24%)	0.02 (-53.67%)	314.00 (+113.90%)	222.50 (+141.22%)	208.10 (+160.12%)	148.30 (+105.12%)	61.26 (+97.18%)
`1531781` — 2026-01-23 18:49:24	0.17 (n/a)	0.14 (n/a)	0.15 (n/a)	0.08 (n/a)	0.03 (n/a)	146.80 (n/a)	92.24 (n/a)	80.00 (n/a)	72.30 (n/a)	31.07 (n/a)

dequant_1_cols_1_channels_2048_tile_2048

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`130b6ea` — 2025-12-05 21:15:49	0.05 (-18.83%)	0.04 (-5.21%)	0.04 (-2.69%)	0.03 (+37.10%)	0.00 (-53.98%)	160.40 (-27.06%)	141.12 (-0.91%)	136.40 (+2.79%)	114.30 (+23.17%)	17.37 (-59.06%)
`0a6c11c` — 2025-12-04 00:39:10	0.06 (n/a)	0.04 (n/a)	0.04 (n/a)	0.02 (n/a)	0.01 (n/a)	219.90 (n/a)	142.42 (n/a)	132.70 (n/a)	92.80 (n/a)	42.43 (n/a)

dequant_1_cols_1_channels_2048_tile_2048_0

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`0685306` — 2026-02-06 13:45:43	0.04 (-24.56%)	0.03 (-29.82%)	0.03 (-28.87%)	0.02 (-46.68%)	0.01 (+40.79%)	235.30 (+87.49%)	175.22 (+52.26%)	170.90 (+40.54%)	116.80 (+32.58%)	55.40 (+259.75%)
`1531781` — 2026-01-23 18:49:24	0.06 (n/a)	0.05 (n/a)	0.04 (n/a)	0.04 (n/a)	0.01 (n/a)	125.50 (n/a)	115.08 (n/a)	121.60 (n/a)	88.10 (n/a)	15.40 (n/a)

dequant_1_cols_2_channels_2048_tile_1024

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`130b6ea` — 2025-12-05 21:15:49	0.06 (-13.28%)	0.05 (-8.93%)	0.06 (-13.59%)	0.04 (+4.63%)	0.01 (-39.10%)	133.60 (-4.43%)	106.08 (+5.85%)	95.10 (+15.69%)	87.90 (+15.35%)	17.90 (-31.97%)
`0a6c11c` — 2025-12-04 00:39:10	0.07 (n/a)	0.06 (n/a)	0.06 (n/a)	0.04 (n/a)	0.01 (n/a)	139.80 (n/a)	100.22 (n/a)	82.20 (n/a)	76.20 (n/a)	26.31 (n/a)

dequant_1_cols_2_channels_2048_tile_1024_0

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`0685306` — 2026-02-06 13:45:43	0.04 (-40.13%)	0.03 (-28.56%)	0.04 (-24.57%)	0.02 (-26.86%)	0.01 (-45.86%)	223.90 (+36.77%)	157.66 (+37.50%)	147.60 (+32.61%)	127.80 (+67.06%)	39.26 (+24.12%)
`1531781` — 2026-01-23 18:49:24	0.07 (n/a)	0.05 (n/a)	0.05 (n/a)	0.03 (n/a)	0.01 (n/a)	163.70 (n/a)	114.66 (n/a)	111.30 (n/a)	76.50 (n/a)	31.63 (n/a)

dequant_2_cols_1_channels_2048_tile_1024

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`130b6ea` — 2025-12-05 21:15:49	0.07 (+5.66%)	0.05 (+1.21%)	0.05 (+0.61%)	0.04 (-15.02%)	0.01 (+58.42%)	143.40 (+17.64%)	101.02 (+1.49%)	97.60 (-0.61%)	78.90 (-5.40%)	22.56 (+78.52%)
`0a6c11c` — 2025-12-04 00:39:10	0.06 (n/a)	0.05 (n/a)	0.05 (n/a)	0.04 (n/a)	0.01 (n/a)	121.90 (n/a)	99.54 (n/a)	98.20 (n/a)	83.40 (n/a)	12.64 (n/a)

dequant_2_cols_1_channels_2048_tile_1024_0

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`0685306` — 2026-02-06 13:45:43	0.04 (-25.87%)	0.03 (-34.12%)	0.03 (-37.89%)	0.03 (-27.68%)	0.01 (-20.48%)	204.30 (+38.23%)	172.32 (+52.31%)	171.40 (+60.94%)	136.90 (+35.01%)	28.96 (+47.65%)
`1531781` — 2026-01-23 18:49:24	0.05 (n/a)	0.05 (n/a)	0.05 (n/a)	0.04 (n/a)	0.01 (n/a)	147.80 (n/a)	113.14 (n/a)	106.50 (n/a)	101.40 (n/a)	19.61 (n/a)

dequant_2_cols_2_channels_2048_tile_512

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`130b6ea` — 2025-12-05 21:15:49	0.05 (+17.06%)	0.04 (+8.98%)	0.05 (+21.37%)	0.03 (-11.65%)	0.01 (+301.89%)	159.70 (+13.18%)	126.32 (-5.68%)	109.00 (-17.67%)	106.80 (-14.56%)	22.48 (+280.15%)
`0a6c11c` — 2025-12-04 00:39:10	0.04 (n/a)	0.04 (n/a)	0.04 (n/a)	0.04 (n/a)	0.00 (n/a)	141.10 (n/a)	133.92 (n/a)	132.40 (n/a)	125.00 (n/a)	5.91 (n/a)

dequant_2_cols_2_channels_2048_tile_512_0

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`0685306` — 2026-02-06 13:45:43	0.04 (-22.28%)	0.04 (-20.15%)	0.04 (-24.38%)	0.03 (-11.44%)	0.01 (-37.91%)	178.80 (+12.95%)	149.76 (+23.38%)	140.40 (+32.20%)	126.40 (+28.72%)	23.81 (-8.87%)
`1531781` — 2026-01-23 18:49:24	0.05 (n/a)	0.04 (n/a)	0.05 (n/a)	0.03 (n/a)	0.01 (n/a)	158.30 (n/a)	121.38 (n/a)	106.20 (n/a)	98.20 (n/a)	26.13 (n/a)

dequant_4_cols_1_channels_2048_tile_512

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`130b6ea` — 2025-12-05 21:15:49	0.06 (+20.24%)	0.05 (-3.19%)	0.05 (-6.73%)	0.03 (-20.69%)	0.01 (+194.26%)	156.10 (+26.09%)	120.86 (+7.15%)	114.90 (+7.18%)	88.30 (-16.85%)	24.40 (+214.82%)
`0a6c11c` — 2025-12-04 00:39:10	0.05 (n/a)	0.05 (n/a)	0.05 (n/a)	0.04 (n/a)	0.00 (n/a)	123.80 (n/a)	112.80 (n/a)	107.20 (n/a)	106.20 (n/a)	7.75 (n/a)

dequant_4_cols_1_channels_2048_tile_512_0

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`0685306` — 2026-02-06 13:45:43	0.04 (-28.56%)	0.03 (-30.32%)	0.03 (-34.79%)	0.03 (-8.83%)	0.01 (-50.27%)	180.70 (+9.65%)	157.38 (+38.51%)	158.30 (+53.39%)	118.70 (+39.98%)	24.61 (-24.39%)
`1531781` — 2026-01-23 18:49:24	0.06 (n/a)	0.05 (n/a)	0.05 (n/a)	0.03 (n/a)	0.01 (n/a)	164.80 (n/a)	113.62 (n/a)	103.20 (n/a)	84.80 (n/a)	32.55 (n/a)

dequant_4_cols_2_channels_2048_tile_256

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`130b6ea` — 2025-12-05 21:15:49	0.05 (-6.09%)	0.04 (-8.04%)	0.04 (-12.46%)	0.03 (-0.41%)	0.01 (-20.19%)	168.10 (+0.42%)	135.74 (+7.44%)	132.20 (+14.26%)	107.40 (+6.44%)	22.91 (-12.67%)
`0a6c11c` — 2025-12-04 00:39:10	0.05 (n/a)	0.04 (n/a)	0.05 (n/a)	0.03 (n/a)	0.01 (n/a)	167.40 (n/a)	126.34 (n/a)	115.70 (n/a)	100.90 (n/a)	26.23 (n/a)

dequant_4_cols_2_channels_2048_tile_256_0

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`0685306` — 2026-02-06 13:45:43	0.03 (-52.54%)	0.03 (-41.28%)	0.03 (-45.93%)	0.03 (-3.64%)	0.00 (-87.45%)	195.50 (+3.77%)	176.18 (+56.30%)	173.60 (+85.07%)	167.80 (+110.54%)	11.32 (-74.03%)
`1531781` — 2026-01-23 18:49:24	0.07 (n/a)	0.05 (n/a)	0.06 (n/a)	0.03 (n/a)	0.01 (n/a)	188.40 (n/a)	112.72 (n/a)	93.80 (n/a)	79.70 (n/a)	43.58 (n/a)

dequant_8_cols_1_channels_2048_tile_256

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`130b6ea` — 2025-12-05 21:15:49	0.04 (-13.65%)	0.04 (-13.67%)	0.04 (-18.56%)	0.03 (-13.52%)	0.00 (-23.88%)	170.40 (+15.68%)	140.12 (+15.33%)	135.80 (+22.78%)	119.60 (+15.78%)	17.94 (+1.87%)
`0a6c11c` — 2025-12-04 00:39:10	0.05 (n/a)	0.04 (n/a)	0.05 (n/a)	0.04 (n/a)	0.01 (n/a)	147.30 (n/a)	121.50 (n/a)	110.60 (n/a)	103.30 (n/a)	17.61 (n/a)

dequant_8_cols_1_channels_2048_tile_256_0

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`0685306` — 2026-02-06 13:45:43	0.04 (-16.97%)	0.03 (-23.19%)	0.03 (-23.39%)	0.02 (-25.82%)	0.01 (+6.47%)	225.10 (+34.87%)	182.34 (+31.50%)	177.40 (+30.54%)	141.60 (+20.41%)	30.48 (+69.87%)
`1531781` — 2026-01-23 18:49:24	0.04 (n/a)	0.04 (n/a)	0.04 (n/a)	0.03 (n/a)	0.00 (n/a)	166.90 (n/a)	138.66 (n/a)	135.90 (n/a)	117.60 (n/a)	17.94 (n/a)

dequant_8_cols_2_channels_2048_tile_128

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`130b6ea` — 2025-12-05 21:15:49	0.05 (+17.21%)	0.03 (-1.70%)	0.03 (-3.04%)	0.03 (-16.20%)	0.01 (+95.59%)	198.20 (+19.33%)	160.00 (+4.92%)	166.20 (+3.17%)	106.50 (-14.66%)	29.86 (+87.19%)
`0a6c11c` — 2025-12-04 00:39:10	0.04 (n/a)	0.03 (n/a)	0.03 (n/a)	0.03 (n/a)	0.00 (n/a)	166.10 (n/a)	152.50 (n/a)	161.10 (n/a)	124.80 (n/a)	15.95 (n/a)

dequant_8_cols_2_channels_2048_tile_128_0

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`0685306` — 2026-02-06 13:45:43	0.03 (-44.41%)	0.02 (-35.83%)	0.03 (-33.89%)	0.02 (-18.01%)	0.00 (-63.49%)	280.40 (+21.97%)	215.82 (+48.23%)	198.10 (+51.22%)	191.00 (+79.85%)	37.62 (-23.28%)
`1531781` — 2026-01-23 18:49:24	0.05 (n/a)	0.04 (n/a)	0.04 (n/a)	0.02 (n/a)	0.01 (n/a)	229.90 (n/a)	145.60 (n/a)	131.00 (n/a)	106.20 (n/a)	49.03 (n/a)

eltwise_add_1_cols_2_channels_2048_tile_2048

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`0685306` — 2026-02-06 13:45:43	0.10 (-39.19%)	0.08 (-37.78%)	0.07 (-27.29%)	0.04 (-45.26%)	0.02 (-44.97%)	284.70 (+82.73%)	176.44 (+60.72%)	165.70 (+37.51%)	121.40 (+64.50%)	63.46 (+84.58%)
`1531781` — 2026-01-23 18:49:24	0.17 (n/a)	0.12 (n/a)	0.10 (n/a)	0.08 (n/a)	0.04 (n/a)	155.80 (n/a)	109.78 (n/a)	120.50 (n/a)	73.80 (n/a)	34.38 (n/a)

eltwise_add_2_cols_2_channels_2048_tile_1024

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`0685306` — 2026-02-06 13:45:43	0.10 (-36.45%)	0.08 (-35.37%)	0.09 (-35.38%)	0.07 (-18.07%)	0.01 (-51.93%)	179.10 (+22.09%)	152.98 (+51.17%)	141.90 (+54.74%)	128.00 (+57.25%)	23.52 (-9.54%)
`1531781` — 2026-01-23 18:49:24	0.15 (n/a)	0.13 (n/a)	0.13 (n/a)	0.08 (n/a)	0.03 (n/a)	146.70 (n/a)	101.20 (n/a)	91.70 (n/a)	81.40 (n/a)	26.00 (n/a)

eltwise_add_4_cols_2_channels_2048_tile_512

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`0685306` — 2026-02-06 13:45:43	0.08 (-50.65%)	0.07 (-54.95%)	0.06 (-60.13%)	0.05 (-52.47%)	0.01 (-43.00%)	244.10 (+110.43%)	193.54 (+123.85%)	202.90 (+150.80%)	147.30 (+102.61%)	39.86 (+131.44%)
`1531781` — 2026-01-23 18:49:24	0.17 (n/a)	0.15 (n/a)	0.15 (n/a)	0.11 (n/a)	0.02 (n/a)	116.00 (n/a)	86.46 (n/a)	80.90 (n/a)	72.70 (n/a)	17.22 (n/a)

eltwise_add_8_cols_2_channels_2048_tile_256

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`0685306` — 2026-02-06 13:45:43	0.06 (-60.17%)	0.06 (-57.97%)	0.06 (-58.27%)	0.05 (-54.13%)	0.01 (-70.52%)	238.70 (+117.99%)	212.26 (+135.84%)	206.00 (+139.53%)	189.10 (+151.13%)	21.65 (+61.10%)
`1531781` — 2026-01-23 18:49:24	0.16 (n/a)	0.14 (n/a)	0.14 (n/a)	0.11 (n/a)	0.02 (n/a)	109.50 (n/a)	90.00 (n/a)	86.00 (n/a)	75.30 (n/a)	13.44 (n/a)

eltwise_mul_1_cols_2_channels_2048_tile_2048

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`0685306` — 2026-02-06 13:45:43	0.12 (-13.41%)	0.10 (-17.68%)	0.10 (-25.32%)	0.09 (-8.77%)	0.01 (-25.81%)	129.50 (+9.65%)	118.54 (+21.11%)	124.00 (+33.91%)	104.40 (+15.49%)	10.71 (-7.24%)
`1531781` — 2026-01-23 18:49:24	0.14 (n/a)	0.13 (n/a)	0.13 (n/a)	0.10 (n/a)	0.01 (n/a)	118.10 (n/a)	97.88 (n/a)	92.60 (n/a)	90.40 (n/a)	11.55 (n/a)

eltwise_mul_2_cols_2_channels_2048_tile_1024

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`0685306` — 2026-02-06 13:45:43	0.12 (-11.76%)	0.09 (-27.34%)	0.08 (-41.04%)	0.06 (-37.80%)	0.03 (+49.43%)	209.10 (+60.85%)	147.76 (+44.72%)	158.30 (+69.67%)	103.20 (+13.28%)	43.22 (+161.49%)
`1531781` — 2026-01-23 18:49:24	0.13 (n/a)	0.12 (n/a)	0.13 (n/a)	0.09 (n/a)	0.02 (n/a)	130.00 (n/a)	102.10 (n/a)	93.30 (n/a)	91.10 (n/a)	16.53 (n/a)

eltwise_mul_4_cols_2_channels_2048_tile_512

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`0685306` — 2026-02-06 13:45:43	0.12 (-28.79%)	0.09 (-36.28%)	0.09 (-37.08%)	0.07 (-37.15%)	0.02 (-18.13%)	181.40 (+59.12%)	136.58 (+58.48%)	129.50 (+58.90%)	104.20 (+40.43%)	28.44 (+79.39%)
`1531781` — 2026-01-23 18:49:24	0.17 (n/a)	0.15 (n/a)	0.15 (n/a)	0.11 (n/a)	0.02 (n/a)	114.00 (n/a)	86.18 (n/a)	81.50 (n/a)	74.20 (n/a)	15.85 (n/a)

eltwise_mul_8_cols_2_channels_2048_tile_256

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`0685306` — 2026-02-06 13:45:43	0.08 (-41.35%)	0.07 (-36.97%)	0.07 (-39.75%)	0.07 (-16.75%)	0.00 (-81.53%)	186.00 (+20.16%)	173.24 (+54.29%)	173.60 (+65.97%)	162.80 (+70.47%)	8.84 (-63.38%)
`1531781` — 2026-01-23 18:49:24	0.13 (n/a)	0.11 (n/a)	0.12 (n/a)	0.08 (n/a)	0.02 (n/a)	154.80 (n/a)	112.28 (n/a)	104.60 (n/a)	95.50 (n/a)	24.14 (n/a)

gelu_1_cols_1_channels_2048_tile_2048

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`0685306` — 2026-02-06 13:45:43	0.05 (-55.97%)	0.05 (-52.56%)	0.05 (-58.03%)	0.05 (-43.84%)	0.00 (-87.45%)	181.00 (+78.15%)	172.32 (+105.63%)	172.60 (+138.40%)	164.00 (+127.15%)	8.02 (-49.08%)
`1531781` — 2026-01-23 18:49:24	0.11 (n/a)	0.10 (n/a)	0.11 (n/a)	0.08 (n/a)	0.02 (n/a)	101.60 (n/a)	83.80 (n/a)	72.40 (n/a)	72.20 (n/a)	15.75 (n/a)

gelu_1_cols_2_channels_2048_tile_1024

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`0685306` — 2026-02-06 13:45:43	0.07 (-43.26%)	0.05 (-48.60%)	0.05 (-55.23%)	0.04 (-46.74%)	0.01 (-40.98%)	206.50 (+87.73%)	157.00 (+95.61%)	154.30 (+123.30%)	117.10 (+76.36%)	36.26 (+95.63%)
`1531781` — 2026-01-23 18:49:24	0.12 (n/a)	0.11 (n/a)	0.12 (n/a)	0.07 (n/a)	0.02 (n/a)	110.00 (n/a)	80.26 (n/a)	69.10 (n/a)	66.40 (n/a)	18.54 (n/a)

gelu_2_cols_1_channels_2048_tile_1024

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`0685306` — 2026-02-06 13:45:43	0.05 (-57.90%)	0.04 (-60.48%)	0.04 (-67.37%)	0.04 (-38.98%)	0.01 (-78.68%)	225.80 (+63.86%)	202.28 (+137.92%)	209.10 (+206.60%)	160.90 (+137.32%)	24.40 (-19.32%)
`1531781` — 2026-01-23 18:49:24	0.12 (n/a)	0.10 (n/a)	0.12 (n/a)	0.06 (n/a)	0.03 (n/a)	137.80 (n/a)	85.02 (n/a)	68.20 (n/a)	67.80 (n/a)	30.24 (n/a)

gelu_2_cols_2_channels_2048_tile_512

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`0685306` — 2026-02-06 13:45:43	0.06 (-47.12%)	0.05 (-54.85%)	0.05 (-58.98%)	0.04 (-46.72%)	0.01 (-53.43%)	202.40 (+87.76%)	168.64 (+119.58%)	169.50 (+143.88%)	126.80 (+88.97%)	27.32 (+57.02%)
`1531781` — 2026-01-23 18:49:24	0.12 (n/a)	0.11 (n/a)	0.12 (n/a)	0.08 (n/a)	0.02 (n/a)	107.80 (n/a)	76.80 (n/a)	69.50 (n/a)	67.10 (n/a)	17.40 (n/a)

gelu_4_cols_1_channels_2048_tile_512

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`0685306` — 2026-02-06 13:45:43	0.07 (-49.54%)	0.06 (-48.16%)	0.05 (-53.80%)	0.04 (-18.22%)	0.01 (-70.74%)	187.20 (+22.27%)	151.74 (+77.35%)	150.00 (+116.45%)	124.90 (+98.25%)	25.15 (-33.72%)
`1531781` — 2026-01-23 18:49:24	0.13 (n/a)	0.11 (n/a)	0.12 (n/a)	0.05 (n/a)	0.03 (n/a)	153.10 (n/a)	85.56 (n/a)	69.30 (n/a)	63.00 (n/a)	37.94 (n/a)

gelu_4_cols_2_channels_2048_tile_256

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`0685306` — 2026-02-06 13:45:43	0.06 (-46.79%)	0.05 (-53.61%)	0.05 (-59.90%)	0.04 (-49.03%)	0.01 (-42.21%)	216.70 (+96.11%)	173.08 (+116.89%)	175.70 (+149.22%)	131.60 (+88.00%)	36.72 (+110.03%)
`1531781` — 2026-01-23 18:49:24	0.12 (n/a)	0.11 (n/a)	0.12 (n/a)	0.07 (n/a)	0.02 (n/a)	110.50 (n/a)	79.80 (n/a)	70.50 (n/a)	70.00 (n/a)	17.48 (n/a)

gelu_8_cols_1_channels_2048_tile_256

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`0685306` — 2026-02-06 13:45:43	0.04 (-64.29%)	0.04 (-61.54%)	0.04 (-67.06%)	0.04 (-9.86%)	0.00 (-92.29%)	231.00 (+10.95%)	214.44 (+117.18%)	218.20 (+203.48%)	196.70 (+180.20%)	14.36 (-76.53%)
`1531781` — 2026-01-23 18:49:24	0.12 (n/a)	0.10 (n/a)	0.11 (n/a)	0.04 (n/a)	0.03 (n/a)	208.20 (n/a)	98.74 (n/a)	71.90 (n/a)	70.20 (n/a)	61.20 (n/a)

gelu_8_cols_2_channels_2048_tile_128

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`0685306` — 2026-02-06 13:45:43	0.04 (-60.88%)	0.03 (-63.49%)	0.04 (-64.08%)	0.02 (-69.30%)	0.01 (-51.12%)	434.60 (+225.79%)	265.54 (+184.18%)	229.10 (+178.37%)	209.30 (+155.56%)	94.99 (+322.81%)
`1531781` — 2026-01-23 18:49:24	0.10 (n/a)	0.09 (n/a)	0.10 (n/a)	0.06 (n/a)	0.02 (n/a)	133.40 (n/a)	93.44 (n/a)	82.30 (n/a)	81.90 (n/a)	22.47 (n/a)

gemm_1792x896x1152_64x32x48_8cols_ccolmaj

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)	Throughput (max)	Throughput (mean)	Throughput (median)	Throughput (min)	Throughput (stddev)
`0685306` — 2026-02-06 13:45:43	5.15 (-0.00%)	4.57 (-6.69%)	4.46 (-10.52%)	4.17 (-2.35%)	0.37 (+3.56%)	2254.70 (+2.41%)	2068.96 (+7.20%)	2106.70 (+11.76%)	1826.80 (+0.01%)	159.79 (+3.73%)	2025.10 (-0.00%)	1796.98 (-6.69%)	1756.03 (-10.52%)	1640.72 (-2.35%)	144.95 (+3.56%)
`1531781` — 2026-01-23 18:49:24	5.15 (n/a)	4.90 (n/a)	4.99 (n/a)	4.27 (n/a)	0.36 (n/a)	2201.70 (n/a)	1929.94 (n/a)	1885.00 (n/a)	1826.70 (n/a)	154.05 (n/a)	2025.19 (n/a)	1925.76 (n/a)	1962.51 (n/a)	1680.20 (n/a)	139.97 (n/a)

gemm_192x384x64_48x96x16_4cols

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)	Throughput (max)	Throughput (mean)	Throughput (median)	Throughput (min)	Throughput (stddev)
`0685306` — 2026-02-06 13:45:43	1.03 (-56.26%)	0.74 (-60.05%)	0.65 (-63.34%)	0.54 (-58.75%)	0.22 (-43.13%)	406.70 (+142.37%)	320.66 (+158.10%)	339.30 (+172.75%)	214.30 (+128.71%)	89.48 (+218.28%)	44.03 (-56.26%)	31.54 (-60.05%)	27.82 (-63.34%)	23.20 (-58.75%)	9.52 (-43.13%)
`1531781` — 2026-01-23 18:49:24	2.36 (n/a)	1.85 (n/a)	1.78 (n/a)	1.32 (n/a)	0.39 (n/a)	167.80 (n/a)	124.24 (n/a)	124.40 (n/a)	93.70 (n/a)	28.11 (n/a)	100.67 (n/a)	78.94 (n/a)	75.88 (n/a)	56.25 (n/a)	16.73 (n/a)

gemm_192x384x64_48x96x16_4cols_bcolmaj_ccolmaj

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)	Throughput (max)	Throughput (mean)	Throughput (median)	Throughput (min)	Throughput (stddev)
`0685306` — 2026-02-06 13:45:43	1.52 (-31.37%)	1.08 (-47.70%)	0.98 (-55.54%)	0.83 (-48.11%)	0.27 (+1.61%)	266.30 (+92.69%)	214.48 (+96.41%)	226.00 (+124.88%)	145.40 (+45.69%)	45.48 (+174.98%)	64.89 (-31.37%)	45.91 (-47.70%)	41.75 (-55.54%)	35.44 (-48.11%)	11.44 (+1.61%)
`1531781` — 2026-01-23 18:49:24	2.22 (n/a)	2.06 (n/a)	2.20 (n/a)	1.60 (n/a)	0.26 (n/a)	138.20 (n/a)	109.20 (n/a)	100.50 (n/a)	99.80 (n/a)	16.54 (n/a)	94.54 (n/a)	87.77 (n/a)	93.91 (n/a)	68.30 (n/a)	11.26 (n/a)

gemm_2048x2048x2048_64x64x32_8_cols_0_bcolmaj_0_ccolmaj_0

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	GFLOP/s (max)	GFLOP/s (mean)	GFLOP/s (median)	GFLOP/s (min)	GFLOP/s (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`3a053d3` — 2025-11-26 01:12:16	3.72 (-0.11%)	3.55 (-2.75%)	3.56 (-2.75%)	3.36 (-5.33%)	0.12 (+76.10%)	2725.23 (-0.11%)	2602.52 (-2.75%)	2612.11 (-2.75%)	2459.89 (-5.33%)	86.27 (+76.10%)	6984.00 (+5.63%)	6608.60 (+2.91%)	6577.00 (+2.83%)	6304.00 (+0.11%)	221.87 (+87.15%)
`9e98049` — 2025-11-25 20:50:55	3.72 (n/a)	3.65 (n/a)	3.66 (n/a)	3.54 (n/a)	0.07 (n/a)	2728.26 (n/a)	2676.06 (n/a)	2686.03 (n/a)	2598.29 (n/a)	48.99 (n/a)	6612.00 (n/a)	6422.00 (n/a)	6396.00 (n/a)	6297.00 (n/a)	118.55 (n/a)

gemm_2048x2048x2048_64x64x32_8_cols_0_bcolmaj_1_ccolmaj_0

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	GFLOP/s (max)	GFLOP/s (mean)	GFLOP/s (median)	GFLOP/s (min)	GFLOP/s (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`3a053d3` — 2025-11-26 01:12:16	3.40 (-3.14%)	3.38 (-0.82%)	3.38 (-0.82%)	3.35 (+4.93%)	0.02 (-84.51%)	2493.81 (-3.14%)	2475.27 (-0.82%)	2479.06 (-0.82%)	2453.92 (+4.93%)	13.33 (-84.51%)	7001.00 (-4.70%)	6940.80 (+0.71%)	6930.00 (+0.83%)	6889.00 (+3.24%)	37.43 (-84.76%)
`9e98049` — 2025-11-25 20:50:55	3.51 (n/a)	3.40 (n/a)	3.41 (n/a)	3.19 (n/a)	0.12 (n/a)	2574.53 (n/a)	2495.72 (n/a)	2499.62 (n/a)	2338.67 (n/a)	86.02 (n/a)	7346.00 (n/a)	6892.20 (n/a)	6873.00 (n/a)	6673.00 (n/a)	245.57 (n/a)

gemm_2048x2048x2048_64x64x32_8_cols_1_bcolmaj_0_ccolmaj_0

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	GFLOP/s (max)	GFLOP/s (mean)	GFLOP/s (median)	GFLOP/s (min)	GFLOP/s (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`3a053d3` — 2025-11-26 01:12:16	3.02 (+0.52%)	2.87 (-2.05%)	2.86 (-2.24%)	2.70 (-6.16%)	0.11 (+131.66%)	2216.19 (+0.52%)	2104.44 (-2.05%)	2094.08 (-2.24%)	1980.84 (-6.16%)	80.25 (+131.67%)	8673.00 (+6.56%)	8175.60 (+2.21%)	8204.00 (+2.29%)	7752.00 (-0.51%)	314.03 (+145.36%)
`9e98049` — 2025-11-25 20:50:55	3.01 (n/a)	2.93 (n/a)	2.92 (n/a)	2.88 (n/a)	0.05 (n/a)	2204.81 (n/a)	2148.41 (n/a)	2142.13 (n/a)	2110.81 (n/a)	34.64 (n/a)	8139.00 (n/a)	7998.60 (n/a)	8020.00 (n/a)	7792.00 (n/a)	127.99 (n/a)

gemm_2048x2048x2048_64x64x64_1cols

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)	Throughput (max)	Throughput (mean)	Throughput (median)	Throughput (min)	Throughput (stddev)
`0685306` — 2026-02-06 13:45:43	0.55 (+5.08%)	0.54 (+4.51%)	0.54 (+4.17%)	0.54 (+4.25%)	0.00 (+671.17%)	46441.60 (-4.08%)	46284.84 (-4.31%)	46426.30 (-4.00%)	46008.60 (-4.83%)	206.97 (+604.09%)	373.41 (+5.08%)	371.18 (+4.51%)	370.05 (+4.17%)	369.92 (+4.25%)	1.66 (+671.14%)
`1531781` — 2026-01-23 18:49:24	0.52 (n/a)	0.52 (n/a)	0.52 (n/a)	0.52 (n/a)	0.00 (n/a)	48415.10 (n/a)	48370.52 (n/a)	48362.20 (n/a)	48344.40 (n/a)	29.40 (n/a)	355.36 (n/a)	355.17 (n/a)	355.23 (n/a)	354.85 (n/a)	0.22 (n/a)

gemm_2048x2048x2048_64x64x64_2_cols_0_bcolmaj_0_ccolmaj_0

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`130b6ea` — 2025-12-05 21:15:49	1.04 (+0.94%)	1.03 (+0.23%)	1.03 (+0.09%)	1.02 (+0.07%)	0.00 (+197.51%)	24576.00 (-0.07%)	24485.82 (-0.23%)	24530.40 (-0.09%)	24266.40 (-0.94%)	111.26 (+194.35%)
`0a6c11c` — 2025-12-04 00:39:10	1.03 (n/a)	1.03 (n/a)	1.03 (n/a)	1.02 (n/a)	0.00 (n/a)	24594.30 (n/a)	24542.14 (n/a)	24551.90 (n/a)	24495.50 (n/a)	37.80 (n/a)

gemm_2048x2048x2048_64x64x64_2_cols_0_bcolmaj_0_ccolmaj_0_0

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)	Throughput (max)	Throughput (mean)	Throughput (median)	Throughput (min)	Throughput (stddev)
`a4b6ffe` — 2026-01-05 19:24:05	1.04 (+0.28%)	1.03 (-0.15%)	1.03 (+0.05%)	1.02 (-0.52%)	0.01 (+42.72%)	24719.80 (+0.52%)	24512.42 (+0.16%)	24529.20 (-0.05%)	24092.60 (-0.28%)	252.92 (+43.15%)	713.08 (+0.28%)	700.92 (-0.15%)	700.38 (+0.05%)	694.98 (-0.52%)	7.30 (+42.71%)
`50ad37c` — 2025-12-29 20:35:53	1.04 (n/a)	1.03 (n/a)	1.03 (n/a)	1.02 (n/a)	0.01 (n/a)	24592.30 (n/a)	24473.70 (n/a)	24540.90 (n/a)	24160.80 (n/a)	176.68 (n/a)	711.06 (n/a)	702.00 (n/a)	700.05 (n/a)	698.59 (n/a)	5.12 (n/a)

gemm_2048x2048x2048_64x64x64_2_cols_0_bcolmaj_1_ccolmaj_0

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`130b6ea` — 2025-12-05 21:15:49	0.99 (+1.69%)	0.97 (+0.29%)	0.97 (-0.04%)	0.97 (-0.15%)	0.01 (+434.92%)	26025.30 (+0.15%)	25839.34 (-0.29%)	25916.70 (+0.04%)	25446.00 (-1.67%)	202.05 (+425.44%)
`0a6c11c` — 2025-12-04 00:39:10	0.97 (n/a)	0.97 (n/a)	0.97 (n/a)	0.97 (n/a)	0.00 (n/a)	25986.90 (n/a)	25913.52 (n/a)	25905.70 (n/a)	25877.00 (n/a)	38.45 (n/a)

gemm_2048x2048x2048_64x64x64_2_cols_0_bcolmaj_1_ccolmaj_0_0

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)	Throughput (max)	Throughput (mean)	Throughput (median)	Throughput (min)	Throughput (stddev)
`a4b6ffe` — 2026-01-05 19:24:05	0.97 (-0.01%)	0.97 (+0.03%)	0.97 (+0.04%)	0.97 (+0.00%)	0.00 (-6.43%)	25927.90 (-0.00%)	25903.20 (-0.03%)	25898.60 (-0.04%)	25892.80 (+0.01%)	14.05 (-6.56%)	663.50 (-0.01%)	663.23 (+0.03%)	663.35 (+0.04%)	662.60 (+0.00%)	0.36 (-6.43%)
`50ad37c` — 2025-12-29 20:35:53	0.97 (n/a)	0.97 (n/a)	0.97 (n/a)	0.97 (n/a)	0.00 (n/a)	25928.10 (n/a)	25910.14 (n/a)	25908.10 (n/a)	25891.20 (n/a)	15.04 (n/a)	663.54 (n/a)	663.06 (n/a)	663.11 (n/a)	662.60 (n/a)	0.38 (n/a)

gemm_2048x2048x2048_64x64x64_2_cols_1_bcolmaj_0_ccolmaj_0

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`130b6ea` — 2025-12-05 21:15:49	0.92 (-1.02%)	0.92 (-0.66%)	0.92 (-0.10%)	0.91 (-0.78%)	0.01 (-4.54%)	27751.40 (+0.79%)	27480.74 (+0.66%)	27354.80 (+0.10%)	27320.50 (+1.03%)	182.69 (-2.84%)
`0a6c11c` — 2025-12-04 00:39:10	0.93 (n/a)	0.92 (n/a)	0.92 (n/a)	0.91 (n/a)	0.01 (n/a)	27534.80 (n/a)	27299.32 (n/a)	27326.70 (n/a)	27042.30 (n/a)	188.04 (n/a)

gemm_2048x2048x2048_64x64x64_2_cols_1_bcolmaj_0_ccolmaj_0_0

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)	Throughput (max)	Throughput (mean)	Throughput (median)	Throughput (min)	Throughput (stddev)
`a4b6ffe` — 2026-01-05 19:24:05	0.92 (+0.11%)	0.91 (-0.12%)	0.91 (-0.21%)	0.91 (-0.24%)	0.00 (+30.34%)	27723.50 (+0.24%)	27548.94 (+0.12%)	27522.90 (+0.21%)	27403.60 (-0.11%)	124.26 (+30.58%)	626.92 (+0.11%)	623.62 (-0.12%)	624.20 (-0.21%)	619.69 (-0.24%)	2.81 (+30.34%)
`50ad37c` — 2025-12-29 20:35:53	0.92 (n/a)	0.91 (n/a)	0.92 (n/a)	0.91 (n/a)	0.00 (n/a)	27657.10 (n/a)	27515.58 (n/a)	27464.70 (n/a)	27434.60 (n/a)	95.16 (n/a)	626.21 (n/a)	624.37 (n/a)	625.52 (n/a)	621.17 (n/a)	2.16 (n/a)

gemm_2048x2048x2048_64x64x64_2cols_bcolmaj

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)	Throughput (max)	Throughput (mean)	Throughput (median)	Throughput (min)	Throughput (stddev)
`0685306` — 2026-02-06 13:45:43	0.96 (+3.81%)	0.94 (+3.30%)	0.95 (+3.66%)	0.93 (+2.50%)	0.01 (+58.17%)	27149.20 (-2.43%)	26656.06 (-3.19%)	26547.00 (-3.54%)	26110.20 (-3.67%)	405.59 (+48.92%)	657.98 (+3.81%)	644.62 (+3.30%)	647.15 (+3.66%)	632.79 (+2.50%)	9.82 (+58.17%)
`1531781` — 2026-01-23 18:49:24	0.93 (n/a)	0.91 (n/a)	0.91 (n/a)	0.90 (n/a)	0.01 (n/a)	27826.70 (n/a)	27533.84 (n/a)	27519.90 (n/a)	27105.20 (n/a)	272.35 (n/a)	633.82 (n/a)	624.00 (n/a)	624.27 (n/a)	617.39 (n/a)	6.21 (n/a)

gemm_2048x2048x2048_64x64x64_8_cols_0_bcolmaj_0_ccolmaj_0

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	GFLOP/s (max)	GFLOP/s (mean)	GFLOP/s (median)	GFLOP/s (min)	GFLOP/s (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`31ab24e` — 2025-11-22 02:01:56	3.65 (-1.06%)	3.59 (+0.83%)	3.63 (-0.26%)	3.45 (+2.66%)	0.07 (-44.71%)	2678.08 (-1.06%)	2632.13 (+0.83%)	2658.60 (-0.26%)	2528.31 (+2.66%)	53.86 (-44.71%)	6795.00 (-2.59%)	6529.80 (-0.92%)	6462.00 (+0.26%)	6415.00 (+1.07%)	137.14 (-45.27%)
`d9a64c5` — 2025-11-21 20:49:35	3.69 (n/a)	3.56 (n/a)	3.64 (n/a)	3.36 (n/a)	0.13 (n/a)	2706.77 (n/a)	2610.50 (n/a)	2665.61 (n/a)	2462.71 (n/a)	97.41 (n/a)	6976.00 (n/a)	6590.40 (n/a)	6445.00 (n/a)	6347.00 (n/a)	250.58 (n/a)

gemm_2048x2048x2048_64x64x64_8_cols_0_bcolmaj_0_ccolmaj_0_0

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)	Throughput (max)	Throughput (mean)	Throughput (median)	Throughput (min)	Throughput (stddev)
`d48746f` — 2026-01-12 22:17:10	3.98 (+0.20%)	3.84 (-2.95%)	3.76 (-5.27%)	3.72 (-4.97%)	0.13 (+473.97%)	6760.00 (+5.23%)	6559.50 (+3.13%)	6697.10 (+5.56%)	6321.00 (-0.20%)	218.68 (+499.40%)	2717.91 (+0.20%)	2621.44 (-2.95%)	2565.27 (-5.27%)	2541.40 (-4.97%)	88.37 (+473.97%)
`331dcca` — 2026-01-08 18:34:10	3.97 (n/a)	3.96 (n/a)	3.97 (n/a)	3.92 (n/a)	0.02 (n/a)	6423.80 (n/a)	6360.32 (n/a)	6344.30 (n/a)	6333.90 (n/a)	36.48 (n/a)	2712.39 (n/a)	2701.17 (n/a)	2707.91 (n/a)	2674.40 (n/a)	15.40 (n/a)

gemm_2048x2048x2048_64x64x64_8_cols_0_bcolmaj_1_ccolmaj_0_0

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)	Throughput (max)	Throughput (mean)	Throughput (median)	Throughput (min)	Throughput (stddev)
`d48746f` — 2026-01-12 22:17:10	3.75 (+0.28%)	3.70 (+1.25%)	3.71 (-0.21%)	3.65 (+4.77%)	0.04 (-66.90%)	6901.80 (-4.56%)	6795.46 (-1.30%)	6790.20 (+0.21%)	6715.90 (-0.28%)	67.05 (-68.40%)	2558.08 (+0.28%)	2528.33 (+1.25%)	2530.10 (-0.21%)	2489.18 (+4.77%)	24.81 (-66.90%)
`331dcca` — 2026-01-08 18:34:10	3.74 (n/a)	3.66 (n/a)	3.71 (n/a)	3.48 (n/a)	0.11 (n/a)	7231.20 (n/a)	6885.18 (n/a)	6775.70 (n/a)	6734.90 (n/a)	212.21 (n/a)	2550.87 (n/a)	2497.04 (n/a)	2535.50 (n/a)	2375.78 (n/a)	74.96 (n/a)

gemm_2048x2048x2048_64x64x64_8_cols_1_bcolmaj_0_ccolmaj_0_0

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)	Throughput (max)	Throughput (mean)	Throughput (median)	Throughput (min)	Throughput (stddev)
`d48746f` — 2026-01-12 22:17:10	3.62 (-0.77%)	3.54 (+0.14%)	3.60 (-0.04%)	3.44 (+2.17%)	0.09 (-25.77%)	7307.40 (-2.13%)	7103.46 (-0.18%)	6982.50 (+0.04%)	6948.20 (+0.77%)	182.95 (-26.55%)	2472.55 (-0.77%)	2419.80 (+0.14%)	2460.41 (-0.04%)	2351.04 (+2.17%)	61.76 (-25.77%)
`331dcca` — 2026-01-08 18:34:10	3.65 (n/a)	3.54 (n/a)	3.61 (n/a)	3.37 (n/a)	0.12 (n/a)	7466.20 (n/a)	7116.40 (n/a)	6979.60 (n/a)	6895.10 (n/a)	249.09 (n/a)	2491.61 (n/a)	2416.45 (n/a)	2461.45 (n/a)	2301.00 (n/a)	83.20 (n/a)

gemm_2048x2048x2048_64x64x64_8cols_bcolmaj_ccolmaj

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)	Throughput (max)	Throughput (mean)	Throughput (median)	Throughput (min)	Throughput (stddev)
`0685306` — 2026-02-06 13:45:43	3.46 (-0.11%)	3.36 (-1.65%)	3.33 (-3.35%)	3.27 (-0.71%)	0.08 (+12.78%)	7705.90 (+0.71%)	7492.40 (+1.68%)	7553.60 (+3.46%)	7271.50 (+0.11%)	184.00 (+13.21%)	2362.62 (-0.11%)	2294.08 (-1.65%)	2274.40 (-3.35%)	2229.44 (-0.71%)	56.56 (+12.78%)
`1531781` — 2026-01-23 18:49:24	3.46 (n/a)	3.42 (n/a)	3.45 (n/a)	3.29 (n/a)	0.07 (n/a)	7651.20 (n/a)	7368.34 (n/a)	7300.80 (n/a)	7263.70 (n/a)	162.53 (n/a)	2365.17 (n/a)	2332.46 (n/a)	2353.14 (n/a)	2245.39 (n/a)	50.15 (n/a)

gemm_384x1536x1792_32x48x64_4cols_bcolmaj

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)	Throughput (max)	Throughput (mean)	Throughput (median)	Throughput (min)	Throughput (stddev)
`0685306` — 2026-02-06 13:45:43	4.47 (+2.72%)	4.11 (-2.96%)	4.09 (-5.49%)	3.79 (-1.55%)	0.30 (+39.05%)	2127.90 (+1.58%)	1967.92 (+3.27%)	1970.30 (+5.82%)	1802.90 (-2.65%)	145.31 (+37.24%)	1172.50 (+2.72%)	1078.92 (-2.96%)	1072.90 (-5.50%)	993.44 (-1.55%)	79.97 (+39.05%)
`1531781` — 2026-01-23 18:49:24	4.35 (n/a)	4.24 (n/a)	4.33 (n/a)	3.85 (n/a)	0.22 (n/a)	2094.90 (n/a)	1905.64 (n/a)	1862.00 (n/a)	1852.00 (n/a)	105.88 (n/a)	1141.45 (n/a)	1111.86 (n/a)	1135.29 (n/a)	1009.08 (n/a)	57.51 (n/a)

gemm_896x1792x640_32x64x80_8cols_ccolmaj

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)	Throughput (max)	Throughput (mean)	Throughput (median)	Throughput (min)	Throughput (stddev)
`0685306` — 2026-02-06 13:45:43	6.25 (-8.08%)	4.90 (-23.81%)	4.82 (-28.08%)	3.86 (-27.56%)	0.87 (+40.10%)	1721.60 (+38.05%)	1389.78 (+33.29%)	1380.70 (+39.06%)	1064.30 (+8.79%)	235.47 (+105.49%)	1931.04 (-8.08%)	1514.46 (-23.81%)	1488.55 (-28.08%)	1193.76 (-27.56%)	267.35 (+40.10%)
`1531781` — 2026-01-23 18:49:24	6.80 (n/a)	6.43 (n/a)	6.70 (n/a)	5.33 (n/a)	0.62 (n/a)	1247.10 (n/a)	1042.70 (n/a)	992.90 (n/a)	978.30 (n/a)	114.59 (n/a)	2100.72 (n/a)	1987.76 (n/a)	2069.85 (n/a)	1647.93 (n/a)	190.83 (n/a)

layer_norm_1_cols_1_channels_2048_tile_2048

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`0685306` — 2026-02-06 13:45:43	0.06 (-43.84%)	0.06 (-35.98%)	0.05 (-32.96%)	0.05 (-29.66%)	0.01 (-65.31%)	170.20 (+42.19%)	147.40 (+51.27%)	150.00 (+49.25%)	127.30 (+78.04%)	19.27 (-14.98%)
`1531781` — 2026-01-23 18:49:24	0.11 (n/a)	0.09 (n/a)	0.08 (n/a)	0.07 (n/a)	0.02 (n/a)	119.70 (n/a)	97.44 (n/a)	100.50 (n/a)	71.50 (n/a)	22.67 (n/a)

layer_norm_1_cols_2_channels_2048_tile_1024

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`0685306` — 2026-02-06 13:45:43	0.08 (-26.82%)	0.05 (-38.30%)	0.05 (-35.87%)	0.03 (-43.08%)	0.02 (-2.89%)	242.70 (+75.74%)	170.02 (+72.12%)	151.40 (+55.92%)	104.40 (+36.65%)	61.32 (+146.47%)
`1531781` — 2026-01-23 18:49:24	0.11 (n/a)	0.09 (n/a)	0.08 (n/a)	0.06 (n/a)	0.02 (n/a)	138.10 (n/a)	98.78 (n/a)	97.10 (n/a)	76.40 (n/a)	24.88 (n/a)

layer_norm_2_cols_1_channels_2048_tile_1024

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`0685306` — 2026-02-06 13:45:43	0.07 (-40.39%)	0.06 (-36.26%)	0.06 (-31.75%)	0.04 (-33.28%)	0.01 (-41.28%)	210.00 (+49.89%)	153.40 (+55.93%)	138.30 (+46.50%)	114.20 (+67.69%)	38.40 (+47.10%)
`1531781` — 2026-01-23 18:49:24	0.12 (n/a)	0.09 (n/a)	0.09 (n/a)	0.06 (n/a)	0.02 (n/a)	140.10 (n/a)	98.38 (n/a)	94.40 (n/a)	68.10 (n/a)	26.11 (n/a)

layer_norm_2_cols_2_channels_2048_tile_512

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`0685306` — 2026-02-06 13:45:43	0.05 (-60.53%)	0.04 (-52.98%)	0.05 (-46.20%)	0.03 (-51.59%)	0.01 (-67.83%)	269.40 (+106.60%)	190.64 (+107.44%)	173.80 (+85.88%)	157.10 (+153.39%)	45.05 (+76.04%)
`1531781` — 2026-01-23 18:49:24	0.13 (n/a)	0.09 (n/a)	0.09 (n/a)	0.06 (n/a)	0.03 (n/a)	130.40 (n/a)	91.90 (n/a)	93.50 (n/a)	62.00 (n/a)	25.59 (n/a)

layer_norm_4_cols_1_channels_2048_tile_512

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`0685306` — 2026-02-06 13:45:43	0.07 (-41.42%)	0.05 (-46.77%)	0.05 (-56.03%)	0.03 (-46.44%)	0.01 (-33.70%)	263.20 (+86.67%)	174.84 (+91.00%)	177.10 (+127.34%)	124.70 (+70.82%)	56.38 (+98.34%)
`1531781` — 2026-01-23 18:49:24	0.11 (n/a)	0.09 (n/a)	0.11 (n/a)	0.06 (n/a)	0.02 (n/a)	141.00 (n/a)	91.54 (n/a)	77.90 (n/a)	73.00 (n/a)	28.43 (n/a)

layer_norm_4_cols_2_channels_2048_tile_256

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`0685306` — 2026-02-06 13:45:43	0.06 (-42.19%)	0.04 (-48.63%)	0.04 (-54.82%)	0.03 (-51.29%)	0.02 (-36.20%)	304.70 (+105.32%)	206.72 (+100.43%)	194.30 (+121.30%)	127.60 (+73.13%)	73.01 (+126.05%)
`1531781` — 2026-01-23 18:49:24	0.11 (n/a)	0.09 (n/a)	0.09 (n/a)	0.06 (n/a)	0.02 (n/a)	148.40 (n/a)	103.14 (n/a)	87.80 (n/a)	73.70 (n/a)	32.30 (n/a)

layer_norm_8_cols_1_channels_2048_tile_256

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`0685306` — 2026-02-06 13:45:43	0.06 (-50.04%)	0.05 (-52.27%)	0.05 (-57.45%)	0.04 (-22.87%)	0.01 (-69.47%)	199.60 (+29.69%)	176.96 (+95.84%)	181.90 (+135.01%)	134.70 (+100.15%)	26.78 (-25.33%)
`1531781` — 2026-01-23 18:49:24	0.12 (n/a)	0.10 (n/a)	0.11 (n/a)	0.05 (n/a)	0.03 (n/a)	153.90 (n/a)	90.36 (n/a)	77.40 (n/a)	67.30 (n/a)	35.86 (n/a)

layer_norm_8_cols_2_channels_2048_tile_128

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`0685306` — 2026-02-06 13:45:43	0.04 (-51.26%)	0.04 (-55.42%)	0.03 (-58.67%)	0.02 (-63.12%)	0.01 (-30.39%)	380.50 (+171.20%)	248.28 (+133.13%)	236.30 (+142.11%)	189.50 (+105.09%)	77.39 (+290.15%)
`1531781` — 2026-01-23 18:49:24	0.09 (n/a)	0.08 (n/a)	0.08 (n/a)	0.06 (n/a)	0.01 (n/a)	140.30 (n/a)	106.50 (n/a)	97.60 (n/a)	92.40 (n/a)	19.84 (n/a)

matrix_vector_mul_128x128_32_1col

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)
`130b6ea` — 2025-12-05 21:15:49	0.37 (+17.13%)	0.30 (+24.87%)	0.30 (+38.03%)	0.23 (+12.47%)	0.06 (+35.23%)
`0a6c11c` — 2025-12-04 00:39:10	0.32 (n/a)	0.24 (n/a)	0.22 (n/a)	0.20 (n/a)	0.04 (n/a)

matrix_vector_mul_128x128_32_1col0

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Throughput (max)	Throughput (mean)	Throughput (median)	Throughput (min)	Throughput (stddev)
`331dcca` — 2026-01-08 18:34:10	0.26 (-1.81%)	0.24 (+4.06%)	0.24 (+0.52%)	0.21 (+16.97%)	0.02 (-48.16%)	0.26 (-1.81%)	0.24 (+4.06%)	0.24 (+0.52%)	0.20 (+16.97%)	0.02 (-48.16%)
`a4b6ffe` — 2026-01-05 19:24:05	0.26 (n/a)	0.23 (n/a)	0.24 (n/a)	0.18 (n/a)	0.04 (n/a)	0.26 (n/a)	0.23 (n/a)	0.24 (n/a)	0.17 (n/a)	0.04 (n/a)

matrix_vector_mul_128x128_32tsi_128tso_1col0

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Throughput (max)	Throughput (mean)	Throughput (median)	Throughput (min)	Throughput (stddev)
`0685306` — 2026-02-06 13:45:43	0.28 (-36.38%)	0.24 (-29.81%)	0.25 (-28.18%)	0.20 (-9.91%)	0.04 (-49.83%)	0.28 (-36.38%)	0.24 (-29.81%)	0.25 (-28.18%)	0.20 (-9.91%)	0.04 (-49.83%)
`1531781` — 2026-01-23 18:49:24	0.45 (n/a)	0.35 (n/a)	0.35 (n/a)	0.22 (n/a)	0.08 (n/a)	0.44 (n/a)	0.34 (n/a)	0.34 (n/a)	0.22 (n/a)	0.08 (n/a)

matrix_vector_mul_2048x8192_1_1col

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)
`130b6ea` — 2025-12-05 21:15:49	13.57 (+3.47%)	13.15 (+1.09%)	13.10 (+0.16%)	12.52 (-1.58%)	0.37 (+153.19%)
`0a6c11c` — 2025-12-04 00:39:10	13.11 (n/a)	13.01 (n/a)	13.07 (n/a)	12.72 (n/a)	0.15 (n/a)

matrix_vector_mul_2048x8192_1_1col0

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Throughput (max)	Throughput (mean)	Throughput (median)	Throughput (min)	Throughput (stddev)
`331dcca` — 2026-01-08 18:34:10	13.43 (-0.07%)	13.30 (+4.62%)	13.27 (+1.29%)	13.16 (+16.70%)	0.12 (-86.69%)	13.42 (-0.07%)	13.29 (+4.62%)	13.26 (+1.29%)	13.16 (+16.70%)	0.12 (-86.69%)
`a4b6ffe` — 2026-01-05 19:24:05	13.44 (n/a)	12.71 (n/a)	13.10 (n/a)	11.28 (n/a)	0.87 (n/a)	13.43 (n/a)	12.70 (n/a)	13.09 (n/a)	11.27 (n/a)	0.87 (n/a)

matrix_vector_mul_2048x8192_1_2col

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)
`130b6ea` — 2025-12-05 21:15:49	25.06 (+2.47%)	23.71 (+1.81%)	24.13 (+0.50%)	20.90 (+5.41%)	1.48 (-15.60%)
`0a6c11c` — 2025-12-04 00:39:10	24.46 (n/a)	23.29 (n/a)	24.01 (n/a)	19.83 (n/a)	1.75 (n/a)

matrix_vector_mul_2048x8192_1_2col0

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Throughput (max)	Throughput (mean)	Throughput (median)	Throughput (min)	Throughput (stddev)
`331dcca` — 2026-01-08 18:34:10	24.46 (-0.60%)	21.07 (-8.03%)	19.75 (-17.83%)	18.08 (+1.18%)	3.03 (+7.07%)	24.44 (-0.60%)	21.06 (-8.03%)	19.74 (-17.83%)	18.07 (+1.18%)	3.03 (+7.07%)
`a4b6ffe` — 2026-01-05 19:24:05	24.60 (n/a)	22.91 (n/a)	24.04 (n/a)	17.87 (n/a)	2.83 (n/a)	24.59 (n/a)	22.90 (n/a)	24.02 (n/a)	17.86 (n/a)	2.83 (n/a)

matrix_vector_mul_2048x8192_1_4col

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)
`130b6ea` — 2025-12-05 21:15:49	42.05 (+1.54%)	40.50 (+12.60%)	40.49 (+2.36%)	38.98 (+97.53%)	0.97 (-88.09%)
`0a6c11c` — 2025-12-04 00:39:10	41.41 (n/a)	35.97 (n/a)	39.55 (n/a)	19.73 (n/a)	8.15 (n/a)

matrix_vector_mul_2048x8192_1_4col0

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Throughput (max)	Throughput (mean)	Throughput (median)	Throughput (min)	Throughput (stddev)
`331dcca` — 2026-01-08 18:34:10	42.58 (+1.51%)	41.37 (+14.29%)	41.25 (+4.89%)	40.33 (+91.42%)	0.92 (-89.18%)	42.56 (+1.51%)	41.34 (+14.29%)	41.23 (+4.89%)	40.30 (+91.42%)	0.92 (-89.18%)
`a4b6ffe` — 2026-01-05 19:24:05	41.95 (n/a)	36.20 (n/a)	39.33 (n/a)	21.07 (n/a)	8.54 (n/a)	41.92 (n/a)	36.17 (n/a)	39.31 (n/a)	21.05 (n/a)	8.53 (n/a)

matrix_vector_mul_2048x8192_1_8col

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)
`130b6ea` — 2025-12-05 21:15:49	46.62 (+3.11%)	44.82 (+0.17%)	44.78 (+0.17%)	42.60 (-3.80%)	1.60 (+367.72%)
`0a6c11c` — 2025-12-04 00:39:10	45.21 (n/a)	44.75 (n/a)	44.70 (n/a)	44.29 (n/a)	0.34 (n/a)

matrix_vector_mul_2048x8192_1_8col0

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Throughput (max)	Throughput (mean)	Throughput (median)	Throughput (min)	Throughput (stddev)
`331dcca` — 2026-01-08 18:34:10	46.04 (+4.14%)	45.06 (+3.47%)	44.55 (+2.76%)	44.30 (+2.84%)	0.87 (+66.58%)	46.01 (+4.14%)	45.03 (+3.47%)	44.52 (+2.76%)	44.27 (+2.84%)	0.87 (+66.58%)
`a4b6ffe` — 2026-01-05 19:24:05	44.21 (n/a)	43.55 (n/a)	43.35 (n/a)	43.07 (n/a)	0.52 (n/a)	44.18 (n/a)	43.52 (n/a)	43.32 (n/a)	43.05 (n/a)	0.52 (n/a)

matrix_vector_mul_2048x8192_1tsi_1024tso_2col0

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Throughput (max)	Throughput (mean)	Throughput (median)	Throughput (min)	Throughput (stddev)
`0685306` — 2026-02-06 13:45:43	24.91 (+0.41%)	24.23 (-1.48%)	24.15 (-1.61%)	23.90 (-2.06%)	0.41 (+129.88%)	24.89 (+0.41%)	24.21 (-1.48%)	24.14 (-1.61%)	23.88 (-2.06%)	0.40 (+129.88%)
`1531781` — 2026-01-23 18:49:24	24.80 (n/a)	24.59 (n/a)	24.55 (n/a)	24.40 (n/a)	0.18 (n/a)	24.79 (n/a)	24.58 (n/a)	24.53 (n/a)	24.38 (n/a)	0.18 (n/a)

matrix_vector_mul_2048x8192_1tsi_2048tso_1col0

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Throughput (max)	Throughput (mean)	Throughput (median)	Throughput (min)	Throughput (stddev)
`0685306` — 2026-02-06 13:45:43	13.51 (-0.23%)	12.83 (-3.86%)	13.21 (-1.74%)	10.91 (-16.75%)	1.09 (+427.10%)	13.50 (-0.23%)	12.82 (-3.86%)	13.20 (-1.74%)	10.90 (-16.75%)	1.09 (+427.10%)
`1531781` — 2026-01-23 18:49:24	13.54 (n/a)	13.34 (n/a)	13.45 (n/a)	13.10 (n/a)	0.21 (n/a)	13.53 (n/a)	13.33 (n/a)	13.44 (n/a)	13.09 (n/a)	0.21 (n/a)

matrix_vector_mul_2048x8192_1tsi_256tso_8col0

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Throughput (max)	Throughput (mean)	Throughput (median)	Throughput (min)	Throughput (stddev)
`0685306` — 2026-02-06 13:45:43	43.10 (-7.92%)	41.99 (-8.54%)	41.85 (-9.76%)	41.33 (-7.77%)	0.70 (-21.25%)	43.08 (-7.92%)	41.96 (-8.54%)	41.82 (-9.76%)	41.31 (-7.77%)	0.70 (-21.25%)
`1531781` — 2026-01-23 18:49:24	46.81 (n/a)	45.91 (n/a)	46.38 (n/a)	44.81 (n/a)	0.89 (n/a)	46.79 (n/a)	45.88 (n/a)	46.35 (n/a)	44.79 (n/a)	0.89 (n/a)

matrix_vector_mul_2048x8192_1tsi_512tso_4col0

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Throughput (max)	Throughput (mean)	Throughput (median)	Throughput (min)	Throughput (stddev)
`0685306` — 2026-02-06 13:45:43	43.32 (+4.40%)	38.47 (-5.74%)	39.46 (-4.06%)	31.40 (-21.41%)	4.40 (+576.86%)	43.29 (+4.40%)	38.45 (-5.74%)	39.43 (-4.06%)	31.38 (-21.41%)	4.40 (+576.86%)
`1531781` — 2026-01-23 18:49:24	41.49 (n/a)	40.81 (n/a)	41.13 (n/a)	39.95 (n/a)	0.65 (n/a)	41.46 (n/a)	40.79 (n/a)	41.10 (n/a)	39.93 (n/a)	0.65 (n/a)

matrix_vector_mul_8192x2048_4_1col

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)
`130b6ea` — 2025-12-05 21:15:49	13.29 (+1.63%)	13.19 (+6.08%)	13.19 (+1.17%)	13.09 (+26.86%)	0.08 (-92.94%)
`0a6c11c` — 2025-12-04 00:39:10	13.08 (n/a)	12.44 (n/a)	13.03 (n/a)	10.32 (n/a)	1.07 (n/a)

matrix_vector_mul_8192x2048_4_1col0

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Throughput (max)	Throughput (mean)	Throughput (median)	Throughput (min)	Throughput (stddev)
`331dcca` — 2026-01-08 18:34:10	13.68 (+3.42%)	13.30 (+8.54%)	13.38 (+7.25%)	12.72 (+14.96%)	0.35 (-66.09%)	13.67 (+3.42%)	13.29 (+8.54%)	13.38 (+7.25%)	12.71 (+14.96%)	0.35 (-66.09%)
`a4b6ffe` — 2026-01-05 19:24:05	13.23 (n/a)	12.25 (n/a)	12.48 (n/a)	11.07 (n/a)	1.04 (n/a)	13.22 (n/a)	12.24 (n/a)	12.47 (n/a)	11.06 (n/a)	1.04 (n/a)

matrix_vector_mul_8192x2048_4_2col

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)
`130b6ea` — 2025-12-05 21:15:49	25.58 (+4.24%)	24.23 (+6.59%)	24.03 (+0.29%)	23.58 (+37.29%)	0.70 (-74.97%)
`0a6c11c` — 2025-12-04 00:39:10	24.54 (n/a)	22.73 (n/a)	23.96 (n/a)	17.18 (n/a)	2.79 (n/a)

matrix_vector_mul_8192x2048_4_2col0

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Throughput (max)	Throughput (mean)	Throughput (median)	Throughput (min)	Throughput (stddev)
`331dcca` — 2026-01-08 18:34:10	25.08 (+1.32%)	24.68 (+13.42%)	24.61 (+3.26%)	24.47 (+48.32%)	0.24 (-93.56%)	25.07 (+1.32%)	24.67 (+13.42%)	24.59 (+3.26%)	24.46 (+48.32%)	0.24 (-93.56%)
`a4b6ffe` — 2026-01-05 19:24:05	24.76 (n/a)	21.76 (n/a)	23.83 (n/a)	16.50 (n/a)	3.74 (n/a)	24.74 (n/a)	21.75 (n/a)	23.82 (n/a)	16.49 (n/a)	3.73 (n/a)

matrix_vector_mul_8192x2048_4_4col

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)
`130b6ea` — 2025-12-05 21:15:49	40.94 (-2.19%)	39.74 (-1.10%)	39.76 (-0.60%)	38.17 (-2.14%)	1.04 (+4.39%)
`0a6c11c` — 2025-12-04 00:39:10	41.86 (n/a)	40.18 (n/a)	40.00 (n/a)	39.00 (n/a)	1.00 (n/a)

matrix_vector_mul_8192x2048_4_4col0

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Throughput (max)	Throughput (mean)	Throughput (median)	Throughput (min)	Throughput (stddev)
`331dcca` — 2026-01-08 18:34:10	40.74 (-1.09%)	37.16 (-7.15%)	40.25 (+1.47%)	24.51 (-37.44%)	7.07 (+736.13%)	40.72 (-1.09%)	37.13 (-7.15%)	40.23 (+1.47%)	24.50 (-37.44%)	7.07 (+736.13%)
`a4b6ffe` — 2026-01-05 19:24:05	41.19 (n/a)	40.02 (n/a)	39.67 (n/a)	39.18 (n/a)	0.85 (n/a)	41.17 (n/a)	39.99 (n/a)	39.64 (n/a)	39.16 (n/a)	0.85 (n/a)

matrix_vector_mul_8192x2048_4_8col

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)
`130b6ea` — 2025-12-05 21:15:49	46.71 (-1.04%)	44.91 (-2.34%)	44.71 (-2.98%)	43.80 (-1.44%)	0.96 (+6.93%)
`0a6c11c` — 2025-12-04 00:39:10	47.20 (n/a)	45.98 (n/a)	46.08 (n/a)	44.44 (n/a)	0.90 (n/a)

matrix_vector_mul_8192x2048_4_8col0

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Throughput (max)	Throughput (mean)	Throughput (median)	Throughput (min)	Throughput (stddev)
`331dcca` — 2026-01-08 18:34:10	46.28 (+5.68%)	44.89 (+14.59%)	44.74 (+4.26%)	44.20 (+62.80%)	0.84 (-87.96%)	46.25 (+5.68%)	44.86 (+14.59%)	44.71 (+4.26%)	44.17 (+62.80%)	0.84 (-87.96%)
`a4b6ffe` — 2026-01-05 19:24:05	43.79 (n/a)	39.17 (n/a)	42.91 (n/a)	27.15 (n/a)	6.98 (n/a)	43.77 (n/a)	39.15 (n/a)	42.88 (n/a)	27.13 (n/a)	6.98 (n/a)

matrix_vector_mul_8192x2048_4tsi_1024tso_1col0

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Throughput (max)	Throughput (mean)	Throughput (median)	Throughput (min)	Throughput (stddev)
`0685306` — 2026-02-06 13:45:43	13.40 (-0.42%)	13.11 (-1.59%)	13.20 (-0.88%)	12.47 (-5.52%)	0.37 (+228.29%)	13.40 (-0.42%)	13.11 (-1.59%)	13.19 (-0.88%)	12.46 (-5.52%)	0.37 (+228.29%)
`1531781` — 2026-01-23 18:49:24	13.46 (n/a)	13.33 (n/a)	13.32 (n/a)	13.20 (n/a)	0.11 (n/a)	13.45 (n/a)	13.32 (n/a)	13.31 (n/a)	13.19 (n/a)	0.11 (n/a)

matrix_vector_mul_8192x2048_4tsi_1024tso_2col0

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Throughput (max)	Throughput (mean)	Throughput (median)	Throughput (min)	Throughput (stddev)
`0685306` — 2026-02-06 13:45:43	24.31 (-3.79%)	23.79 (-3.89%)	23.81 (-3.26%)	23.03 (-6.20%)	0.47 (+59.41%)	24.29 (-3.79%)	23.77 (-3.89%)	23.79 (-3.26%)	23.01 (-6.20%)	0.47 (+59.41%)
`1531781` — 2026-01-23 18:49:24	25.27 (n/a)	24.75 (n/a)	24.61 (n/a)	24.55 (n/a)	0.30 (n/a)	25.25 (n/a)	24.74 (n/a)	24.59 (n/a)	24.53 (n/a)	0.30 (n/a)

matrix_vector_mul_8192x2048_4tsi_1024tso_4col0

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Throughput (max)	Throughput (mean)	Throughput (median)	Throughput (min)	Throughput (stddev)
`0685306` — 2026-02-06 13:45:43	41.32 (+1.13%)	39.70 (-0.76%)	39.31 (-2.76%)	38.56 (-0.78%)	1.07 (+20.25%)	41.30 (+1.13%)	39.68 (-0.76%)	39.29 (-2.76%)	38.54 (-0.78%)	1.06 (+20.25%)
`1531781` — 2026-01-23 18:49:24	40.86 (n/a)	40.01 (n/a)	40.43 (n/a)	38.86 (n/a)	0.89 (n/a)	40.84 (n/a)	39.98 (n/a)	40.41 (n/a)	38.84 (n/a)	0.89 (n/a)

matrix_vector_mul_8192x2048_4tsi_1024tso_8col0

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Throughput (max)	Throughput (mean)	Throughput (median)	Throughput (min)	Throughput (stddev)
`0685306` — 2026-02-06 13:45:43	43.77 (-5.01%)	40.93 (-9.25%)	42.02 (-7.62%)	36.70 (-16.38%)	3.02 (+212.65%)	43.75 (-5.01%)	40.91 (-9.25%)	42.00 (-7.62%)	36.68 (-16.38%)	3.02 (+212.65%)
`1531781` — 2026-01-23 18:49:24	46.08 (n/a)	45.11 (n/a)	45.49 (n/a)	43.89 (n/a)	0.97 (n/a)	46.05 (n/a)	45.08 (n/a)	45.46 (n/a)	43.87 (n/a)	0.97 (n/a)

mem_copy_16_cores_2_chans_2048_tile_128_False

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`130b6ea` — 2025-12-05 21:15:49	0.07 (+22.03%)	0.06 (+27.86%)	0.06 (+35.49%)	0.05 (+15.00%)	0.01 (+49.69%)	173.30 (-13.00%)	137.94 (-21.32%)	133.60 (-26.19%)	123.40 (-18.06%)	18.40 (+8.53%)
`0a6c11c` — 2025-12-04 00:39:10	0.05 (n/a)	0.05 (n/a)	0.05 (n/a)	0.04 (n/a)	0.00 (n/a)	199.20 (n/a)	175.32 (n/a)	181.00 (n/a)	150.60 (n/a)	16.96 (n/a)

mem_copy_16_cores_2_chans_2048_tile_128_False0

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`0685306` — 2026-02-06 13:45:43	0.06 (-41.52%)	0.05 (-41.37%)	0.05 (-48.03%)	0.03 (-17.75%)	0.01 (-65.69%)	237.50 (+21.61%)	176.58 (+58.20%)	166.70 (+92.49%)	144.70 (+71.04%)	35.40 (-25.72%)
`1531781` — 2026-01-23 18:49:24	0.10 (n/a)	0.08 (n/a)	0.09 (n/a)	0.04 (n/a)	0.02 (n/a)	195.30 (n/a)	111.62 (n/a)	86.60 (n/a)	84.60 (n/a)	47.65 (n/a)

mem_copy_1_cols_1_channels_2048_tile_2048

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`003b49d` — 2025-11-14 19:30:54	0.05 (-16.99%)	0.05 (-12.47%)	0.05 (-7.32%)	0.04 (-11.17%)	0.00 (-29.61%)	197.00 (+12.57%)	171.00 (+13.70%)	164.00 (+7.89%)	153.00 (+20.47%)	17.49 (-4.32%)
`0999a9a` — 2025-11-13 19:24:08	0.06 (n/a)	0.05 (n/a)	0.05 (n/a)	0.04 (n/a)	0.01 (n/a)	175.00 (n/a)	150.40 (n/a)	152.00 (n/a)	127.00 (n/a)	18.28 (n/a)

mem_copy_1_cols_2_channels_2048_tile_1024

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`003b49d` — 2025-11-14 19:30:54	0.06 (+5.56%)	0.05 (-1.59%)	0.05 (-1.39%)	0.03 (-14.07%)	0.01 (+25.03%)	263.00 (+16.37%)	166.00 (+4.53%)	144.00 (+1.41%)	126.00 (-5.26%)	49.38 (+43.17%)
`0999a9a` — 2025-11-13 19:24:08	0.06 (n/a)	0.05 (n/a)	0.05 (n/a)	0.03 (n/a)	0.01 (n/a)	226.00 (n/a)	158.80 (n/a)	142.00 (n/a)	133.00 (n/a)	34.49 (n/a)

mem_copy_1_cores_1_chans_2048_tile_2048_False

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`130b6ea` — 2025-12-05 21:15:49	0.08 (+10.93%)	0.07 (+7.79%)	0.07 (+3.55%)	0.05 (-1.73%)	0.01 (+36.73%)	150.20 (+1.76%)	121.68 (-6.36%)	125.60 (-3.38%)	97.90 (-9.85%)	18.56 (+21.98%)
`0a6c11c` — 2025-12-04 00:39:10	0.08 (n/a)	0.06 (n/a)	0.06 (n/a)	0.06 (n/a)	0.01 (n/a)	147.60 (n/a)	129.94 (n/a)	130.00 (n/a)	108.60 (n/a)	15.21 (n/a)

mem_copy_1_cores_1_chans_2048_tile_2048_False0

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`0685306` — 2026-02-06 13:45:43	0.06 (-42.13%)	0.04 (-48.78%)	0.04 (-54.42%)	0.03 (-52.23%)	0.01 (-33.43%)	247.20 (+109.31%)	191.50 (+98.40%)	185.30 (+119.55%)	142.30 (+72.69%)	44.35 (+142.48%)
`1531781` — 2026-01-23 18:49:24	0.10 (n/a)	0.09 (n/a)	0.10 (n/a)	0.07 (n/a)	0.02 (n/a)	118.10 (n/a)	96.52 (n/a)	84.40 (n/a)	82.40 (n/a)	18.29 (n/a)

mem_copy_2_cols_1_channels_2048_tile_1024

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`003b49d` — 2025-11-14 19:30:54	0.06 (-15.32%)	0.05 (-17.09%)	0.05 (-23.81%)	0.04 (-15.54%)	0.01 (-18.18%)	193.00 (+18.40%)	160.80 (+20.36%)	168.00 (+31.25%)	124.00 (+18.10%)	22.92 (+11.39%)
`0999a9a` — 2025-11-13 19:24:08	0.07 (n/a)	0.06 (n/a)	0.06 (n/a)	0.05 (n/a)	0.01 (n/a)	163.00 (n/a)	133.60 (n/a)	128.00 (n/a)	105.00 (n/a)	20.58 (n/a)

mem_copy_2_cols_2_channels_2048_tile_512

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`003b49d` — 2025-11-14 19:30:54	0.06 (-23.02%)	0.05 (-6.98%)	0.05 (-2.13%)	0.04 (-15.31%)	0.01 (-33.29%)	196.00 (+18.07%)	147.00 (+6.52%)	141.00 (+2.17%)	126.00 (+29.90%)	25.27 (+8.13%)
`0999a9a` — 2025-11-13 19:24:08	0.08 (n/a)	0.06 (n/a)	0.06 (n/a)	0.05 (n/a)	0.01 (n/a)	166.00 (n/a)	138.00 (n/a)	138.00 (n/a)	97.00 (n/a)	23.37 (n/a)

mem_copy_2_cores_1_chans_2048_tile_1024_False

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`130b6ea` — 2025-12-05 21:15:49	0.10 (+31.37%)	0.08 (+29.12%)	0.08 (+29.06%)	0.07 (+41.80%)	0.01 (+0.23%)	109.90 (-29.46%)	98.30 (-23.38%)	100.80 (-22.52%)	79.40 (-23.87%)	10.54 (-46.17%)
`0a6c11c` — 2025-12-04 00:39:10	0.08 (n/a)	0.07 (n/a)	0.06 (n/a)	0.05 (n/a)	0.01 (n/a)	155.80 (n/a)	128.30 (n/a)	130.10 (n/a)	104.30 (n/a)	19.58 (n/a)

mem_copy_2_cores_1_chans_2048_tile_1024_False0

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`0685306` — 2026-02-06 13:45:43	0.07 (-38.13%)	0.06 (-32.06%)	0.06 (-32.89%)	0.06 (-21.01%)	0.01 (-73.61%)	148.30 (+26.64%)	129.94 (+40.93%)	129.50 (+49.02%)	113.70 (+61.51%)	12.58 (-46.05%)
`1531781` — 2026-01-23 18:49:24	0.12 (n/a)	0.09 (n/a)	0.09 (n/a)	0.07 (n/a)	0.02 (n/a)	117.10 (n/a)	92.20 (n/a)	86.90 (n/a)	70.40 (n/a)	23.32 (n/a)

mem_copy_2_cores_2_chans_2048_tile_1024_False

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`130b6ea` — 2025-12-05 21:15:49	0.10 (+25.02%)	0.07 (+10.01%)	0.07 (+7.53%)	0.05 (-1.99%)	0.01 (+77.08%)	150.10 (+2.04%)	114.60 (-7.22%)	111.70 (-6.99%)	83.90 (-20.02%)	21.23 (+43.54%)
`0a6c11c` — 2025-12-04 00:39:10	0.08 (n/a)	0.07 (n/a)	0.07 (n/a)	0.06 (n/a)	0.01 (n/a)	147.10 (n/a)	123.52 (n/a)	120.10 (n/a)	104.90 (n/a)	14.79 (n/a)

mem_copy_2_cores_2_chans_2048_tile_1024_False0

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`0685306` — 2026-02-06 13:45:43	0.07 (-32.31%)	0.05 (-38.96%)	0.06 (-35.76%)	0.03 (-52.85%)	0.02 (-4.96%)	297.30 (+112.05%)	178.38 (+75.47%)	143.30 (+55.59%)	122.50 (+47.77%)	73.01 (+201.34%)
`1531781` — 2026-01-23 18:49:24	0.10 (n/a)	0.08 (n/a)	0.09 (n/a)	0.06 (n/a)	0.02 (n/a)	140.20 (n/a)	101.66 (n/a)	92.10 (n/a)	82.90 (n/a)	24.23 (n/a)

mem_copy_4_cols_1_channels_2048_tile_512

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`003b49d` — 2025-11-14 19:30:54	0.07 (+11.97%)	0.05 (+3.09%)	0.05 (+2.10%)	0.03 (-11.35%)	0.01 (+41.65%)	229.00 (+12.81%)	156.00 (-0.51%)	143.00 (-2.05%)	117.00 (-10.69%)	38.21 (+48.71%)
`0999a9a` — 2025-11-13 19:24:08	0.06 (n/a)	0.05 (n/a)	0.05 (n/a)	0.04 (n/a)	0.01 (n/a)	203.00 (n/a)	156.80 (n/a)	146.00 (n/a)	131.00 (n/a)	25.69 (n/a)

mem_copy_4_cols_2_channels_2048_tile_256

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`003b49d` — 2025-11-14 19:30:54	0.06 (+9.02%)	0.05 (+5.59%)	0.05 (+0.00%)	0.04 (+3.63%)	0.01 (+25.48%)	193.00 (-3.50%)	147.40 (-4.66%)	147.00 (+0.00%)	122.00 (-8.27%)	25.24 (+7.48%)
`0999a9a` — 2025-11-13 19:24:08	0.06 (n/a)	0.05 (n/a)	0.05 (n/a)	0.04 (n/a)	0.01 (n/a)	200.00 (n/a)	154.60 (n/a)	147.00 (n/a)	133.00 (n/a)	23.48 (n/a)

mem_copy_4_cores_1_chans_2048_tile_512_False

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`130b6ea` — 2025-12-05 21:15:49	0.09 (-3.98%)	0.07 (+4.00%)	0.07 (+7.04%)	0.06 (+3.67%)	0.01 (-15.76%)	129.70 (-3.57%)	113.30 (-4.50%)	116.90 (-6.55%)	90.40 (+4.15%)	15.04 (-13.80%)
`0a6c11c` — 2025-12-04 00:39:10	0.09 (n/a)	0.07 (n/a)	0.07 (n/a)	0.06 (n/a)	0.01 (n/a)	134.50 (n/a)	118.64 (n/a)	125.10 (n/a)	86.80 (n/a)	17.45 (n/a)

mem_copy_4_cores_1_chans_2048_tile_512_False0

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`0685306` — 2026-02-06 13:45:43	0.06 (-49.00%)	0.05 (-50.55%)	0.05 (-56.31%)	0.04 (-39.17%)	0.01 (-56.65%)	211.00 (+64.46%)	166.84 (+97.30%)	163.80 (+128.77%)	128.70 (+95.89%)	36.90 (+39.70%)
`1531781` — 2026-01-23 18:49:24	0.12 (n/a)	0.10 (n/a)	0.11 (n/a)	0.06 (n/a)	0.03 (n/a)	128.30 (n/a)	84.56 (n/a)	71.60 (n/a)	65.70 (n/a)	26.42 (n/a)

mem_copy_4_cores_2_chans_2048_tile_512_False

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`130b6ea` — 2025-12-05 21:15:49	0.08 (+9.57%)	0.07 (+16.85%)	0.07 (+19.58%)	0.06 (+24.52%)	0.01 (+1.05%)	134.50 (-19.70%)	116.64 (-14.92%)	116.80 (-16.39%)	98.00 (-8.75%)	15.71 (-25.13%)
`0a6c11c` — 2025-12-04 00:39:10	0.08 (n/a)	0.06 (n/a)	0.06 (n/a)	0.05 (n/a)	0.01 (n/a)	167.50 (n/a)	137.10 (n/a)	139.70 (n/a)	107.40 (n/a)	20.99 (n/a)

mem_copy_4_cores_2_chans_2048_tile_512_False0

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`0685306` — 2026-02-06 13:45:43	0.06 (-43.54%)	0.05 (-49.74%)	0.04 (-55.52%)	0.04 (-35.13%)	0.01 (-62.80%)	205.30 (+54.13%)	180.82 (+94.47%)	187.40 (+124.70%)	145.40 (+77.10%)	22.05 (-1.95%)
`1531781` — 2026-01-23 18:49:24	0.10 (n/a)	0.09 (n/a)	0.10 (n/a)	0.06 (n/a)	0.02 (n/a)	133.20 (n/a)	92.98 (n/a)	83.40 (n/a)	82.10 (n/a)	22.49 (n/a)

mem_copy_8_cols_1_channels_2048_tile_256

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`003b49d` — 2025-11-14 19:30:54	0.05 (-13.48%)	0.05 (-17.79%)	0.05 (-15.33%)	0.03 (-25.09%)	0.01 (+6.67%)	267.00 (+33.50%)	176.80 (+24.51%)	150.00 (+18.11%)	141.00 (+15.57%)	47.41 (+61.41%)
`0999a9a` — 2025-11-13 19:24:08	0.06 (n/a)	0.06 (n/a)	0.06 (n/a)	0.04 (n/a)	0.01 (n/a)	200.00 (n/a)	142.00 (n/a)	127.00 (n/a)	122.00 (n/a)	29.37 (n/a)

mem_copy_8_cols_2_channels_2048_tile_128

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`003b49d` — 2025-11-14 19:30:54	0.05 (+13.38%)	0.05 (+23.85%)	0.05 (+37.91%)	0.05 (+30.72%)	0.00 (-56.85%)	166.00 (-23.50%)	153.40 (-20.60%)	153.00 (-27.49%)	142.00 (-11.80%)	7.63 (-70.60%)
`0999a9a` — 2025-11-13 19:24:08	0.05 (n/a)	0.04 (n/a)	0.04 (n/a)	0.04 (n/a)	0.01 (n/a)	217.00 (n/a)	193.20 (n/a)	211.00 (n/a)	161.00 (n/a)	25.96 (n/a)

mem_copy_8_cores_1_chans_2048_tile_256_False

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`130b6ea` — 2025-12-05 21:15:49	0.09 (+16.31%)	0.08 (+15.34%)	0.08 (+14.60%)	0.07 (+19.30%)	0.01 (-5.41%)	117.30 (-16.15%)	107.80 (-13.61%)	108.50 (-12.78%)	94.10 (-13.99%)	8.29 (-32.01%)
`0a6c11c` — 2025-12-04 00:39:10	0.07 (n/a)	0.07 (n/a)	0.07 (n/a)	0.06 (n/a)	0.01 (n/a)	139.90 (n/a)	124.78 (n/a)	124.40 (n/a)	109.40 (n/a)	12.20 (n/a)

mem_copy_8_cores_1_chans_2048_tile_256_False0

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`0685306` — 2026-02-06 13:45:43	0.07 (-22.26%)	0.06 (-33.56%)	0.06 (-38.32%)	0.04 (-35.28%)	0.01 (+3.60%)	202.90 (+54.53%)	154.50 (+54.84%)	143.80 (+62.12%)	111.80 (+28.65%)	39.94 (+111.26%)
`1531781` — 2026-01-23 18:49:24	0.09 (n/a)	0.08 (n/a)	0.09 (n/a)	0.06 (n/a)	0.01 (n/a)	131.30 (n/a)	99.78 (n/a)	88.70 (n/a)	86.90 (n/a)	18.90 (n/a)

mem_copy_8_cores_2_chans_2048_tile_256_False

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`130b6ea` — 2025-12-05 21:15:49	0.09 (+12.61%)	0.07 (-2.24%)	0.07 (-8.85%)	0.06 (+5.13%)	0.01 (+35.50%)	136.50 (-4.88%)	120.86 (+2.96%)	125.10 (+9.64%)	91.60 (-11.15%)	15.72 (+10.46%)
`0a6c11c` — 2025-12-04 00:39:10	0.08 (n/a)	0.07 (n/a)	0.07 (n/a)	0.06 (n/a)	0.01 (n/a)	143.50 (n/a)	117.38 (n/a)	114.10 (n/a)	103.10 (n/a)	14.23 (n/a)

mem_copy_8_cores_2_chans_2048_tile_256_False0

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`0685306` — 2026-02-06 13:45:43	0.06 (-41.12%)	0.05 (-47.25%)	0.05 (-47.93%)	0.04 (-36.64%)	0.01 (-43.82%)	227.90 (+57.83%)	183.82 (+88.15%)	169.30 (+92.17%)	135.90 (+69.88%)	39.01 (+48.07%)
`1531781` — 2026-01-23 18:49:24	0.10 (n/a)	0.09 (n/a)	0.09 (n/a)	0.06 (n/a)	0.02 (n/a)	144.40 (n/a)	97.70 (n/a)	88.10 (n/a)	80.00 (n/a)	26.35 (n/a)

mha

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`130b6ea` — 2025-12-05 21:15:49	0.18 (+0.16%)	0.18 (+0.19%)	0.18 (+0.12%)	0.18 (+0.32%)	0.00 (-52.24%)	47253.90 (-0.32%)	47213.70 (-0.19%)	47210.90 (-0.12%)	47167.70 (-0.16%)	28.71 (-52.47%)
`0a6c11c` — 2025-12-04 00:39:10	0.18 (n/a)	0.18 (n/a)	0.18 (n/a)	0.18 (n/a)	0.00 (n/a)	47405.00 (n/a)	47301.72 (n/a)	47265.90 (n/a)	47241.80 (n/a)	60.40 (n/a)

mha0

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`0685306` — 2026-02-06 13:45:43	0.18 (-0.89%)	0.18 (-0.70%)	0.18 (-0.68%)	0.18 (-0.64%)	0.00 (-31.67%)	47674.20 (+0.65%)	47565.12 (+0.70%)	47557.40 (+0.69%)	47493.90 (+0.89%)	73.21 (-30.58%)
`1531781` — 2026-01-23 18:49:24	0.18 (n/a)	0.18 (n/a)	0.18 (n/a)	0.18 (n/a)	0.00 (n/a)	47367.50 (n/a)	47232.68 (n/a)	47232.00 (n/a)	47073.00 (n/a)	105.45 (n/a)

relu_1_cols_1_channels_2048_tile_2048

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`0685306` — 2026-02-06 13:45:43	0.07 (-35.65%)	0.05 (-44.05%)	0.04 (-54.96%)	0.04 (-27.03%)	0.01 (-49.83%)	209.60 (+37.08%)	174.54 (+71.76%)	182.40 (+121.90%)	113.50 (+55.48%)	38.49 (+7.68%)
`1531781` — 2026-01-23 18:49:24	0.11 (n/a)	0.09 (n/a)	0.10 (n/a)	0.05 (n/a)	0.03 (n/a)	152.90 (n/a)	101.62 (n/a)	82.20 (n/a)	73.00 (n/a)	35.75 (n/a)

relu_2_cols_1_channels_2048_tile_1024

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`0685306` — 2026-02-06 13:45:43	0.07 (-32.63%)	0.05 (-38.28%)	0.05 (-40.60%)	0.04 (-25.45%)	0.01 (-43.08%)	201.90 (+34.06%)	161.22 (+59.21%)	153.80 (+68.46%)	123.30 (+48.55%)	30.89 (+10.26%)
`1531781` — 2026-01-23 18:49:24	0.10 (n/a)	0.08 (n/a)	0.09 (n/a)	0.05 (n/a)	0.02 (n/a)	150.60 (n/a)	101.26 (n/a)	91.30 (n/a)	83.00 (n/a)	28.02 (n/a)

relu_4_cols_1_channels_2048_tile_512

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`0685306` — 2026-02-06 13:45:43	0.06 (-36.41%)	0.05 (-41.93%)	0.05 (-47.94%)	0.04 (-32.09%)	0.01 (-49.30%)	200.00 (+47.28%)	163.32 (+69.77%)	165.70 (+92.00%)	130.70 (+57.28%)	25.90 (+15.29%)
`1531781` — 2026-01-23 18:49:24	0.10 (n/a)	0.09 (n/a)	0.09 (n/a)	0.06 (n/a)	0.02 (n/a)	135.80 (n/a)	96.20 (n/a)	86.30 (n/a)	83.10 (n/a)	22.46 (n/a)

relu_8_cols_1_channels_2048_tile_256

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`0685306` — 2026-02-06 13:45:43	0.07 (-34.05%)	0.05 (-45.44%)	0.05 (-54.60%)	0.04 (-21.50%)	0.01 (-50.11%)	196.70 (+27.40%)	164.26 (+76.02%)	172.60 (+120.15%)	114.90 (+51.58%)	31.54 (-7.89%)
`1531781` — 2026-01-23 18:49:24	0.11 (n/a)	0.09 (n/a)	0.10 (n/a)	0.05 (n/a)	0.02 (n/a)	154.40 (n/a)	93.32 (n/a)	78.40 (n/a)	75.80 (n/a)	34.24 (n/a)

rms_norm_1_cols_1_channels_2048_tile_2048

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`0685306` — 2026-02-06 13:45:43	0.06 (-29.95%)	0.05 (-44.81%)	0.04 (-55.00%)	0.03 (-53.92%)	0.02 (+30.51%)	285.40 (+117.03%)	195.98 (+94.62%)	205.30 (+122.19%)	129.70 (+42.68%)	65.24 (+276.13%)
`1531781` — 2026-01-23 18:49:24	0.09 (n/a)	0.08 (n/a)	0.09 (n/a)	0.06 (n/a)	0.01 (n/a)	131.50 (n/a)	100.70 (n/a)	92.40 (n/a)	90.90 (n/a)	17.35 (n/a)

rms_norm_1_cols_2_channels_2048_tile_1024

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`0685306` — 2026-02-06 13:45:43	0.07 (-39.62%)	0.05 (-46.75%)	0.05 (-51.72%)	0.04 (-23.78%)	0.01 (-61.09%)	190.50 (+31.20%)	154.02 (+77.98%)	149.30 (+107.07%)	116.80 (+65.67%)	26.99 (-17.73%)
`1531781` — 2026-01-23 18:49:24	0.12 (n/a)	0.10 (n/a)	0.11 (n/a)	0.06 (n/a)	0.03 (n/a)	145.20 (n/a)	86.54 (n/a)	72.10 (n/a)	70.50 (n/a)	32.80 (n/a)

rms_norm_2_cols_1_channels_2048_tile_1024

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`0685306` — 2026-02-06 13:45:43	0.07 (-20.20%)	0.06 (-31.55%)	0.06 (-36.05%)	0.05 (-20.87%)	0.01 (-24.47%)	163.80 (+26.39%)	143.18 (+45.77%)	142.10 (+56.33%)	112.50 (+25.28%)	21.13 (+20.40%)
`1531781` — 2026-01-23 18:49:24	0.09 (n/a)	0.09 (n/a)	0.09 (n/a)	0.06 (n/a)	0.01 (n/a)	129.60 (n/a)	98.22 (n/a)	90.90 (n/a)	89.80 (n/a)	17.55 (n/a)

rms_norm_2_cols_2_channels_2048_tile_512

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`0685306` — 2026-02-06 13:45:43	0.08 (-39.98%)	0.06 (-47.59%)	0.06 (-45.14%)	0.04 (-55.50%)	0.02 (-2.50%)	210.00 (+124.60%)	147.98 (+99.11%)	133.30 (+82.35%)	104.00 (+66.67%)	41.82 (+260.45%)
`1531781` — 2026-01-23 18:49:24	0.13 (n/a)	0.11 (n/a)	0.11 (n/a)	0.09 (n/a)	0.02 (n/a)	93.50 (n/a)	74.32 (n/a)	73.10 (n/a)	62.40 (n/a)	11.60 (n/a)

rms_norm_4_cols_1_channels_2048_tile_512

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`0685306` — 2026-02-06 13:45:43	0.06 (-33.56%)	0.05 (-39.27%)	0.05 (-41.33%)	0.04 (-34.71%)	0.01 (-23.53%)	200.30 (+53.25%)	161.72 (+65.90%)	153.80 (+70.51%)	129.50 (+50.58%)	32.16 (+72.30%)
`1531781` — 2026-01-23 18:49:24	0.10 (n/a)	0.09 (n/a)	0.09 (n/a)	0.06 (n/a)	0.01 (n/a)	130.70 (n/a)	97.48 (n/a)	90.20 (n/a)	86.00 (n/a)	18.66 (n/a)

rms_norm_4_cols_2_channels_2048_tile_256

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`0685306` — 2026-02-06 13:45:43	0.06 (-46.05%)	0.05 (-47.40%)	0.05 (-54.14%)	0.05 (-28.07%)	0.01 (-70.68%)	174.50 (+39.04%)	153.88 (+83.41%)	160.20 (+117.96%)	135.20 (+85.46%)	17.09 (-26.53%)
`1531781` — 2026-01-23 18:49:24	0.11 (n/a)	0.10 (n/a)	0.11 (n/a)	0.07 (n/a)	0.02 (n/a)	125.50 (n/a)	83.90 (n/a)	73.50 (n/a)	72.90 (n/a)	23.26 (n/a)

rms_norm_8_cols_1_channels_2048_tile_256

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`0685306` — 2026-02-06 13:45:43	0.06 (-28.41%)	0.05 (-33.53%)	0.05 (-39.37%)	0.05 (-13.93%)	0.01 (-59.49%)	179.30 (+16.20%)	156.02 (+46.28%)	156.40 (+64.98%)	131.40 (+39.79%)	17.23 (-35.32%)
`1531781` — 2026-01-23 18:49:24	0.09 (n/a)	0.08 (n/a)	0.09 (n/a)	0.05 (n/a)	0.01 (n/a)	154.30 (n/a)	106.66 (n/a)	94.80 (n/a)	94.00 (n/a)	26.64 (n/a)

rms_norm_8_cols_2_channels_2048_tile_128

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`0685306` — 2026-02-06 13:45:43	0.05 (-42.97%)	0.05 (-42.02%)	0.05 (-44.97%)	0.04 (-20.53%)	0.01 (-64.26%)	211.00 (+25.82%)	181.08 (+65.58%)	166.40 (+81.66%)	155.90 (+75.37%)	26.81 (-20.02%)
`1531781` — 2026-01-23 18:49:24	0.09 (n/a)	0.08 (n/a)	0.09 (n/a)	0.05 (n/a)	0.02 (n/a)	167.70 (n/a)	109.36 (n/a)	91.60 (n/a)	88.90 (n/a)	33.52 (n/a)

rope_1_cols_2_channels_4096_tile_4096_0

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`1b2d230` — 2026-01-13 22:38:03	0.19 (-21.66%)	0.17 (-8.53%)	0.18 (+2.09%)	0.14 (-6.71%)	0.02 (-38.07%)	174.10 (+7.20%)	150.06 (+8.11%)	139.60 (-2.04%)	129.80 (+27.63%)	20.36 (-10.65%)
`d48746f` — 2026-01-12 22:17:10	0.24 (n/a)	0.18 (n/a)	0.17 (n/a)	0.15 (n/a)	0.04 (n/a)	162.40 (n/a)	138.80 (n/a)	142.50 (n/a)	101.70 (n/a)	22.79 (n/a)

rope_1c_32rows_512cols_32arows_0m

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`0685306` — 2026-02-06 13:45:43	0.73 (-33.85%)	0.55 (-43.96%)	0.57 (-45.70%)	0.39 (-47.16%)	0.16 (+7.33%)	253.20 (+89.24%)	190.64 (+87.75%)	171.90 (+84.24%)	134.70 (+51.18%)	58.18 (+212.84%)
`1531781` — 2026-01-23 18:49:24	1.10 (n/a)	0.99 (n/a)	1.05 (n/a)	0.73 (n/a)	0.15 (n/a)	133.80 (n/a)	101.54 (n/a)	93.30 (n/a)	89.10 (n/a)	18.60 (n/a)

rope_1c_32rows_512cols_8arows_0m

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`0685306` — 2026-02-06 13:45:43	0.62 (+12.03%)	0.50 (-4.90%)	0.58 (+10.91%)	0.30 (-37.89%)	0.15 (+380.56%)	247.50 (+61.03%)	160.80 (+14.17%)	126.40 (-9.78%)	118.30 (-10.72%)	56.54 (+573.07%)
`1531781` — 2026-01-23 18:49:24	0.56 (n/a)	0.52 (n/a)	0.53 (n/a)	0.48 (n/a)	0.03 (n/a)	153.70 (n/a)	140.84 (n/a)	140.10 (n/a)	132.50 (n/a)	8.40 (n/a)

rope_2_cols_2_channels_4096_tile_2048_0

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`1b2d230` — 2026-01-13 22:38:03	0.21 (-0.56%)	0.17 (-10.62%)	0.16 (-13.58%)	0.13 (-13.33%)	0.03 (+17.05%)	183.70 (+15.39%)	151.76 (+12.77%)	155.20 (+15.73%)	117.30 (+0.60%)	24.24 (+35.73%)
`d48746f` — 2026-01-12 22:17:10	0.21 (n/a)	0.19 (n/a)	0.18 (n/a)	0.15 (n/a)	0.02 (n/a)	159.20 (n/a)	134.58 (n/a)	134.10 (n/a)	116.60 (n/a)	17.86 (n/a)

rope_2c_32rows_512cols_32arows_0m

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`0685306` — 2026-02-06 13:45:43	0.89 (-29.00%)	0.75 (-22.79%)	0.82 (-9.68%)	0.57 (-8.95%)	0.14 (-43.50%)	172.40 (+9.81%)	136.04 (+25.85%)	119.90 (+10.71%)	110.90 (+40.91%)	27.69 (-10.94%)
`1531781` — 2026-01-23 18:49:24	1.25 (n/a)	0.97 (n/a)	0.91 (n/a)	0.63 (n/a)	0.25 (n/a)	157.00 (n/a)	108.10 (n/a)	108.30 (n/a)	78.70 (n/a)	31.09 (n/a)

rope_2c_32rows_512cols_8arows_0m

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`0685306` — 2026-02-06 13:45:43	0.61 (-3.01%)	0.52 (-9.36%)	0.52 (-7.64%)	0.38 (-22.41%)	0.09 (+65.62%)	194.10 (+28.88%)	146.78 (+12.41%)	141.10 (+8.21%)	120.40 (+3.08%)	28.21 (+124.51%)
`1531781` — 2026-01-23 18:49:24	0.63 (n/a)	0.57 (n/a)	0.57 (n/a)	0.49 (n/a)	0.05 (n/a)	150.60 (n/a)	130.58 (n/a)	130.40 (n/a)	116.80 (n/a)	12.57 (n/a)

rope_4_cols_2_channels_4096_tile_1024_0

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`1b2d230` — 2026-01-13 22:38:03	0.22 (+18.92%)	0.17 (+6.24%)	0.17 (+5.57%)	0.14 (-0.50%)	0.03 (+90.77%)	171.10 (+0.47%)	147.40 (-4.39%)	148.10 (-5.31%)	110.10 (-15.89%)	23.59 (+60.67%)
`d48746f` — 2026-01-12 22:17:10	0.19 (n/a)	0.16 (n/a)	0.16 (n/a)	0.14 (n/a)	0.02 (n/a)	170.30 (n/a)	154.16 (n/a)	156.40 (n/a)	130.90 (n/a)	14.69 (n/a)

rope_8_cols_2_channels_4096_tile_512_0

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`1b2d230` — 2026-01-13 22:38:03	0.17 (-1.30%)	0.15 (+6.63%)	0.15 (-8.78%)	0.14 (+39.68%)	0.01 (-67.35%)	174.90 (-28.41%)	163.42 (-10.36%)	167.40 (+9.63%)	146.50 (+1.31%)	11.19 (-76.17%)
`d48746f` — 2026-01-12 22:17:10	0.17 (n/a)	0.14 (n/a)	0.16 (n/a)	0.10 (n/a)	0.03 (n/a)	244.30 (n/a)	182.30 (n/a)	152.70 (n/a)	144.60 (n/a)	46.98 (n/a)

rope_8c_32rows_512cols_32arows_0m

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`0685306` — 2026-02-06 13:45:43	0.88 (-33.79%)	0.63 (-29.25%)	0.56 (-34.54%)	0.49 (-18.36%)	0.17 (-38.88%)	200.30 (+22.51%)	163.12 (+38.97%)	176.40 (+52.73%)	111.10 (+50.95%)	37.51 (+16.94%)
`1531781` — 2026-01-23 18:49:24	1.34 (n/a)	0.89 (n/a)	0.85 (n/a)	0.60 (n/a)	0.27 (n/a)	163.50 (n/a)	117.38 (n/a)	115.50 (n/a)	73.60 (n/a)	32.08 (n/a)

rope_8c_32rows_512cols_8arows_0m

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`0685306` — 2026-02-06 13:45:43	0.61 (+11.43%)	0.44 (-14.05%)	0.42 (-19.93%)	0.34 (-30.53%)	0.10 (+292.10%)	220.10 (+43.95%)	173.20 (+20.85%)	174.50 (+24.91%)	121.60 (-10.26%)	37.02 (+395.94%)
`1531781` — 2026-01-23 18:49:24	0.54 (n/a)	0.52 (n/a)	0.53 (n/a)	0.48 (n/a)	0.03 (n/a)	152.90 (n/a)	143.32 (n/a)	139.70 (n/a)	135.50 (n/a)	7.46 (n/a)

sigmoid_1_cols_1_channels_2048_tile_2048

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`0685306` — 2026-02-06 13:45:43	0.07 (-39.79%)	0.06 (-39.44%)	0.06 (-49.64%)	0.05 (-10.99%)	0.01 (-66.92%)	172.10 (+12.34%)	138.88 (+52.75%)	136.10 (+98.40%)	113.50 (+66.18%)	23.41 (-36.39%)
`1531781` — 2026-01-23 18:49:24	0.12 (n/a)	0.10 (n/a)	0.12 (n/a)	0.05 (n/a)	0.03 (n/a)	153.20 (n/a)	90.92 (n/a)	68.60 (n/a)	68.30 (n/a)	36.80 (n/a)

sigmoid_2_cols_1_channels_2048_tile_1024

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`0685306` — 2026-02-06 13:45:43	0.07 (-41.92%)	0.06 (-39.24%)	0.06 (-47.33%)	0.04 (-32.25%)	0.01 (-57.48%)	207.90 (+47.66%)	141.92 (+57.97%)	130.90 (+89.99%)	117.30 (+72.25%)	37.45 (+16.87%)
`1531781` — 2026-01-23 18:49:24	0.12 (n/a)	0.10 (n/a)	0.12 (n/a)	0.06 (n/a)	0.03 (n/a)	140.80 (n/a)	89.84 (n/a)	68.90 (n/a)	68.10 (n/a)	32.04 (n/a)

sigmoid_4_cols_1_channels_2048_tile_512

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`0685306` — 2026-02-06 13:45:43	0.07 (-25.98%)	0.06 (-27.96%)	0.06 (-39.58%)	0.05 (-0.05%)	0.01 (-44.25%)	169.50 (+0.06%)	142.10 (+33.33%)	144.70 (+65.56%)	113.50 (+35.12%)	27.15 (-25.33%)
`1531781` — 2026-01-23 18:49:24	0.10 (n/a)	0.08 (n/a)	0.09 (n/a)	0.05 (n/a)	0.02 (n/a)	169.40 (n/a)	106.58 (n/a)	87.40 (n/a)	84.00 (n/a)	36.37 (n/a)

sigmoid_8_cols_1_channels_2048_tile_256

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`0685306` — 2026-02-06 13:45:43	0.08 (-32.37%)	0.05 (-47.50%)	0.05 (-60.03%)	0.04 (-25.11%)	0.02 (-42.03%)	210.00 (+33.50%)	164.20 (+82.97%)	171.90 (+150.22%)	97.50 (+47.95%)	41.19 (+6.02%)
`1531781` — 2026-01-23 18:49:24	0.12 (n/a)	0.10 (n/a)	0.12 (n/a)	0.05 (n/a)	0.03 (n/a)	157.30 (n/a)	89.74 (n/a)	68.70 (n/a)	65.90 (n/a)	38.85 (n/a)

silu_1_cols_1_channels_2048_tile_2048

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`0685306` — 2026-02-06 13:45:43	0.06 (-44.86%)	0.05 (-43.57%)	0.05 (-49.64%)	0.04 (-38.72%)	0.01 (-62.46%)	233.90 (+63.22%)	178.58 (+70.27%)	163.30 (+98.66%)	146.20 (+81.39%)	36.23 (+12.38%)
`1531781` — 2026-01-23 18:49:24	0.10 (n/a)	0.08 (n/a)	0.10 (n/a)	0.06 (n/a)	0.02 (n/a)	143.30 (n/a)	104.88 (n/a)	82.20 (n/a)	80.60 (n/a)	32.24 (n/a)

silu_2_cols_1_channels_2048_tile_1024

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`0685306` — 2026-02-06 13:45:43	0.05 (-51.06%)	0.04 (-49.12%)	0.04 (-57.89%)	0.04 (-34.97%)	0.00 (-79.73%)	213.60 (+53.78%)	193.08 (+87.27%)	197.60 (+137.50%)	166.70 (+104.29%)	18.20 (-35.71%)
`1531781` — 2026-01-23 18:49:24	0.10 (n/a)	0.08 (n/a)	0.10 (n/a)	0.06 (n/a)	0.02 (n/a)	138.90 (n/a)	103.10 (n/a)	83.20 (n/a)	81.60 (n/a)	28.30 (n/a)

silu_4_cols_1_channels_2048_tile_512

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`0685306` — 2026-02-06 13:45:43	0.05 (-60.82%)	0.04 (-58.72%)	0.04 (-62.09%)	0.04 (-35.84%)	0.00 (-85.77%)	214.10 (+55.82%)	188.62 (+126.27%)	184.50 (+163.57%)	173.80 (+155.21%)	16.74 (-44.69%)
`1531781` — 2026-01-23 18:49:24	0.12 (n/a)	0.11 (n/a)	0.12 (n/a)	0.06 (n/a)	0.03 (n/a)	137.40 (n/a)	83.36 (n/a)	70.00 (n/a)	68.10 (n/a)	30.27 (n/a)

silu_8_cols_1_channels_2048_tile_256

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`0685306` — 2026-02-06 13:45:43	0.05 (-59.92%)	0.04 (-56.60%)	0.04 (-54.63%)	0.03 (-55.21%)	0.01 (-65.11%)	245.10 (+123.22%)	204.18 (+128.90%)	193.80 (+120.48%)	177.40 (+149.51%)	28.15 (+94.06%)
`1531781` — 2026-01-23 18:49:24	0.12 (n/a)	0.09 (n/a)	0.09 (n/a)	0.07 (n/a)	0.02 (n/a)	109.80 (n/a)	89.20 (n/a)	87.90 (n/a)	71.10 (n/a)	14.50 (n/a)

softmax_1_cols_2_channels_4096_tile_2048

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`1531781` — 2026-01-23 18:49:24	0.21 (+20.19%)	0.16 (+13.84%)	0.16 (+8.19%)	0.12 (+26.61%)	0.04 (+17.18%)	135.50 (-20.99%)	104.68 (-12.65%)	102.70 (-7.56%)	76.50 (-16.85%)	22.55 (-26.21%)
`0f0da0b` — 2026-01-14 16:31:00	0.18 (n/a)	0.14 (n/a)	0.15 (n/a)	0.10 (n/a)	0.03 (n/a)	171.50 (n/a)	119.84 (n/a)	111.10 (n/a)	92.00 (n/a)	30.56 (n/a)

softmax_2_cols_2_channels_32768_tile_1024

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`0685306` — 2026-02-06 13:45:43	0.80 (n/a)	0.76 (n/a)	0.77 (n/a)	0.68 (n/a)	0.05 (n/a)	191.50 (n/a)	172.88 (n/a)	169.30 (n/a)	163.30 (n/a)	11.75 (n/a)

softmax_2_cols_2_channels_32768_tile_512

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`0685306` — 2026-02-06 13:45:43	0.90 (n/a)	0.72 (n/a)	0.73 (n/a)	0.54 (n/a)	0.15 (n/a)	242.70 (n/a)	188.58 (n/a)	178.60 (n/a)	145.90 (n/a)	40.18 (n/a)

softmax_2_cols_2_channels_4096_tile_1024

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`1531781` — 2026-01-23 18:49:24	0.22 (+40.26%)	0.18 (+43.33%)	0.20 (+64.08%)	0.13 (+24.34%)	0.04 (+68.98%)	127.80 (-19.62%)	94.74 (-28.95%)	83.40 (-39.08%)	73.90 (-28.67%)	24.46 (-5.31%)
`0f0da0b` — 2026-01-14 16:31:00	0.16 (n/a)	0.13 (n/a)	0.12 (n/a)	0.10 (n/a)	0.03 (n/a)	159.00 (n/a)	133.34 (n/a)	136.90 (n/a)	103.60 (n/a)	25.83 (n/a)

softmax_2_cols_2_channels_4096_tile_512

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`1531781` — 2026-01-23 18:49:24	0.24 (+85.21%)	0.17 (+69.97%)	0.15 (+54.16%)	0.11 (+37.21%)	0.05 (+203.82%)	148.10 (-27.12%)	102.78 (-37.88%)	106.70 (-35.10%)	69.60 (-45.96%)	31.81 (+15.98%)
`0f0da0b` — 2026-01-14 16:31:00	0.13 (n/a)	0.10 (n/a)	0.10 (n/a)	0.08 (n/a)	0.02 (n/a)	203.20 (n/a)	165.46 (n/a)	164.40 (n/a)	128.80 (n/a)	27.43 (n/a)

softmax_4_cols_4_channels_32768_tile_2048

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`0685306` — 2026-02-06 13:45:43	0.77 (n/a)	0.62 (n/a)	0.64 (n/a)	0.47 (n/a)	0.11 (n/a)	278.70 (n/a)	216.70 (n/a)	206.20 (n/a)	170.90 (n/a)	40.32 (n/a)

swiglu

No metrics available.

swiglu_decode_1x2048x2048

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`130b6ea` — 2025-12-05 21:15:49	0.00 (-15.46%)	0.00 (-27.20%)	0.00 (-28.17%)	0.00 (-33.90%)	0.00 (+1834.83%)	5187.20 (+51.28%)	4720.84 (+38.38%)	4750.90 (+39.22%)	4015.10 (+18.29%)	390.23 (+3298.45%)
`0a6c11c` — 2025-12-04 00:39:10	0.00 (n/a)	0.00 (n/a)	0.00 (n/a)	0.00 (n/a)	0.00 (n/a)	3428.80 (n/a)	3411.62 (n/a)	3412.60 (n/a)	3394.20 (n/a)	11.48 (n/a)

swiglu_decode_1x2048x2048_0

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`0685306` — 2026-02-06 13:45:43	n/a (n/a)	n/a (n/a)	n/a (n/a)	n/a (n/a)	n/a (n/a)	n/a (n/a)	n/a (n/a)	n/a (n/a)	n/a (n/a)	n/a (n/a)
`1531781` — 2026-01-23 18:49:24	0.00 (n/a)	0.00 (n/a)	0.00 (n/a)	0.00 (n/a)	0.00 (n/a)	5111.40 (n/a)	4324.88 (n/a)	4736.10 (n/a)	3410.70 (n/a)	843.97 (n/a)

swiglu_prefill_256x2048x2048_0

No metrics available.

tanh_1_cols_1_channels_2048_tile_2048

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`0685306` — 2026-02-06 13:45:43	0.08 (-17.48%)	0.06 (-20.98%)	0.07 (-7.21%)	0.05 (-21.03%)	0.01 (-19.33%)	178.30 (+26.63%)	138.52 (+26.92%)	124.90 (+7.77%)	103.70 (+21.14%)	31.38 (+33.42%)
`1531781` — 2026-01-23 18:49:24	0.10 (n/a)	0.08 (n/a)	0.07 (n/a)	0.06 (n/a)	0.02 (n/a)	140.80 (n/a)	109.14 (n/a)	115.90 (n/a)	85.60 (n/a)	23.52 (n/a)

tanh_2_cols_1_channels_2048_tile_1024

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`0685306` — 2026-02-06 13:45:43	0.07 (-34.88%)	0.05 (-45.84%)	0.05 (-49.07%)	0.04 (-47.07%)	0.01 (+5.45%)	195.30 (+89.06%)	165.56 (+88.09%)	172.40 (+96.36%)	119.40 (+53.47%)	27.98 (+190.35%)
`1531781` — 2026-01-23 18:49:24	0.11 (n/a)	0.09 (n/a)	0.09 (n/a)	0.08 (n/a)	0.01 (n/a)	103.30 (n/a)	88.02 (n/a)	87.80 (n/a)	77.80 (n/a)	9.64 (n/a)

tanh_4_cols_1_channels_2048_tile_512

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`0685306` — 2026-02-06 13:45:43	0.07 (-27.11%)	0.06 (-36.11%)	0.06 (-34.30%)	0.03 (-61.70%)	0.02 (+43.01%)	316.90 (+161.04%)	164.62 (+76.25%)	131.40 (+52.26%)	116.10 (+37.23%)	85.38 (+440.21%)
`1531781` — 2026-01-23 18:49:24	0.10 (n/a)	0.09 (n/a)	0.09 (n/a)	0.07 (n/a)	0.01 (n/a)	121.40 (n/a)	93.40 (n/a)	86.30 (n/a)	84.60 (n/a)	15.80 (n/a)

tanh_8_cols_1_channels_2048_tile_256

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`0685306` — 2026-02-06 13:45:43	0.05 (-58.14%)	0.04 (-51.64%)	0.04 (-59.96%)	0.04 (-20.03%)	0.00 (-88.43%)	204.80 (+25.03%)	192.60 (+87.17%)	200.70 (+149.63%)	171.20 (+138.77%)	14.11 (-64.71%)
`1531781` — 2026-01-23 18:49:24	0.11 (n/a)	0.09 (n/a)	0.10 (n/a)	0.05 (n/a)	0.03 (n/a)	163.80 (n/a)	102.90 (n/a)	80.40 (n/a)	71.70 (n/a)	39.99 (n/a)

transpose_2048_M_64_N_1_cols_1_channels_64_m_64_n_8_s

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`130b6ea` — 2025-12-05 21:15:49	3.93 (-7.02%)	3.68 (+2.48%)	3.66 (+1.82%)	3.45 (+15.62%)	0.15 (-63.64%)	151.80 (-13.50%)	142.70 (-3.59%)	143.40 (-1.78%)	133.50 (+7.49%)	5.90 (-66.22%)
`0a6c11c` — 2025-12-04 00:39:10	4.22 (n/a)	3.59 (n/a)	3.59 (n/a)	2.99 (n/a)	0.42 (n/a)	175.50 (n/a)	148.02 (n/a)	146.00 (n/a)	124.20 (n/a)	17.46 (n/a)

transpose_2048_M_64_N_1_cols_1_channels_64_m_64_n_8_s0

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`0685306` — 2026-02-06 13:45:43	3.56 (-28.07%)	2.97 (-31.65%)	3.32 (-24.05%)	2.29 (-38.75%)	0.62 (+24.87%)	229.10 (+63.29%)	183.46 (+50.40%)	157.80 (+31.61%)	147.30 (+38.96%)	41.30 (+191.78%)
`1531781` — 2026-01-23 18:49:24	4.95 (n/a)	4.34 (n/a)	4.37 (n/a)	3.74 (n/a)	0.50 (n/a)	140.30 (n/a)	121.98 (n/a)	119.90 (n/a)	106.00 (n/a)	14.15 (n/a)

transpose_2048_M_64_N_1_cols_2_channels_64_m_64_n_8_s

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`130b6ea` — 2025-12-05 21:15:49	5.11 (+15.57%)	3.85 (-3.32%)	3.40 (-18.64%)	3.29 (+7.09%)	0.71 (+49.14%)	159.60 (-6.61%)	140.18 (+4.80%)	154.20 (+22.97%)	102.60 (-13.49%)	22.70 (+19.77%)
`0a6c11c` — 2025-12-04 00:39:10	4.42 (n/a)	3.99 (n/a)	4.18 (n/a)	3.07 (n/a)	0.48 (n/a)	170.90 (n/a)	133.76 (n/a)	125.40 (n/a)	118.60 (n/a)	18.95 (n/a)

transpose_2048_M_64_N_1_cols_2_channels_64_m_64_n_8_s0

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`0685306` — 2026-02-06 13:45:43	3.91 (-35.70%)	3.35 (-32.89%)	3.23 (-36.44%)	2.78 (-32.01%)	0.44 (-40.13%)	188.40 (+47.07%)	158.74 (+48.55%)	162.10 (+57.38%)	134.20 (+55.50%)	21.29 (+35.84%)
`1531781` — 2026-01-23 18:49:24	6.07 (n/a)	4.99 (n/a)	5.09 (n/a)	4.09 (n/a)	0.74 (n/a)	128.10 (n/a)	106.86 (n/a)	103.00 (n/a)	86.30 (n/a)	15.67 (n/a)

weighted_rms_norm_1_cols_2_channels_2048_weights_2048

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`0685306` — 2026-02-06 13:45:43	n/a (n/a)	n/a (n/a)	n/a (n/a)	n/a (n/a)	n/a (n/a)	n/a (n/a)	n/a (n/a)	n/a (n/a)	n/a (n/a)	n/a (n/a)
`1531781` — 2026-01-23 18:49:24	0.11 (n/a)	0.09 (n/a)	0.09 (n/a)	0.05 (n/a)	0.02 (n/a)	155.00 (n/a)	99.92 (n/a)	89.60 (n/a)	76.50 (n/a)	31.36 (n/a)

weighted_rms_norm_2_cols_2_channels_2048_weights_1024

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`0685306` — 2026-02-06 13:45:43	n/a (n/a)	n/a (n/a)	n/a (n/a)	n/a (n/a)	n/a (n/a)	n/a (n/a)	n/a (n/a)	n/a (n/a)	n/a (n/a)	n/a (n/a)
`1531781` — 2026-01-23 18:49:24	0.10 (n/a)	0.09 (n/a)	0.09 (n/a)	0.05 (n/a)	0.02 (n/a)	166.70 (n/a)	102.72 (n/a)	88.50 (n/a)	83.10 (n/a)	35.87 (n/a)

weighted_rms_norm_4_cols_2_channels_2048_weights_512

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`0685306` — 2026-02-06 13:45:43	n/a (n/a)	n/a (n/a)	n/a (n/a)	n/a (n/a)	n/a (n/a)	n/a (n/a)	n/a (n/a)	n/a (n/a)	n/a (n/a)	n/a (n/a)
`1531781` — 2026-01-23 18:49:24	0.09 (n/a)	0.09 (n/a)	0.09 (n/a)	0.07 (n/a)	0.01 (n/a)	119.00 (n/a)	96.48 (n/a)	89.90 (n/a)	89.30 (n/a)	12.81 (n/a)

weighted_rms_norm_8_cols_2_channels_2048_weights_256

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`0685306` — 2026-02-06 13:45:43	n/a (n/a)	n/a (n/a)	n/a (n/a)	n/a (n/a)	n/a (n/a)	n/a (n/a)	n/a (n/a)	n/a (n/a)	n/a (n/a)	n/a (n/a)
`1531781` — 2026-01-23 18:49:24	0.11 (n/a)	0.10 (n/a)	0.10 (n/a)	0.06 (n/a)	0.02 (n/a)	135.80 (n/a)	89.92 (n/a)	80.00 (n/a)	74.30 (n/a)	25.76 (n/a)

github-actions · 2026-02-06T13:58:14Z

📊 Test Results for Test Example Applications

0685306 (2026_02_06_13_57_42)

IRONCLAD

Tested on 2026_02_06_13_57_42 at commit 0685306.

Test	Checks	TTFT (mean)	TPS (mean)

📈 Trends (vs main branch) for Test Example Applications

0685306 (2026_02_06_13_57_42)

IRONCLAD Trends

llama_3.2_1b

Commit/Date	Num Tokens (max)	Num Tokens (mean)	Num Tokens (median)	Num Tokens (min)	Num Tokens (stddev)	TPS (max)	TPS (mean)	TPS (median)	TPS (min)	TPS (stddev)	TTFT (max)	TTFT (mean)	TTFT (median)	TTFT (min)	TTFT (stddev)	Total (max)	Total (mean)	Total (median)	Total (min)	Total (stddev)
`130b6ea` — 2025-12-05 21:33:12	40.00 (+0.00%)	40.00 (+0.00%)	40.00 (+0.00%)	40.00 (+0.00%)	0.00 (n/a)	4.71 (-0.42%)	4.64 (-0.09%)	4.64 (+0.65%)	4.55 (-0.22%)	0.05 (-17.66%)	4.41 (-0.34%)	4.39 (-0.19%)	4.38 (-0.33%)	4.37 (-0.15%)	0.01 (-25.90%)	12.96 (-0.00%)	12.80 (+0.07%)	12.80 (-0.23%)	12.67 (+0.44%)	0.09 (-21.12%)
`0a6c11c` — 2025-12-03 23:35:15	40.00 (n/a)	40.00 (n/a)	40.00 (n/a)	40.00 (n/a)	0.00 (n/a)	4.73 (n/a)	4.64 (n/a)	4.61 (n/a)	4.56 (n/a)	0.06 (n/a)	4.42 (n/a)	4.40 (n/a)	4.40 (n/a)	4.37 (n/a)	0.02 (n/a)	12.96 (n/a)	12.79 (n/a)	12.83 (n/a)	12.62 (n/a)	0.12 (n/a)

llama_3.2_1b_prompt_13_tokens_1

Commit/Date	Num_Tokens (max)	Num_Tokens (mean)	Num_Tokens (median)	Num_Tokens (min)	Num_Tokens (stddev)	TPS (max)	TPS (mean)	TPS (median)	TPS (min)	TPS (stddev)	TTFT (max)	TTFT (mean)	TTFT (median)	TTFT (min)	TTFT (stddev)
`1531781` — 2026-01-23 18:40:34	1.00 (+0.00%)	1.00 (+0.00%)	1.00 (+0.00%)	1.00 (+0.00%)	0.00 (n/a)	0.00 (n/a)	0.00 (n/a)	0.00 (n/a)	0.00 (n/a)	0.00 (n/a)	0.61 (-2.51%)	0.61 (-2.51%)	0.61 (-2.51%)	0.61 (-2.51%)	0.00 (n/a)
`0f0da0b` — 2026-01-14 16:41:39	1.00 (n/a)	1.00 (n/a)	1.00 (n/a)	1.00 (n/a)	0.00 (n/a)	0.00 (n/a)	0.00 (n/a)	0.00 (n/a)	0.00 (n/a)	0.00 (n/a)	0.63 (n/a)	0.63 (n/a)	0.63 (n/a)	0.63 (n/a)	0.00 (n/a)

llama_3.2_1b_prompt_13_tokens_40

Commit/Date	Num_Tokens (max)	Num_Tokens (mean)	Num_Tokens (median)	Num_Tokens (min)	Num_Tokens (stddev)	TPS (max)	TPS (mean)	TPS (median)	TPS (min)	TPS (stddev)	TTFT (max)	TTFT (mean)	TTFT (median)	TTFT (min)	TTFT (stddev)
`1531781` — 2026-01-23 18:40:34	40.00 (+0.00%)	40.00 (+0.00%)	40.00 (+0.00%)	40.00 (+0.00%)	0.00 (n/a)	4.34 (-1.81%)	4.34 (-1.81%)	4.34 (-1.81%)	4.34 (-1.81%)	0.00 (n/a)	0.61 (-0.68%)	0.61 (-0.68%)	0.61 (-0.68%)	0.61 (-0.68%)	0.00 (n/a)
`0f0da0b` — 2026-01-14 16:41:39	40.00 (n/a)	40.00 (n/a)	40.00 (n/a)	40.00 (n/a)	0.00 (n/a)	4.42 (n/a)	4.42 (n/a)	4.42 (n/a)	4.42 (n/a)	0.00 (n/a)	0.62 (n/a)	0.62 (n/a)	0.62 (n/a)	0.62 (n/a)	0.00 (n/a)

llama_3.2_1b_prompt_2048_tokens_1

Commit/Date	Num_Tokens (max)	Num_Tokens (mean)	Num_Tokens (median)	Num_Tokens (min)	Num_Tokens (stddev)	TPS (max)	TPS (mean)	TPS (median)	TPS (min)	TPS (stddev)	TTFT (max)	TTFT (mean)	TTFT (median)	TTFT (min)	TTFT (stddev)
`1531781` — 2026-01-23 18:40:34	1.00 (+0.00%)	1.00 (+0.00%)	1.00 (+0.00%)	1.00 (+0.00%)	0.00 (n/a)	0.00 (n/a)	0.00 (n/a)	0.00 (n/a)	0.00 (n/a)	0.00 (n/a)	2.70 (-0.29%)	2.70 (-0.29%)	2.70 (-0.29%)	2.70 (-0.29%)	0.00 (n/a)
`0f0da0b` — 2026-01-14 16:41:39	1.00 (n/a)	1.00 (n/a)	1.00 (n/a)	1.00 (n/a)	0.00 (n/a)	0.00 (n/a)	0.00 (n/a)	0.00 (n/a)	0.00 (n/a)	0.00 (n/a)	2.71 (n/a)	2.71 (n/a)	2.71 (n/a)	2.71 (n/a)	0.00 (n/a)

llama_3.2_1b_prompt_2048_tokens_40

Commit/Date	Num_Tokens (max)	Num_Tokens (mean)	Num_Tokens (median)	Num_Tokens (min)	Num_Tokens (stddev)	TPS (max)	TPS (mean)	TPS (median)	TPS (min)	TPS (stddev)	TTFT (max)	TTFT (mean)	TTFT (median)	TTFT (min)	TTFT (stddev)
`1531781` — 2026-01-23 18:40:34	40.00 (+0.00%)	40.00 (+0.00%)	40.00 (+0.00%)	40.00 (+0.00%)	0.00 (n/a)	4.22 (+2.18%)	4.22 (+2.18%)	4.22 (+2.18%)	4.22 (+2.18%)	0.00 (n/a)	2.70 (-0.24%)	2.70 (-0.24%)	2.70 (-0.24%)	2.70 (-0.24%)	0.00 (n/a)
`0f0da0b` — 2026-01-14 16:41:39	40.00 (n/a)	40.00 (n/a)	40.00 (n/a)	40.00 (n/a)	0.00 (n/a)	4.13 (n/a)	4.13 (n/a)	4.13 (n/a)	4.13 (n/a)	0.00 (n/a)	2.71 (n/a)	2.71 (n/a)	2.71 (n/a)	2.71 (n/a)	0.00 (n/a)

github-actions · 2026-02-06T16:24:54Z

📊 Test Results for Small Benchmark/Test Suite

b0ad401 (2026_02_06_16_24_33)

IRONCLAD

Tested on 2026_02_06_16_24_33 at commit b0ad401.

Test	Checks	Latency (mean)	Bandwidth (mean)	GFLOP/s (mean)

📈 Trends (vs main branch) for Small Benchmark/Test Suite

b0ad401 (2026_02_06_16_24_33)

IRONCLAD Trends

axpy_1_cols_2_channels_2048_tile_2048_3.0

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`130b6ea` — 2025-12-05 21:15:49	0.13 (-1.96%)	0.11 (+2.52%)	0.10 (+1.86%)	0.08 (-1.91%)	0.02 (-1.72%)	162.40 (+1.95%)	120.96 (-2.36%)	117.00 (-1.85%)	94.40 (+2.05%)	23.11 (+4.09%)
`0a6c11c` — 2025-12-04 00:39:10	0.13 (n/a)	0.10 (n/a)	0.10 (n/a)	0.08 (n/a)	0.02 (n/a)	159.30 (n/a)	123.88 (n/a)	119.20 (n/a)	92.50 (n/a)	22.21 (n/a)

axpy_1_cols_2_channels_2048_tile_2048_3.0_0

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`1531781` — 2026-01-23 18:49:24	0.16 (+80.10%)	0.13 (+54.74%)	0.11 (+34.32%)	0.10 (+39.85%)	0.03 (+432.10%)	118.80 (-28.48%)	100.30 (-33.22%)	110.40 (-25.56%)	78.00 (-44.48%)	20.49 (+105.98%)
`0f0da0b` — 2026-01-14 16:31:00	0.09 (n/a)	0.08 (n/a)	0.08 (n/a)	0.07 (n/a)	0.01 (n/a)	166.10 (n/a)	150.20 (n/a)	148.30 (n/a)	140.50 (n/a)	9.95 (n/a)

axpy_2_cols_2_channels_2048_tile_1024_3.0

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`130b6ea` — 2025-12-05 21:15:49	0.11 (-26.77%)	0.09 (-9.09%)	0.09 (-10.81%)	0.08 (+8.25%)	0.01 (-55.42%)	152.00 (-7.60%)	135.46 (+6.28%)	142.80 (+12.18%)	115.00 (+36.58%)	15.03 (-41.36%)
`0a6c11c` — 2025-12-04 00:39:10	0.15 (n/a)	0.10 (n/a)	0.10 (n/a)	0.07 (n/a)	0.02 (n/a)	164.50 (n/a)	127.46 (n/a)	127.30 (n/a)	84.20 (n/a)	25.64 (n/a)

axpy_2_cols_2_channels_2048_tile_1024_3.0_0

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`1531781` — 2026-01-23 18:49:24	0.14 (+54.25%)	0.12 (+45.65%)	0.14 (+61.31%)	0.08 (-2.79%)	0.03 (+557.87%)	156.90 (+2.89%)	104.46 (-28.09%)	90.30 (-37.98%)	87.90 (-35.22%)	29.52 (+351.33%)
`0f0da0b` — 2026-01-14 16:31:00	0.09 (n/a)	0.08 (n/a)	0.08 (n/a)	0.08 (n/a)	0.00 (n/a)	152.50 (n/a)	145.26 (n/a)	145.60 (n/a)	135.70 (n/a)	6.54 (n/a)

axpy_4_cols_2_channels_2048_tile_512_3.0

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`130b6ea` — 2025-12-05 21:15:49	0.12 (-10.21%)	0.10 (-8.41%)	0.10 (-6.20%)	0.07 (+3.52%)	0.02 (-23.48%)	166.70 (-3.36%)	128.78 (+7.33%)	121.80 (+6.65%)	103.90 (+11.36%)	22.95 (-18.82%)
`0a6c11c` — 2025-12-04 00:39:10	0.13 (n/a)	0.11 (n/a)	0.11 (n/a)	0.07 (n/a)	0.02 (n/a)	172.50 (n/a)	119.98 (n/a)	114.20 (n/a)	93.30 (n/a)	28.27 (n/a)

axpy_4_cols_2_channels_2048_tile_512_3.0_0

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`1531781` — 2026-01-23 18:49:24	0.16 (+65.19%)	0.14 (+57.55%)	0.15 (+63.48%)	0.09 (+24.56%)	0.03 (+220.28%)	134.10 (-19.70%)	92.44 (-34.51%)	83.40 (-38.81%)	78.60 (-39.49%)	23.38 (+58.01%)
`0f0da0b` — 2026-01-14 16:31:00	0.09 (n/a)	0.09 (n/a)	0.09 (n/a)	0.07 (n/a)	0.01 (n/a)	167.00 (n/a)	141.16 (n/a)	136.30 (n/a)	129.90 (n/a)	14.80 (n/a)

axpy_8_cols_2_channels_2048_tile_256_3.0

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`130b6ea` — 2025-12-05 21:15:49	0.10 (-16.19%)	0.08 (-10.40%)	0.09 (-2.09%)	0.07 (-17.38%)	0.01 (-10.61%)	184.40 (+21.00%)	153.18 (+11.99%)	144.10 (+2.20%)	127.10 (+19.34%)	21.26 (+34.76%)
`0a6c11c` — 2025-12-04 00:39:10	0.12 (n/a)	0.09 (n/a)	0.09 (n/a)	0.08 (n/a)	0.01 (n/a)	152.40 (n/a)	136.78 (n/a)	141.00 (n/a)	106.50 (n/a)	15.78 (n/a)

axpy_8_cols_2_channels_2048_tile_256_3.0_0

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`1531781` — 2026-01-23 18:49:24	0.17 (+67.03%)	0.14 (+66.86%)	0.15 (+90.72%)	0.08 (+20.53%)	0.03 (+162.97%)	146.80 (-17.06%)	92.24 (-37.11%)	80.00 (-47.54%)	72.30 (-40.15%)	31.07 (+36.92%)
`0f0da0b` — 2026-01-14 16:31:00	0.10 (n/a)	0.09 (n/a)	0.08 (n/a)	0.07 (n/a)	0.01 (n/a)	177.00 (n/a)	146.68 (n/a)	152.50 (n/a)	120.80 (n/a)	22.69 (n/a)

dequant_1_cols_1_channels_2048_tile_2048

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`130b6ea` — 2025-12-05 21:15:49	0.05 (-18.83%)	0.04 (-5.21%)	0.04 (-2.69%)	0.03 (+37.10%)	0.00 (-53.98%)	160.40 (-27.06%)	141.12 (-0.91%)	136.40 (+2.79%)	114.30 (+23.17%)	17.37 (-59.06%)
`0a6c11c` — 2025-12-04 00:39:10	0.06 (n/a)	0.04 (n/a)	0.04 (n/a)	0.02 (n/a)	0.01 (n/a)	219.90 (n/a)	142.42 (n/a)	132.70 (n/a)	92.80 (n/a)	42.43 (n/a)

dequant_1_cols_1_channels_2048_tile_2048_0

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`1531781` — 2026-01-23 18:49:24	0.06 (+13.11%)	0.05 (+1.55%)	0.04 (-0.75%)	0.04 (+7.05%)	0.01 (+25.85%)	125.50 (-6.55%)	115.08 (-1.12%)	121.60 (+0.75%)	88.10 (-11.63%)	15.40 (+4.09%)
`0f0da0b` — 2026-01-14 16:31:00	0.05 (n/a)	0.05 (n/a)	0.04 (n/a)	0.04 (n/a)	0.01 (n/a)	134.30 (n/a)	116.38 (n/a)	120.70 (n/a)	99.70 (n/a)	14.79 (n/a)

dequant_1_cols_2_channels_2048_tile_1024

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`130b6ea` — 2025-12-05 21:15:49	0.06 (-13.28%)	0.05 (-8.93%)	0.06 (-13.59%)	0.04 (+4.63%)	0.01 (-39.10%)	133.60 (-4.43%)	106.08 (+5.85%)	95.10 (+15.69%)	87.90 (+15.35%)	17.90 (-31.97%)
`0a6c11c` — 2025-12-04 00:39:10	0.07 (n/a)	0.06 (n/a)	0.06 (n/a)	0.04 (n/a)	0.01 (n/a)	139.80 (n/a)	100.22 (n/a)	82.20 (n/a)	76.20 (n/a)	26.31 (n/a)

dequant_1_cols_2_channels_2048_tile_1024_0

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`1531781` — 2026-01-23 18:49:24	0.07 (+24.45%)	0.05 (+9.74%)	0.05 (-2.94%)	0.03 (+16.88%)	0.01 (+14.24%)	163.70 (-14.47%)	114.66 (-9.57%)	111.30 (+3.06%)	76.50 (-19.64%)	31.63 (-21.04%)
`0f0da0b` — 2026-01-14 16:31:00	0.06 (n/a)	0.04 (n/a)	0.05 (n/a)	0.03 (n/a)	0.01 (n/a)	191.40 (n/a)	126.80 (n/a)	108.00 (n/a)	95.20 (n/a)	40.06 (n/a)

dequant_2_cols_1_channels_2048_tile_1024

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`130b6ea` — 2025-12-05 21:15:49	0.07 (+5.66%)	0.05 (+1.21%)	0.05 (+0.61%)	0.04 (-15.02%)	0.01 (+58.42%)	143.40 (+17.64%)	101.02 (+1.49%)	97.60 (-0.61%)	78.90 (-5.40%)	22.56 (+78.52%)
`0a6c11c` — 2025-12-04 00:39:10	0.06 (n/a)	0.05 (n/a)	0.05 (n/a)	0.04 (n/a)	0.01 (n/a)	121.90 (n/a)	99.54 (n/a)	98.20 (n/a)	83.40 (n/a)	12.64 (n/a)

dequant_2_cols_1_channels_2048_tile_1024_0

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`1531781` — 2026-01-23 18:49:24	0.05 (+2.33%)	0.05 (+22.41%)	0.05 (+32.19%)	0.04 (+7.46%)	0.01 (-3.28%)	147.80 (-6.93%)	113.14 (-18.51%)	106.50 (-24.31%)	101.40 (-2.31%)	19.61 (-8.79%)
`0f0da0b` — 2026-01-14 16:31:00	0.05 (n/a)	0.04 (n/a)	0.04 (n/a)	0.03 (n/a)	0.01 (n/a)	158.80 (n/a)	138.84 (n/a)	140.70 (n/a)	103.80 (n/a)	21.51 (n/a)

dequant_2_cols_2_channels_2048_tile_512

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`130b6ea` — 2025-12-05 21:15:49	0.05 (+17.06%)	0.04 (+8.98%)	0.05 (+21.37%)	0.03 (-11.65%)	0.01 (+301.89%)	159.70 (+13.18%)	126.32 (-5.68%)	109.00 (-17.67%)	106.80 (-14.56%)	22.48 (+280.15%)
`0a6c11c` — 2025-12-04 00:39:10	0.04 (n/a)	0.04 (n/a)	0.04 (n/a)	0.04 (n/a)	0.00 (n/a)	141.10 (n/a)	133.92 (n/a)	132.40 (n/a)	125.00 (n/a)	5.91 (n/a)

dequant_2_cols_2_channels_2048_tile_512_0

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`1531781` — 2026-01-23 18:49:24	0.05 (+14.83%)	0.04 (+10.04%)	0.05 (+21.96%)	0.03 (-12.01%)	0.01 (+143.96%)	158.30 (+13.64%)	121.38 (-6.49%)	106.20 (-17.99%)	98.20 (-12.94%)	26.13 (+141.52%)
`0f0da0b` — 2026-01-14 16:31:00	0.05 (n/a)	0.04 (n/a)	0.04 (n/a)	0.04 (n/a)	0.00 (n/a)	139.30 (n/a)	129.80 (n/a)	129.50 (n/a)	112.80 (n/a)	10.82 (n/a)

dequant_4_cols_1_channels_2048_tile_512

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`130b6ea` — 2025-12-05 21:15:49	0.06 (+20.24%)	0.05 (-3.19%)	0.05 (-6.73%)	0.03 (-20.69%)	0.01 (+194.26%)	156.10 (+26.09%)	120.86 (+7.15%)	114.90 (+7.18%)	88.30 (-16.85%)	24.40 (+214.82%)
`0a6c11c` — 2025-12-04 00:39:10	0.05 (n/a)	0.05 (n/a)	0.05 (n/a)	0.04 (n/a)	0.00 (n/a)	123.80 (n/a)	112.80 (n/a)	107.20 (n/a)	106.20 (n/a)	7.75 (n/a)

dequant_4_cols_1_channels_2048_tile_512_0

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`1531781` — 2026-01-23 18:49:24	0.06 (+28.85%)	0.05 (+16.79%)	0.05 (+23.60%)	0.03 (-12.66%)	0.01 (+157.83%)	164.80 (+14.52%)	113.62 (-10.20%)	103.20 (-19.12%)	84.80 (-22.42%)	32.55 (+130.11%)
`0f0da0b` — 2026-01-14 16:31:00	0.05 (n/a)	0.04 (n/a)	0.04 (n/a)	0.04 (n/a)	0.00 (n/a)	143.90 (n/a)	126.52 (n/a)	127.60 (n/a)	109.30 (n/a)	14.15 (n/a)

dequant_4_cols_2_channels_2048_tile_256

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`130b6ea` — 2025-12-05 21:15:49	0.05 (-6.09%)	0.04 (-8.04%)	0.04 (-12.46%)	0.03 (-0.41%)	0.01 (-20.19%)	168.10 (+0.42%)	135.74 (+7.44%)	132.20 (+14.26%)	107.40 (+6.44%)	22.91 (-12.67%)
`0a6c11c` — 2025-12-04 00:39:10	0.05 (n/a)	0.04 (n/a)	0.05 (n/a)	0.03 (n/a)	0.01 (n/a)	167.40 (n/a)	126.34 (n/a)	115.70 (n/a)	100.90 (n/a)	26.23 (n/a)

dequant_4_cols_2_channels_2048_tile_256_0

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`1531781` — 2026-01-23 18:49:24	0.07 (+42.90%)	0.05 (+60.94%)	0.06 (+84.55%)	0.03 (+50.15%)	0.01 (+27.03%)	188.40 (-33.40%)	112.72 (-39.08%)	93.80 (-45.84%)	79.70 (-29.96%)	43.58 (-36.67%)
`0f0da0b` — 2026-01-14 16:31:00	0.05 (n/a)	0.03 (n/a)	0.03 (n/a)	0.02 (n/a)	0.01 (n/a)	282.90 (n/a)	185.04 (n/a)	173.20 (n/a)	113.80 (n/a)	68.82 (n/a)

dequant_8_cols_1_channels_2048_tile_256

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`130b6ea` — 2025-12-05 21:15:49	0.04 (-13.65%)	0.04 (-13.67%)	0.04 (-18.56%)	0.03 (-13.52%)	0.00 (-23.88%)	170.40 (+15.68%)	140.12 (+15.33%)	135.80 (+22.78%)	119.60 (+15.78%)	17.94 (+1.87%)
`0a6c11c` — 2025-12-04 00:39:10	0.05 (n/a)	0.04 (n/a)	0.05 (n/a)	0.04 (n/a)	0.01 (n/a)	147.30 (n/a)	121.50 (n/a)	110.60 (n/a)	103.30 (n/a)	17.61 (n/a)

dequant_8_cols_1_channels_2048_tile_256_0

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`1531781` — 2026-01-23 18:49:24	0.04 (-5.92%)	0.04 (-11.07%)	0.04 (-12.74%)	0.03 (-9.17%)	0.00 (-4.85%)	166.90 (+10.09%)	138.66 (+12.51%)	135.90 (+14.59%)	117.60 (+6.33%)	17.94 (+10.10%)
`0f0da0b` — 2026-01-14 16:31:00	0.05 (n/a)	0.04 (n/a)	0.04 (n/a)	0.03 (n/a)	0.00 (n/a)	151.60 (n/a)	123.24 (n/a)	118.60 (n/a)	110.60 (n/a)	16.30 (n/a)

dequant_8_cols_2_channels_2048_tile_128

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`130b6ea` — 2025-12-05 21:15:49	0.05 (+17.21%)	0.03 (-1.70%)	0.03 (-3.04%)	0.03 (-16.20%)	0.01 (+95.59%)	198.20 (+19.33%)	160.00 (+4.92%)	166.20 (+3.17%)	106.50 (-14.66%)	29.86 (+87.19%)
`0a6c11c` — 2025-12-04 00:39:10	0.04 (n/a)	0.03 (n/a)	0.03 (n/a)	0.03 (n/a)	0.00 (n/a)	166.10 (n/a)	152.50 (n/a)	161.10 (n/a)	124.80 (n/a)	15.95 (n/a)

dequant_8_cols_2_channels_2048_tile_128_0

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`1531781` — 2026-01-23 18:49:24	0.05 (+10.58%)	0.04 (+11.15%)	0.04 (+32.75%)	0.02 (-15.30%)	0.01 (+23.92%)	229.90 (+18.08%)	145.60 (-7.34%)	131.00 (-24.67%)	106.20 (-9.54%)	49.03 (+42.68%)
`0f0da0b` — 2026-01-14 16:31:00	0.04 (n/a)	0.03 (n/a)	0.03 (n/a)	0.03 (n/a)	0.01 (n/a)	194.70 (n/a)	157.14 (n/a)	173.90 (n/a)	117.40 (n/a)	34.37 (n/a)

eltwise_add_1_cols_2_channels_2048_tile_2048

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`1531781` — 2026-01-23 18:49:24	0.17 (+59.76%)	0.12 (+36.17%)	0.10 (+5.69%)	0.08 (+17.86%)	0.04 (+129.37%)	155.80 (-15.19%)	109.78 (-22.78%)	120.50 (-5.34%)	73.80 (-37.40%)	34.38 (+16.64%)
`0f0da0b` — 2026-01-14 16:31:00	0.10 (n/a)	0.09 (n/a)	0.10 (n/a)	0.07 (n/a)	0.02 (n/a)	183.70 (n/a)	142.16 (n/a)	127.30 (n/a)	117.90 (n/a)	29.48 (n/a)

eltwise_add_2_cols_2_channels_2048_tile_1024

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`1531781` — 2026-01-23 18:49:24	0.15 (+24.10%)	0.13 (+23.91%)	0.13 (+30.32%)	0.08 (+16.80%)	0.03 (+31.09%)	146.70 (-14.36%)	101.20 (-18.68%)	91.70 (-23.26%)	81.40 (-19.41%)	26.00 (-6.71%)
`0f0da0b` — 2026-01-14 16:31:00	0.12 (n/a)	0.10 (n/a)	0.10 (n/a)	0.07 (n/a)	0.02 (n/a)	171.30 (n/a)	124.44 (n/a)	119.50 (n/a)	101.00 (n/a)	27.87 (n/a)

eltwise_add_4_cols_2_channels_2048_tile_512

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`1531781` — 2026-01-23 18:49:24	0.17 (+52.81%)	0.15 (+77.83%)	0.15 (+81.17%)	0.11 (+106.70%)	0.02 (+14.06%)	116.00 (-51.65%)	86.46 (-45.77%)	80.90 (-44.82%)	72.70 (-34.56%)	17.22 (-64.39%)
`0f0da0b` — 2026-01-14 16:31:00	0.11 (n/a)	0.08 (n/a)	0.08 (n/a)	0.05 (n/a)	0.02 (n/a)	239.90 (n/a)	159.44 (n/a)	146.60 (n/a)	111.10 (n/a)	48.37 (n/a)

eltwise_add_8_cols_2_channels_2048_tile_256

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`1531781` — 2026-01-23 18:49:24	0.16 (+58.53%)	0.14 (+73.02%)	0.14 (+70.95%)	0.11 (+188.94%)	0.02 (-23.24%)	109.50 (-65.38%)	90.00 (-48.14%)	86.00 (-41.50%)	75.30 (-36.93%)	13.44 (-83.58%)
`0f0da0b` — 2026-01-14 16:31:00	0.10 (n/a)	0.08 (n/a)	0.08 (n/a)	0.04 (n/a)	0.03 (n/a)	316.30 (n/a)	173.56 (n/a)	147.00 (n/a)	119.40 (n/a)	81.81 (n/a)

eltwise_mul_1_cols_2_channels_2048_tile_2048

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`1531781` — 2026-01-23 18:49:24	0.14 (+42.19%)	0.13 (+64.04%)	0.13 (+74.51%)	0.10 (+75.35%)	0.01 (+0.96%)	118.10 (-42.97%)	97.88 (-39.87%)	92.60 (-42.70%)	90.40 (-29.70%)	11.55 (-59.47%)
`0f0da0b` — 2026-01-14 16:31:00	0.10 (n/a)	0.08 (n/a)	0.08 (n/a)	0.06 (n/a)	0.01 (n/a)	207.10 (n/a)	162.78 (n/a)	161.60 (n/a)	128.60 (n/a)	28.50 (n/a)

eltwise_mul_2_cols_2_channels_2048_tile_1024

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`1531781` — 2026-01-23 18:49:24	0.13 (+44.68%)	0.12 (+43.52%)	0.13 (+47.71%)	0.09 (+35.56%)	0.02 (+80.76%)	130.00 (-26.26%)	102.10 (-29.85%)	93.30 (-32.34%)	91.10 (-30.88%)	16.53 (-8.94%)
`0f0da0b` — 2026-01-14 16:31:00	0.09 (n/a)	0.09 (n/a)	0.09 (n/a)	0.07 (n/a)	0.01 (n/a)	176.30 (n/a)	145.54 (n/a)	137.90 (n/a)	131.80 (n/a)	18.15 (n/a)

eltwise_mul_4_cols_2_channels_2048_tile_512

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`1531781` — 2026-01-23 18:49:24	0.17 (+63.62%)	0.15 (+73.87%)	0.15 (+72.82%)	0.11 (+104.66%)	0.02 (+10.66%)	114.00 (-51.16%)	86.18 (-44.46%)	81.50 (-42.12%)	74.20 (-38.88%)	15.85 (-65.75%)
`0f0da0b` — 2026-01-14 16:31:00	0.10 (n/a)	0.08 (n/a)	0.09 (n/a)	0.05 (n/a)	0.02 (n/a)	233.40 (n/a)	155.18 (n/a)	140.80 (n/a)	121.40 (n/a)	46.29 (n/a)

eltwise_mul_8_cols_2_channels_2048_tile_256

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`1531781` — 2026-01-23 18:49:24	0.13 (+46.12%)	0.11 (+59.06%)	0.12 (+66.26%)	0.08 (+48.55%)	0.02 (+55.56%)	154.80 (-32.70%)	112.28 (-36.89%)	104.60 (-39.85%)	95.50 (-31.59%)	24.14 (-26.88%)
`0f0da0b` — 2026-01-14 16:31:00	0.09 (n/a)	0.07 (n/a)	0.07 (n/a)	0.05 (n/a)	0.01 (n/a)	230.00 (n/a)	177.92 (n/a)	173.90 (n/a)	139.60 (n/a)	33.01 (n/a)

gelu_1_cols_1_channels_2048_tile_2048

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`1531781` — 2026-01-23 18:49:24	0.11 (+53.19%)	0.10 (+56.21%)	0.11 (+82.01%)	0.08 (+32.44%)	0.02 (+220.41%)	101.60 (-24.52%)	83.80 (-34.61%)	72.40 (-45.11%)	72.20 (-34.72%)	15.75 (+58.78%)
`0f0da0b` — 2026-01-14 16:31:00	0.07 (n/a)	0.06 (n/a)	0.06 (n/a)	0.06 (n/a)	0.01 (n/a)	134.60 (n/a)	128.16 (n/a)	131.90 (n/a)	110.60 (n/a)	9.92 (n/a)

gelu_1_cols_2_channels_2048_tile_1024

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`1531781` — 2026-01-23 18:49:24	0.12 (+63.09%)	0.11 (+76.43%)	0.12 (+100.14%)	0.07 (+62.57%)	0.02 (+62.52%)	110.00 (-38.48%)	80.26 (-43.34%)	69.10 (-50.00%)	66.40 (-38.69%)	18.54 (-39.34%)
`0f0da0b` — 2026-01-14 16:31:00	0.08 (n/a)	0.06 (n/a)	0.06 (n/a)	0.05 (n/a)	0.01 (n/a)	178.80 (n/a)	141.66 (n/a)	138.20 (n/a)	108.30 (n/a)	30.55 (n/a)

gelu_2_cols_1_channels_2048_tile_1024

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`1531781` — 2026-01-23 18:49:24	0.12 (+51.06%)	0.10 (+80.20%)	0.12 (+120.49%)	0.06 (+32.94%)	0.03 (+101.39%)	137.80 (-24.74%)	85.02 (-42.29%)	68.20 (-54.65%)	67.80 (-33.79%)	30.24 (+4.82%)
`0f0da0b` — 2026-01-14 16:31:00	0.08 (n/a)	0.06 (n/a)	0.05 (n/a)	0.04 (n/a)	0.01 (n/a)	183.10 (n/a)	147.32 (n/a)	150.40 (n/a)	102.40 (n/a)	28.85 (n/a)

gelu_2_cols_2_channels_2048_tile_512

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`1531781` — 2026-01-23 18:49:24	0.12 (+49.86%)	0.11 (+81.45%)	0.12 (+78.91%)	0.08 (+119.08%)	0.02 (+8.63%)	107.80 (-54.38%)	76.80 (-47.80%)	69.50 (-44.13%)	67.10 (-33.23%)	17.40 (-67.51%)
`0f0da0b` — 2026-01-14 16:31:00	0.08 (n/a)	0.06 (n/a)	0.07 (n/a)	0.03 (n/a)	0.02 (n/a)	236.30 (n/a)	147.12 (n/a)	124.40 (n/a)	100.50 (n/a)	53.56 (n/a)

gelu_4_cols_1_channels_2048_tile_512

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`1531781` — 2026-01-23 18:49:24	0.13 (+73.67%)	0.11 (+92.11%)	0.12 (+121.10%)	0.05 (+47.52%)	0.03 (+98.64%)	153.10 (-32.20%)	85.56 (-45.77%)	69.30 (-54.77%)	63.00 (-42.47%)	37.94 (-17.18%)
`0f0da0b` — 2026-01-14 16:31:00	0.07 (n/a)	0.06 (n/a)	0.05 (n/a)	0.04 (n/a)	0.02 (n/a)	225.80 (n/a)	157.78 (n/a)	153.20 (n/a)	109.50 (n/a)	45.81 (n/a)

gelu_4_cols_2_channels_2048_tile_256

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`1531781` — 2026-01-23 18:49:24	0.12 (+87.02%)	0.11 (+117.65%)	0.12 (+144.58%)	0.07 (+103.45%)	0.02 (+63.92%)	110.50 (-50.82%)	79.80 (-54.61%)	70.50 (-59.08%)	70.00 (-46.56%)	17.48 (-56.79%)
`0f0da0b` — 2026-01-14 16:31:00	0.06 (n/a)	0.05 (n/a)	0.05 (n/a)	0.04 (n/a)	0.01 (n/a)	224.70 (n/a)	175.80 (n/a)	172.30 (n/a)	131.00 (n/a)	40.46 (n/a)

gelu_8_cols_1_channels_2048_tile_256

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`1531781` — 2026-01-23 18:49:24	0.12 (+61.36%)	0.10 (+65.53%)	0.11 (+79.09%)	0.04 (-5.63%)	0.03 (+170.92%)	208.20 (+5.95%)	98.74 (-30.26%)	71.90 (-44.13%)	70.20 (-38.04%)	61.20 (+79.92%)
`0f0da0b` — 2026-01-14 16:31:00	0.07 (n/a)	0.06 (n/a)	0.06 (n/a)	0.04 (n/a)	0.01 (n/a)	196.50 (n/a)	141.58 (n/a)	128.70 (n/a)	113.30 (n/a)	34.01 (n/a)

gelu_8_cols_2_channels_2048_tile_128

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`1531781` — 2026-01-23 18:49:24	0.10 (+64.20%)	0.09 (+72.61%)	0.10 (+77.79%)	0.06 (+39.91%)	0.02 (+119.55%)	133.40 (-28.55%)	93.44 (-40.99%)	82.30 (-43.75%)	81.90 (-39.11%)	22.47 (-5.49%)
`0f0da0b` — 2026-01-14 16:31:00	0.06 (n/a)	0.05 (n/a)	0.06 (n/a)	0.04 (n/a)	0.01 (n/a)	186.70 (n/a)	158.34 (n/a)	146.30 (n/a)	134.50 (n/a)	23.77 (n/a)

gemm_1792x896x1152_64x32x48_8cols_ccolmaj

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)	Throughput (max)	Throughput (mean)	Throughput (median)	Throughput (min)	Throughput (stddev)
`1531781` — 2026-01-23 18:49:24	5.15 (+4.03%)	4.90 (+9.86%)	4.99 (+12.67%)	4.27 (+4.36%)	0.36 (+11.38%)	2201.70 (-4.17%)	1929.94 (-8.91%)	1885.00 (-11.25%)	1826.70 (-3.87%)	154.05 (+4.53%)	2025.19 (+4.03%)	1925.76 (+9.86%)	1962.51 (+12.67%)	1680.20 (+4.36%)	139.97 (+11.38%)
`0f0da0b` — 2026-01-14 16:31:00	4.95 (n/a)	4.46 (n/a)	4.43 (n/a)	4.09 (n/a)	0.32 (n/a)	2297.60 (n/a)	2118.80 (n/a)	2123.90 (n/a)	1900.20 (n/a)	147.37 (n/a)	1946.79 (n/a)	1752.94 (n/a)	1741.75 (n/a)	1610.08 (n/a)	125.67 (n/a)

gemm_192x384x64_48x96x16_4cols

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)	Throughput (max)	Throughput (mean)	Throughput (median)	Throughput (min)	Throughput (stddev)
`1531781` — 2026-01-23 18:49:24	2.36 (+52.61%)	1.85 (+39.10%)	1.78 (+41.59%)	1.32 (+14.31%)	0.39 (+132.56%)	167.80 (-12.51%)	124.24 (-26.22%)	124.40 (-29.36%)	93.70 (-34.52%)	28.11 (+35.82%)	100.67 (+52.61%)	78.94 (+39.10%)	75.88 (+41.59%)	56.25 (+14.31%)	16.73 (+132.56%)
`0f0da0b` — 2026-01-14 16:31:00	1.55 (n/a)	1.33 (n/a)	1.26 (n/a)	1.15 (n/a)	0.17 (n/a)	191.80 (n/a)	168.40 (n/a)	176.10 (n/a)	143.10 (n/a)	20.70 (n/a)	65.96 (n/a)	56.75 (n/a)	53.59 (n/a)	49.20 (n/a)	7.20 (n/a)

gemm_192x384x64_48x96x16_4cols_bcolmaj_ccolmaj

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)	Throughput (max)	Throughput (mean)	Throughput (median)	Throughput (min)	Throughput (stddev)
`1531781` — 2026-01-23 18:49:24	2.22 (+30.89%)	2.06 (+54.21%)	2.20 (+75.49%)	1.60 (+35.74%)	0.26 (+25.57%)	138.20 (-26.33%)	109.20 (-35.27%)	100.50 (-43.03%)	99.80 (-23.58%)	16.54 (-28.10%)	94.54 (+30.89%)	87.77 (+54.21%)	93.91 (+75.49%)	68.30 (+35.74%)	11.26 (+25.57%)
`0f0da0b` — 2026-01-14 16:31:00	1.69 (n/a)	1.33 (n/a)	1.25 (n/a)	1.18 (n/a)	0.21 (n/a)	187.60 (n/a)	168.70 (n/a)	176.40 (n/a)	130.60 (n/a)	23.00 (n/a)	72.23 (n/a)	56.92 (n/a)	53.51 (n/a)	50.31 (n/a)	8.97 (n/a)

gemm_2048x2048x2048_64x64x32_8_cols_0_bcolmaj_0_ccolmaj_0

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	GFLOP/s (max)	GFLOP/s (mean)	GFLOP/s (median)	GFLOP/s (min)	GFLOP/s (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`3a053d3` — 2025-11-26 01:12:16	3.72 (-0.11%)	3.55 (-2.75%)	3.56 (-2.75%)	3.36 (-5.33%)	0.12 (+76.10%)	2725.23 (-0.11%)	2602.52 (-2.75%)	2612.11 (-2.75%)	2459.89 (-5.33%)	86.27 (+76.10%)	6984.00 (+5.63%)	6608.60 (+2.91%)	6577.00 (+2.83%)	6304.00 (+0.11%)	221.87 (+87.15%)
`9e98049` — 2025-11-25 20:50:55	3.72 (n/a)	3.65 (n/a)	3.66 (n/a)	3.54 (n/a)	0.07 (n/a)	2728.26 (n/a)	2676.06 (n/a)	2686.03 (n/a)	2598.29 (n/a)	48.99 (n/a)	6612.00 (n/a)	6422.00 (n/a)	6396.00 (n/a)	6297.00 (n/a)	118.55 (n/a)

gemm_2048x2048x2048_64x64x32_8_cols_0_bcolmaj_1_ccolmaj_0

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	GFLOP/s (max)	GFLOP/s (mean)	GFLOP/s (median)	GFLOP/s (min)	GFLOP/s (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`3a053d3` — 2025-11-26 01:12:16	3.40 (-3.14%)	3.38 (-0.82%)	3.38 (-0.82%)	3.35 (+4.93%)	0.02 (-84.51%)	2493.81 (-3.14%)	2475.27 (-0.82%)	2479.06 (-0.82%)	2453.92 (+4.93%)	13.33 (-84.51%)	7001.00 (-4.70%)	6940.80 (+0.71%)	6930.00 (+0.83%)	6889.00 (+3.24%)	37.43 (-84.76%)
`9e98049` — 2025-11-25 20:50:55	3.51 (n/a)	3.40 (n/a)	3.41 (n/a)	3.19 (n/a)	0.12 (n/a)	2574.53 (n/a)	2495.72 (n/a)	2499.62 (n/a)	2338.67 (n/a)	86.02 (n/a)	7346.00 (n/a)	6892.20 (n/a)	6873.00 (n/a)	6673.00 (n/a)	245.57 (n/a)

gemm_2048x2048x2048_64x64x32_8_cols_1_bcolmaj_0_ccolmaj_0

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	GFLOP/s (max)	GFLOP/s (mean)	GFLOP/s (median)	GFLOP/s (min)	GFLOP/s (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`3a053d3` — 2025-11-26 01:12:16	3.02 (+0.52%)	2.87 (-2.05%)	2.86 (-2.24%)	2.70 (-6.16%)	0.11 (+131.66%)	2216.19 (+0.52%)	2104.44 (-2.05%)	2094.08 (-2.24%)	1980.84 (-6.16%)	80.25 (+131.67%)	8673.00 (+6.56%)	8175.60 (+2.21%)	8204.00 (+2.29%)	7752.00 (-0.51%)	314.03 (+145.36%)
`9e98049` — 2025-11-25 20:50:55	3.01 (n/a)	2.93 (n/a)	2.92 (n/a)	2.88 (n/a)	0.05 (n/a)	2204.81 (n/a)	2148.41 (n/a)	2142.13 (n/a)	2110.81 (n/a)	34.64 (n/a)	8139.00 (n/a)	7998.60 (n/a)	8020.00 (n/a)	7792.00 (n/a)	127.99 (n/a)

gemm_2048x2048x2048_64x64x64_1cols

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)	Throughput (max)	Throughput (mean)	Throughput (median)	Throughput (min)	Throughput (stddev)
`1531781` — 2026-01-23 18:49:24	0.52 (-0.75%)	0.52 (-0.07%)	0.52 (+0.01%)	0.52 (+0.22%)	0.00 (-86.34%)	48415.10 (-0.22%)	48370.52 (+0.07%)	48362.20 (-0.01%)	48344.40 (+0.76%)	29.40 (-86.25%)	355.36 (-0.75%)	355.17 (-0.07%)	355.23 (+0.01%)	354.85 (+0.22%)	0.22 (-86.34%)
`0f0da0b` — 2026-01-14 16:31:00	0.52 (n/a)	0.52 (n/a)	0.52 (n/a)	0.52 (n/a)	0.00 (n/a)	48520.80 (n/a)	48335.84 (n/a)	48365.40 (n/a)	47981.00 (n/a)	213.79 (n/a)	358.06 (n/a)	355.43 (n/a)	355.21 (n/a)	354.07 (n/a)	1.58 (n/a)

gemm_2048x2048x2048_64x64x64_2_cols_0_bcolmaj_0_ccolmaj_0

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`130b6ea` — 2025-12-05 21:15:49	1.04 (+0.94%)	1.03 (+0.23%)	1.03 (+0.09%)	1.02 (+0.07%)	0.00 (+197.51%)	24576.00 (-0.07%)	24485.82 (-0.23%)	24530.40 (-0.09%)	24266.40 (-0.94%)	111.26 (+194.35%)
`0a6c11c` — 2025-12-04 00:39:10	1.03 (n/a)	1.03 (n/a)	1.03 (n/a)	1.02 (n/a)	0.00 (n/a)	24594.30 (n/a)	24542.14 (n/a)	24551.90 (n/a)	24495.50 (n/a)	37.80 (n/a)

gemm_2048x2048x2048_64x64x64_2_cols_0_bcolmaj_0_ccolmaj_0_0

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)	Throughput (max)	Throughput (mean)	Throughput (median)	Throughput (min)	Throughput (stddev)
`a4b6ffe` — 2026-01-05 19:24:05	1.04 (+0.28%)	1.03 (-0.15%)	1.03 (+0.05%)	1.02 (-0.52%)	0.01 (+42.72%)	24719.80 (+0.52%)	24512.42 (+0.16%)	24529.20 (-0.05%)	24092.60 (-0.28%)	252.92 (+43.15%)	713.08 (+0.28%)	700.92 (-0.15%)	700.38 (+0.05%)	694.98 (-0.52%)	7.30 (+42.71%)
`50ad37c` — 2025-12-29 20:35:53	1.04 (n/a)	1.03 (n/a)	1.03 (n/a)	1.02 (n/a)	0.01 (n/a)	24592.30 (n/a)	24473.70 (n/a)	24540.90 (n/a)	24160.80 (n/a)	176.68 (n/a)	711.06 (n/a)	702.00 (n/a)	700.05 (n/a)	698.59 (n/a)	5.12 (n/a)

gemm_2048x2048x2048_64x64x64_2_cols_0_bcolmaj_1_ccolmaj_0

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`130b6ea` — 2025-12-05 21:15:49	0.99 (+1.69%)	0.97 (+0.29%)	0.97 (-0.04%)	0.97 (-0.15%)	0.01 (+434.92%)	26025.30 (+0.15%)	25839.34 (-0.29%)	25916.70 (+0.04%)	25446.00 (-1.67%)	202.05 (+425.44%)
`0a6c11c` — 2025-12-04 00:39:10	0.97 (n/a)	0.97 (n/a)	0.97 (n/a)	0.97 (n/a)	0.00 (n/a)	25986.90 (n/a)	25913.52 (n/a)	25905.70 (n/a)	25877.00 (n/a)	38.45 (n/a)

gemm_2048x2048x2048_64x64x64_2_cols_0_bcolmaj_1_ccolmaj_0_0

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)	Throughput (max)	Throughput (mean)	Throughput (median)	Throughput (min)	Throughput (stddev)
`a4b6ffe` — 2026-01-05 19:24:05	0.97 (-0.01%)	0.97 (+0.03%)	0.97 (+0.04%)	0.97 (+0.00%)	0.00 (-6.43%)	25927.90 (-0.00%)	25903.20 (-0.03%)	25898.60 (-0.04%)	25892.80 (+0.01%)	14.05 (-6.56%)	663.50 (-0.01%)	663.23 (+0.03%)	663.35 (+0.04%)	662.60 (+0.00%)	0.36 (-6.43%)
`50ad37c` — 2025-12-29 20:35:53	0.97 (n/a)	0.97 (n/a)	0.97 (n/a)	0.97 (n/a)	0.00 (n/a)	25928.10 (n/a)	25910.14 (n/a)	25908.10 (n/a)	25891.20 (n/a)	15.04 (n/a)	663.54 (n/a)	663.06 (n/a)	663.11 (n/a)	662.60 (n/a)	0.38 (n/a)

gemm_2048x2048x2048_64x64x64_2_cols_1_bcolmaj_0_ccolmaj_0

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`130b6ea` — 2025-12-05 21:15:49	0.92 (-1.02%)	0.92 (-0.66%)	0.92 (-0.10%)	0.91 (-0.78%)	0.01 (-4.54%)	27751.40 (+0.79%)	27480.74 (+0.66%)	27354.80 (+0.10%)	27320.50 (+1.03%)	182.69 (-2.84%)
`0a6c11c` — 2025-12-04 00:39:10	0.93 (n/a)	0.92 (n/a)	0.92 (n/a)	0.91 (n/a)	0.01 (n/a)	27534.80 (n/a)	27299.32 (n/a)	27326.70 (n/a)	27042.30 (n/a)	188.04 (n/a)

gemm_2048x2048x2048_64x64x64_2_cols_1_bcolmaj_0_ccolmaj_0_0

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)	Throughput (max)	Throughput (mean)	Throughput (median)	Throughput (min)	Throughput (stddev)
`a4b6ffe` — 2026-01-05 19:24:05	0.92 (+0.11%)	0.91 (-0.12%)	0.91 (-0.21%)	0.91 (-0.24%)	0.00 (+30.34%)	27723.50 (+0.24%)	27548.94 (+0.12%)	27522.90 (+0.21%)	27403.60 (-0.11%)	124.26 (+30.58%)	626.92 (+0.11%)	623.62 (-0.12%)	624.20 (-0.21%)	619.69 (-0.24%)	2.81 (+30.34%)
`50ad37c` — 2025-12-29 20:35:53	0.92 (n/a)	0.91 (n/a)	0.92 (n/a)	0.91 (n/a)	0.00 (n/a)	27657.10 (n/a)	27515.58 (n/a)	27464.70 (n/a)	27434.60 (n/a)	95.16 (n/a)	626.21 (n/a)	624.37 (n/a)	625.52 (n/a)	621.17 (n/a)	2.16 (n/a)

gemm_2048x2048x2048_64x64x64_2cols_bcolmaj

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)	Throughput (max)	Throughput (mean)	Throughput (median)	Throughput (min)	Throughput (stddev)
`1531781` — 2026-01-23 18:49:24	0.93 (+1.26%)	0.91 (+0.29%)	0.91 (+0.24%)	0.90 (-0.34%)	0.01 (+132.37%)	27826.70 (+0.34%)	27533.84 (-0.28%)	27519.90 (-0.24%)	27105.20 (-1.24%)	272.35 (+129.91%)	633.82 (+1.26%)	624.00 (+0.29%)	624.27 (+0.24%)	617.39 (-0.34%)	6.21 (+132.36%)
`0f0da0b` — 2026-01-14 16:31:00	0.92 (n/a)	0.91 (n/a)	0.91 (n/a)	0.91 (n/a)	0.00 (n/a)	27731.40 (n/a)	27612.36 (n/a)	27585.70 (n/a)	27446.50 (n/a)	118.46 (n/a)	625.94 (n/a)	622.19 (n/a)	622.78 (n/a)	619.51 (n/a)	2.67 (n/a)

gemm_2048x2048x2048_64x64x64_8_cols_0_bcolmaj_0_ccolmaj_0

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	GFLOP/s (max)	GFLOP/s (mean)	GFLOP/s (median)	GFLOP/s (min)	GFLOP/s (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`31ab24e` — 2025-11-22 02:01:56	3.65 (-1.06%)	3.59 (+0.83%)	3.63 (-0.26%)	3.45 (+2.66%)	0.07 (-44.71%)	2678.08 (-1.06%)	2632.13 (+0.83%)	2658.60 (-0.26%)	2528.31 (+2.66%)	53.86 (-44.71%)	6795.00 (-2.59%)	6529.80 (-0.92%)	6462.00 (+0.26%)	6415.00 (+1.07%)	137.14 (-45.27%)
`d9a64c5` — 2025-11-21 20:49:35	3.69 (n/a)	3.56 (n/a)	3.64 (n/a)	3.36 (n/a)	0.13 (n/a)	2706.77 (n/a)	2610.50 (n/a)	2665.61 (n/a)	2462.71 (n/a)	97.41 (n/a)	6976.00 (n/a)	6590.40 (n/a)	6445.00 (n/a)	6347.00 (n/a)	250.58 (n/a)

gemm_2048x2048x2048_64x64x64_8_cols_0_bcolmaj_0_ccolmaj_0_0

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)	Throughput (max)	Throughput (mean)	Throughput (median)	Throughput (min)	Throughput (stddev)
`d48746f` — 2026-01-12 22:17:10	3.98 (+0.20%)	3.84 (-2.95%)	3.76 (-5.27%)	3.72 (-4.97%)	0.13 (+473.97%)	6760.00 (+5.23%)	6559.50 (+3.13%)	6697.10 (+5.56%)	6321.00 (-0.20%)	218.68 (+499.40%)	2717.91 (+0.20%)	2621.44 (-2.95%)	2565.27 (-5.27%)	2541.40 (-4.97%)	88.37 (+473.97%)
`331dcca` — 2026-01-08 18:34:10	3.97 (n/a)	3.96 (n/a)	3.97 (n/a)	3.92 (n/a)	0.02 (n/a)	6423.80 (n/a)	6360.32 (n/a)	6344.30 (n/a)	6333.90 (n/a)	36.48 (n/a)	2712.39 (n/a)	2701.17 (n/a)	2707.91 (n/a)	2674.40 (n/a)	15.40 (n/a)

gemm_2048x2048x2048_64x64x64_8_cols_0_bcolmaj_1_ccolmaj_0_0

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)	Throughput (max)	Throughput (mean)	Throughput (median)	Throughput (min)	Throughput (stddev)
`d48746f` — 2026-01-12 22:17:10	3.75 (+0.28%)	3.70 (+1.25%)	3.71 (-0.21%)	3.65 (+4.77%)	0.04 (-66.90%)	6901.80 (-4.56%)	6795.46 (-1.30%)	6790.20 (+0.21%)	6715.90 (-0.28%)	67.05 (-68.40%)	2558.08 (+0.28%)	2528.33 (+1.25%)	2530.10 (-0.21%)	2489.18 (+4.77%)	24.81 (-66.90%)
`331dcca` — 2026-01-08 18:34:10	3.74 (n/a)	3.66 (n/a)	3.71 (n/a)	3.48 (n/a)	0.11 (n/a)	7231.20 (n/a)	6885.18 (n/a)	6775.70 (n/a)	6734.90 (n/a)	212.21 (n/a)	2550.87 (n/a)	2497.04 (n/a)	2535.50 (n/a)	2375.78 (n/a)	74.96 (n/a)

gemm_2048x2048x2048_64x64x64_8_cols_1_bcolmaj_0_ccolmaj_0_0

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)	Throughput (max)	Throughput (mean)	Throughput (median)	Throughput (min)	Throughput (stddev)
`d48746f` — 2026-01-12 22:17:10	3.62 (-0.77%)	3.54 (+0.14%)	3.60 (-0.04%)	3.44 (+2.17%)	0.09 (-25.77%)	7307.40 (-2.13%)	7103.46 (-0.18%)	6982.50 (+0.04%)	6948.20 (+0.77%)	182.95 (-26.55%)	2472.55 (-0.77%)	2419.80 (+0.14%)	2460.41 (-0.04%)	2351.04 (+2.17%)	61.76 (-25.77%)
`331dcca` — 2026-01-08 18:34:10	3.65 (n/a)	3.54 (n/a)	3.61 (n/a)	3.37 (n/a)	0.12 (n/a)	7466.20 (n/a)	7116.40 (n/a)	6979.60 (n/a)	6895.10 (n/a)	249.09 (n/a)	2491.61 (n/a)	2416.45 (n/a)	2461.45 (n/a)	2301.00 (n/a)	83.20 (n/a)

gemm_2048x2048x2048_64x64x64_8cols_bcolmaj_ccolmaj

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)	Throughput (max)	Throughput (mean)	Throughput (median)	Throughput (min)	Throughput (stddev)
`1531781` — 2026-01-23 18:49:24	3.46 (+1.18%)	3.42 (+2.60%)	3.45 (+2.90%)	3.29 (+2.92%)	0.07 (-25.59%)	7651.20 (-2.83%)	7368.34 (-2.57%)	7300.80 (-2.82%)	7263.70 (-1.17%)	162.53 (-28.22%)	2365.17 (+1.18%)	2332.46 (+2.60%)	2353.14 (+2.90%)	2245.39 (+2.92%)	50.15 (-25.59%)
`0f0da0b` — 2026-01-14 16:31:00	3.42 (n/a)	3.33 (n/a)	3.35 (n/a)	3.20 (n/a)	0.10 (n/a)	7874.40 (n/a)	7562.58 (n/a)	7512.80 (n/a)	7349.50 (n/a)	226.42 (n/a)	2337.56 (n/a)	2273.31 (n/a)	2286.73 (n/a)	2181.74 (n/a)	67.40 (n/a)

gemm_384x1536x1792_32x48x64_4cols_bcolmaj

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)	Throughput (max)	Throughput (mean)	Throughput (median)	Throughput (min)	Throughput (stddev)
`1531781` — 2026-01-23 18:49:24	4.35 (-0.80%)	4.24 (+3.81%)	4.33 (+9.58%)	3.85 (-0.98%)	0.22 (-4.90%)	2094.90 (+0.99%)	1905.64 (-3.69%)	1862.00 (-8.75%)	1852.00 (+0.81%)	105.88 (-3.06%)	1141.45 (-0.80%)	1111.86 (+3.81%)	1135.29 (+9.58%)	1009.08 (-0.98%)	57.51 (-4.90%)
`0f0da0b` — 2026-01-14 16:31:00	4.39 (n/a)	4.08 (n/a)	3.95 (n/a)	3.89 (n/a)	0.23 (n/a)	2074.30 (n/a)	1978.70 (n/a)	2040.50 (n/a)	1837.10 (n/a)	109.22 (n/a)	1150.67 (n/a)	1071.02 (n/a)	1036.00 (n/a)	1019.11 (n/a)	60.48 (n/a)

gemm_896x1792x640_32x64x80_8cols_ccolmaj

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)	Throughput (max)	Throughput (mean)	Throughput (median)	Throughput (min)	Throughput (stddev)
`1531781` — 2026-01-23 18:49:24	6.80 (+27.89%)	6.43 (+24.70%)	6.70 (+30.77%)	5.33 (+5.74%)	0.62 (+474.69%)	1247.10 (-5.44%)	1042.70 (-19.16%)	992.90 (-23.53%)	978.30 (-21.81%)	114.59 (+329.89%)	2100.72 (+27.89%)	1987.76 (+24.70%)	2069.85 (+30.77%)	1647.93 (+5.74%)	190.83 (+474.69%)
`0f0da0b` — 2026-01-14 16:31:00	5.32 (n/a)	5.16 (n/a)	5.12 (n/a)	5.04 (n/a)	0.11 (n/a)	1318.80 (n/a)	1289.76 (n/a)	1298.50 (n/a)	1251.20 (n/a)	26.66 (n/a)	1642.57 (n/a)	1594.05 (n/a)	1582.78 (n/a)	1558.44 (n/a)	33.21 (n/a)

layer_norm_1_cols_1_channels_2048_tile_2048

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`1531781` — 2026-01-23 18:49:24	0.11 (+88.61%)	0.09 (+61.18%)	0.08 (+46.00%)	0.07 (+61.81%)	0.02 (+189.47%)	119.70 (-38.20%)	97.44 (-36.14%)	100.50 (-31.54%)	71.50 (-46.96%)	22.67 (-5.36%)
`0f0da0b` — 2026-01-14 16:31:00	0.06 (n/a)	0.05 (n/a)	0.06 (n/a)	0.04 (n/a)	0.01 (n/a)	193.70 (n/a)	152.58 (n/a)	146.80 (n/a)	134.80 (n/a)	23.96 (n/a)

layer_norm_1_cols_2_channels_2048_tile_1024

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`1531781` — 2026-01-23 18:49:24	0.11 (+73.83%)	0.09 (+50.07%)	0.08 (+46.18%)	0.06 (+7.28%)	0.02 (+683.83%)	138.10 (-6.82%)	98.78 (-30.35%)	97.10 (-31.57%)	76.40 (-42.47%)	24.88 (+313.86%)
`0f0da0b` — 2026-01-14 16:31:00	0.06 (n/a)	0.06 (n/a)	0.06 (n/a)	0.06 (n/a)	0.00 (n/a)	148.20 (n/a)	141.82 (n/a)	141.90 (n/a)	132.80 (n/a)	6.01 (n/a)

layer_norm_2_cols_1_channels_2048_tile_1024

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`1531781` — 2026-01-23 18:49:24	0.12 (+58.00%)	0.09 (+35.21%)	0.09 (+33.88%)	0.06 (+12.88%)	0.02 (+142.40%)	140.10 (-11.44%)	98.38 (-23.32%)	94.40 (-25.26%)	68.10 (-36.71%)	26.11 (+36.70%)
`0f0da0b` — 2026-01-14 16:31:00	0.08 (n/a)	0.06 (n/a)	0.06 (n/a)	0.05 (n/a)	0.01 (n/a)	158.20 (n/a)	128.30 (n/a)	126.30 (n/a)	107.60 (n/a)	19.10 (n/a)

layer_norm_2_cols_2_channels_2048_tile_512

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`1531781` — 2026-01-23 18:49:24	0.13 (+115.82%)	0.09 (+79.97%)	0.09 (+57.15%)	0.06 (+84.94%)	0.03 (+142.79%)	130.40 (-45.94%)	91.90 (-43.54%)	93.50 (-36.39%)	62.00 (-53.66%)	25.59 (-42.10%)
`0f0da0b` — 2026-01-14 16:31:00	0.06 (n/a)	0.05 (n/a)	0.06 (n/a)	0.03 (n/a)	0.01 (n/a)	241.20 (n/a)	162.78 (n/a)	147.00 (n/a)	133.80 (n/a)	44.20 (n/a)

layer_norm_4_cols_1_channels_2048_tile_512

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`1531781` — 2026-01-23 18:49:24	0.11 (+63.95%)	0.09 (+43.09%)	0.11 (+55.71%)	0.06 (-6.72%)	0.02 (+806.48%)	141.00 (+7.22%)	91.54 (-25.97%)	77.90 (-35.78%)	73.00 (-39.07%)	28.43 (+505.85%)
`0f0da0b` — 2026-01-14 16:31:00	0.07 (n/a)	0.07 (n/a)	0.07 (n/a)	0.06 (n/a)	0.00 (n/a)	131.50 (n/a)	123.66 (n/a)	121.30 (n/a)	119.80 (n/a)	4.69 (n/a)

layer_norm_4_cols_2_channels_2048_tile_256

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`1531781` — 2026-01-23 18:49:24	0.11 (+76.69%)	0.09 (+61.27%)	0.09 (+70.26%)	0.06 (+36.41%)	0.02 (+195.74%)	148.40 (-26.72%)	103.14 (-34.74%)	87.80 (-41.27%)	73.70 (-43.44%)	32.30 (+19.79%)
`0f0da0b` — 2026-01-14 16:31:00	0.06 (n/a)	0.05 (n/a)	0.05 (n/a)	0.04 (n/a)	0.01 (n/a)	202.50 (n/a)	158.04 (n/a)	149.50 (n/a)	130.30 (n/a)	26.96 (n/a)

layer_norm_8_cols_1_channels_2048_tile_256

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`1531781` — 2026-01-23 18:49:24	0.12 (+75.09%)	0.10 (+78.44%)	0.11 (+94.76%)	0.05 (+12.92%)	0.03 (+213.89%)	153.90 (-11.45%)	90.36 (-39.81%)	77.40 (-48.67%)	67.30 (-42.87%)	35.86 (+72.31%)
`0f0da0b` — 2026-01-14 16:31:00	0.07 (n/a)	0.06 (n/a)	0.05 (n/a)	0.05 (n/a)	0.01 (n/a)	173.80 (n/a)	150.12 (n/a)	150.80 (n/a)	117.80 (n/a)	20.81 (n/a)

layer_norm_8_cols_2_channels_2048_tile_128

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`1531781` — 2026-01-23 18:49:24	0.09 (+88.05%)	0.08 (+72.40%)	0.08 (+83.22%)	0.06 (+32.78%)	0.01 (+723.36%)	140.30 (-24.69%)	106.50 (-40.67%)	97.60 (-45.44%)	92.40 (-46.80%)	19.84 (+234.33%)
`0f0da0b` — 2026-01-14 16:31:00	0.05 (n/a)	0.05 (n/a)	0.05 (n/a)	0.04 (n/a)	0.00 (n/a)	186.30 (n/a)	179.50 (n/a)	178.90 (n/a)	173.70 (n/a)	5.93 (n/a)

matrix_vector_mul_128x128_32_1col

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)
`130b6ea` — 2025-12-05 21:15:49	0.37 (+17.13%)	0.30 (+24.87%)	0.30 (+38.03%)	0.23 (+12.47%)	0.06 (+35.23%)
`0a6c11c` — 2025-12-04 00:39:10	0.32 (n/a)	0.24 (n/a)	0.22 (n/a)	0.20 (n/a)	0.04 (n/a)

matrix_vector_mul_128x128_32_1col0

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Throughput (max)	Throughput (mean)	Throughput (median)	Throughput (min)	Throughput (stddev)
`331dcca` — 2026-01-08 18:34:10	0.26 (-1.81%)	0.24 (+4.06%)	0.24 (+0.52%)	0.21 (+16.97%)	0.02 (-48.16%)	0.26 (-1.81%)	0.24 (+4.06%)	0.24 (+0.52%)	0.20 (+16.97%)	0.02 (-48.16%)
`a4b6ffe` — 2026-01-05 19:24:05	0.26 (n/a)	0.23 (n/a)	0.24 (n/a)	0.18 (n/a)	0.04 (n/a)	0.26 (n/a)	0.23 (n/a)	0.24 (n/a)	0.17 (n/a)	0.04 (n/a)

matrix_vector_mul_128x128_32tsi_128tso_1col0

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Throughput (max)	Throughput (mean)	Throughput (median)	Throughput (min)	Throughput (stddev)
`1531781` — 2026-01-23 18:49:24	0.45 (+63.20%)	0.35 (+43.38%)	0.35 (+38.29%)	0.22 (+12.34%)	0.08 (+156.17%)	0.44 (+63.20%)	0.34 (+43.38%)	0.34 (+38.29%)	0.22 (+12.34%)	0.08 (+156.17%)
`0f0da0b` — 2026-01-14 16:31:00	0.27 (n/a)	0.24 (n/a)	0.25 (n/a)	0.20 (n/a)	0.03 (n/a)	0.27 (n/a)	0.24 (n/a)	0.25 (n/a)	0.19 (n/a)	0.03 (n/a)

matrix_vector_mul_2048x8192_1_1col

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)
`130b6ea` — 2025-12-05 21:15:49	13.57 (+3.47%)	13.15 (+1.09%)	13.10 (+0.16%)	12.52 (-1.58%)	0.37 (+153.19%)
`0a6c11c` — 2025-12-04 00:39:10	13.11 (n/a)	13.01 (n/a)	13.07 (n/a)	12.72 (n/a)	0.15 (n/a)

matrix_vector_mul_2048x8192_1_1col0

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Throughput (max)	Throughput (mean)	Throughput (median)	Throughput (min)	Throughput (stddev)
`331dcca` — 2026-01-08 18:34:10	13.43 (-0.07%)	13.30 (+4.62%)	13.27 (+1.29%)	13.16 (+16.70%)	0.12 (-86.69%)	13.42 (-0.07%)	13.29 (+4.62%)	13.26 (+1.29%)	13.16 (+16.70%)	0.12 (-86.69%)
`a4b6ffe` — 2026-01-05 19:24:05	13.44 (n/a)	12.71 (n/a)	13.10 (n/a)	11.28 (n/a)	0.87 (n/a)	13.43 (n/a)	12.70 (n/a)	13.09 (n/a)	11.27 (n/a)	0.87 (n/a)

matrix_vector_mul_2048x8192_1_2col

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)
`130b6ea` — 2025-12-05 21:15:49	25.06 (+2.47%)	23.71 (+1.81%)	24.13 (+0.50%)	20.90 (+5.41%)	1.48 (-15.60%)
`0a6c11c` — 2025-12-04 00:39:10	24.46 (n/a)	23.29 (n/a)	24.01 (n/a)	19.83 (n/a)	1.75 (n/a)

matrix_vector_mul_2048x8192_1_2col0

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Throughput (max)	Throughput (mean)	Throughput (median)	Throughput (min)	Throughput (stddev)
`331dcca` — 2026-01-08 18:34:10	24.46 (-0.60%)	21.07 (-8.03%)	19.75 (-17.83%)	18.08 (+1.18%)	3.03 (+7.07%)	24.44 (-0.60%)	21.06 (-8.03%)	19.74 (-17.83%)	18.07 (+1.18%)	3.03 (+7.07%)
`a4b6ffe` — 2026-01-05 19:24:05	24.60 (n/a)	22.91 (n/a)	24.04 (n/a)	17.87 (n/a)	2.83 (n/a)	24.59 (n/a)	22.90 (n/a)	24.02 (n/a)	17.86 (n/a)	2.83 (n/a)

matrix_vector_mul_2048x8192_1_4col

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)
`130b6ea` — 2025-12-05 21:15:49	42.05 (+1.54%)	40.50 (+12.60%)	40.49 (+2.36%)	38.98 (+97.53%)	0.97 (-88.09%)
`0a6c11c` — 2025-12-04 00:39:10	41.41 (n/a)	35.97 (n/a)	39.55 (n/a)	19.73 (n/a)	8.15 (n/a)

matrix_vector_mul_2048x8192_1_4col0

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Throughput (max)	Throughput (mean)	Throughput (median)	Throughput (min)	Throughput (stddev)
`331dcca` — 2026-01-08 18:34:10	42.58 (+1.51%)	41.37 (+14.29%)	41.25 (+4.89%)	40.33 (+91.42%)	0.92 (-89.18%)	42.56 (+1.51%)	41.34 (+14.29%)	41.23 (+4.89%)	40.30 (+91.42%)	0.92 (-89.18%)
`a4b6ffe` — 2026-01-05 19:24:05	41.95 (n/a)	36.20 (n/a)	39.33 (n/a)	21.07 (n/a)	8.54 (n/a)	41.92 (n/a)	36.17 (n/a)	39.31 (n/a)	21.05 (n/a)	8.53 (n/a)

matrix_vector_mul_2048x8192_1_8col

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)
`130b6ea` — 2025-12-05 21:15:49	46.62 (+3.11%)	44.82 (+0.17%)	44.78 (+0.17%)	42.60 (-3.80%)	1.60 (+367.72%)
`0a6c11c` — 2025-12-04 00:39:10	45.21 (n/a)	44.75 (n/a)	44.70 (n/a)	44.29 (n/a)	0.34 (n/a)

matrix_vector_mul_2048x8192_1_8col0

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Throughput (max)	Throughput (mean)	Throughput (median)	Throughput (min)	Throughput (stddev)
`331dcca` — 2026-01-08 18:34:10	46.04 (+4.14%)	45.06 (+3.47%)	44.55 (+2.76%)	44.30 (+2.84%)	0.87 (+66.58%)	46.01 (+4.14%)	45.03 (+3.47%)	44.52 (+2.76%)	44.27 (+2.84%)	0.87 (+66.58%)
`a4b6ffe` — 2026-01-05 19:24:05	44.21 (n/a)	43.55 (n/a)	43.35 (n/a)	43.07 (n/a)	0.52 (n/a)	44.18 (n/a)	43.52 (n/a)	43.32 (n/a)	43.05 (n/a)	0.52 (n/a)

matrix_vector_mul_2048x8192_1tsi_1024tso_2col0

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Throughput (max)	Throughput (mean)	Throughput (median)	Throughput (min)	Throughput (stddev)
`1531781` — 2026-01-23 18:49:24	24.80 (+0.16%)	24.59 (-0.08%)	24.55 (-0.22%)	24.40 (+0.04%)	0.18 (+11.75%)	24.79 (+0.16%)	24.58 (-0.08%)	24.53 (-0.22%)	24.38 (+0.04%)	0.18 (+11.75%)
`0f0da0b` — 2026-01-14 16:31:00	24.76 (n/a)	24.61 (n/a)	24.60 (n/a)	24.39 (n/a)	0.16 (n/a)	24.75 (n/a)	24.60 (n/a)	24.58 (n/a)	24.37 (n/a)	0.16 (n/a)

matrix_vector_mul_2048x8192_1tsi_2048tso_1col0

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Throughput (max)	Throughput (mean)	Throughput (median)	Throughput (min)	Throughput (stddev)
`1531781` — 2026-01-23 18:49:24	13.54 (+0.77%)	13.34 (+3.03%)	13.45 (+0.75%)	13.10 (+15.45%)	0.21 (-77.02%)	13.53 (+0.77%)	13.33 (+3.03%)	13.44 (+0.75%)	13.09 (+15.45%)	0.21 (-77.02%)
`0f0da0b` — 2026-01-14 16:31:00	13.44 (n/a)	12.95 (n/a)	13.35 (n/a)	11.35 (n/a)	0.90 (n/a)	13.43 (n/a)	12.94 (n/a)	13.34 (n/a)	11.34 (n/a)	0.90 (n/a)

matrix_vector_mul_2048x8192_1tsi_256tso_8col0

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Throughput (max)	Throughput (mean)	Throughput (median)	Throughput (min)	Throughput (stddev)
`1531781` — 2026-01-23 18:49:24	46.81 (+1.97%)	45.91 (+2.21%)	46.38 (+2.98%)	44.81 (+2.36%)	0.89 (-7.11%)	46.79 (+1.97%)	45.88 (+2.21%)	46.35 (+2.98%)	44.79 (+2.36%)	0.89 (-7.11%)
`0f0da0b` — 2026-01-14 16:31:00	45.91 (n/a)	44.92 (n/a)	45.04 (n/a)	43.78 (n/a)	0.96 (n/a)	45.88 (n/a)	44.89 (n/a)	45.01 (n/a)	43.75 (n/a)	0.96 (n/a)

matrix_vector_mul_2048x8192_1tsi_512tso_4col0

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Throughput (max)	Throughput (mean)	Throughput (median)	Throughput (min)	Throughput (stddev)
`1531781` — 2026-01-23 18:49:24	41.49 (-0.92%)	40.81 (+0.09%)	41.13 (+0.38%)	39.95 (+0.86%)	0.65 (-24.40%)	41.46 (-0.92%)	40.79 (+0.09%)	41.10 (+0.38%)	39.93 (+0.86%)	0.65 (-24.40%)
`0f0da0b` — 2026-01-14 16:31:00	41.88 (n/a)	40.78 (n/a)	40.97 (n/a)	39.61 (n/a)	0.86 (n/a)	41.85 (n/a)	40.75 (n/a)	40.95 (n/a)	39.59 (n/a)	0.86 (n/a)

matrix_vector_mul_8192x2048_4_1col

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)
`130b6ea` — 2025-12-05 21:15:49	13.29 (+1.63%)	13.19 (+6.08%)	13.19 (+1.17%)	13.09 (+26.86%)	0.08 (-92.94%)
`0a6c11c` — 2025-12-04 00:39:10	13.08 (n/a)	12.44 (n/a)	13.03 (n/a)	10.32 (n/a)	1.07 (n/a)

matrix_vector_mul_8192x2048_4_1col0

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Throughput (max)	Throughput (mean)	Throughput (median)	Throughput (min)	Throughput (stddev)
`331dcca` — 2026-01-08 18:34:10	13.68 (+3.42%)	13.30 (+8.54%)	13.38 (+7.25%)	12.72 (+14.96%)	0.35 (-66.09%)	13.67 (+3.42%)	13.29 (+8.54%)	13.38 (+7.25%)	12.71 (+14.96%)	0.35 (-66.09%)
`a4b6ffe` — 2026-01-05 19:24:05	13.23 (n/a)	12.25 (n/a)	12.48 (n/a)	11.07 (n/a)	1.04 (n/a)	13.22 (n/a)	12.24 (n/a)	12.47 (n/a)	11.06 (n/a)	1.04 (n/a)

matrix_vector_mul_8192x2048_4_2col

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)
`130b6ea` — 2025-12-05 21:15:49	25.58 (+4.24%)	24.23 (+6.59%)	24.03 (+0.29%)	23.58 (+37.29%)	0.70 (-74.97%)
`0a6c11c` — 2025-12-04 00:39:10	24.54 (n/a)	22.73 (n/a)	23.96 (n/a)	17.18 (n/a)	2.79 (n/a)

matrix_vector_mul_8192x2048_4_2col0

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Throughput (max)	Throughput (mean)	Throughput (median)	Throughput (min)	Throughput (stddev)
`331dcca` — 2026-01-08 18:34:10	25.08 (+1.32%)	24.68 (+13.42%)	24.61 (+3.26%)	24.47 (+48.32%)	0.24 (-93.56%)	25.07 (+1.32%)	24.67 (+13.42%)	24.59 (+3.26%)	24.46 (+48.32%)	0.24 (-93.56%)
`a4b6ffe` — 2026-01-05 19:24:05	24.76 (n/a)	21.76 (n/a)	23.83 (n/a)	16.50 (n/a)	3.74 (n/a)	24.74 (n/a)	21.75 (n/a)	23.82 (n/a)	16.49 (n/a)	3.73 (n/a)

matrix_vector_mul_8192x2048_4_4col

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)
`130b6ea` — 2025-12-05 21:15:49	40.94 (-2.19%)	39.74 (-1.10%)	39.76 (-0.60%)	38.17 (-2.14%)	1.04 (+4.39%)
`0a6c11c` — 2025-12-04 00:39:10	41.86 (n/a)	40.18 (n/a)	40.00 (n/a)	39.00 (n/a)	1.00 (n/a)

matrix_vector_mul_8192x2048_4_4col0

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Throughput (max)	Throughput (mean)	Throughput (median)	Throughput (min)	Throughput (stddev)
`331dcca` — 2026-01-08 18:34:10	40.74 (-1.09%)	37.16 (-7.15%)	40.25 (+1.47%)	24.51 (-37.44%)	7.07 (+736.13%)	40.72 (-1.09%)	37.13 (-7.15%)	40.23 (+1.47%)	24.50 (-37.44%)	7.07 (+736.13%)
`a4b6ffe` — 2026-01-05 19:24:05	41.19 (n/a)	40.02 (n/a)	39.67 (n/a)	39.18 (n/a)	0.85 (n/a)	41.17 (n/a)	39.99 (n/a)	39.64 (n/a)	39.16 (n/a)	0.85 (n/a)

matrix_vector_mul_8192x2048_4_8col

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)
`130b6ea` — 2025-12-05 21:15:49	46.71 (-1.04%)	44.91 (-2.34%)	44.71 (-2.98%)	43.80 (-1.44%)	0.96 (+6.93%)
`0a6c11c` — 2025-12-04 00:39:10	47.20 (n/a)	45.98 (n/a)	46.08 (n/a)	44.44 (n/a)	0.90 (n/a)

matrix_vector_mul_8192x2048_4_8col0

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Throughput (max)	Throughput (mean)	Throughput (median)	Throughput (min)	Throughput (stddev)
`331dcca` — 2026-01-08 18:34:10	46.28 (+5.68%)	44.89 (+14.59%)	44.74 (+4.26%)	44.20 (+62.80%)	0.84 (-87.96%)	46.25 (+5.68%)	44.86 (+14.59%)	44.71 (+4.26%)	44.17 (+62.80%)	0.84 (-87.96%)
`a4b6ffe` — 2026-01-05 19:24:05	43.79 (n/a)	39.17 (n/a)	42.91 (n/a)	27.15 (n/a)	6.98 (n/a)	43.77 (n/a)	39.15 (n/a)	42.88 (n/a)	27.13 (n/a)	6.98 (n/a)

matrix_vector_mul_8192x2048_4tsi_1024tso_1col0

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Throughput (max)	Throughput (mean)	Throughput (median)	Throughput (min)	Throughput (stddev)
`1531781` — 2026-01-23 18:49:24	13.46 (+0.53%)	13.33 (+2.85%)	13.32 (+1.17%)	13.20 (+10.91%)	0.11 (-81.73%)	13.45 (+0.53%)	13.32 (+2.85%)	13.31 (+1.17%)	13.19 (+10.91%)	0.11 (-81.73%)
`0f0da0b` — 2026-01-14 16:31:00	13.39 (n/a)	12.96 (n/a)	13.17 (n/a)	11.90 (n/a)	0.62 (n/a)	13.38 (n/a)	12.95 (n/a)	13.16 (n/a)	11.89 (n/a)	0.62 (n/a)

matrix_vector_mul_8192x2048_4tsi_1024tso_2col0

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Throughput (max)	Throughput (mean)	Throughput (median)	Throughput (min)	Throughput (stddev)
`1531781` — 2026-01-23 18:49:24	25.27 (+0.55%)	24.75 (-0.75%)	24.61 (-1.28%)	24.55 (-0.37%)	0.30 (+52.22%)	25.25 (+0.55%)	24.74 (-0.75%)	24.59 (-1.28%)	24.53 (-0.37%)	0.30 (+52.22%)
`0f0da0b` — 2026-01-14 16:31:00	25.13 (n/a)	24.94 (n/a)	24.93 (n/a)	24.64 (n/a)	0.20 (n/a)	25.11 (n/a)	24.92 (n/a)	24.91 (n/a)	24.62 (n/a)	0.20 (n/a)

matrix_vector_mul_8192x2048_4tsi_1024tso_4col0

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Throughput (max)	Throughput (mean)	Throughput (median)	Throughput (min)	Throughput (stddev)
`1531781` — 2026-01-23 18:49:24	40.86 (-4.37%)	40.01 (-4.03%)	40.43 (-3.32%)	38.86 (-2.46%)	0.89 (-24.84%)	40.84 (-4.37%)	39.98 (-4.03%)	40.41 (-3.32%)	38.84 (-2.46%)	0.89 (-24.84%)
`0f0da0b` — 2026-01-14 16:31:00	42.73 (n/a)	41.69 (n/a)	41.82 (n/a)	39.85 (n/a)	1.18 (n/a)	42.70 (n/a)	41.66 (n/a)	41.80 (n/a)	39.82 (n/a)	1.18 (n/a)

matrix_vector_mul_8192x2048_4tsi_1024tso_8col0

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Throughput (max)	Throughput (mean)	Throughput (median)	Throughput (min)	Throughput (stddev)
`1531781` — 2026-01-23 18:49:24	46.08 (-1.82%)	45.11 (-1.49%)	45.49 (-0.25%)	43.89 (-1.05%)	0.97 (-6.46%)	46.05 (-1.82%)	45.08 (-1.49%)	45.46 (-0.25%)	43.87 (-1.05%)	0.97 (-6.46%)
`0f0da0b` — 2026-01-14 16:31:00	46.94 (n/a)	45.79 (n/a)	45.60 (n/a)	44.36 (n/a)	1.03 (n/a)	46.91 (n/a)	45.76 (n/a)	45.58 (n/a)	44.33 (n/a)	1.03 (n/a)

mem_copy_16_cores_2_chans_2048_tile_128_False

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`130b6ea` — 2025-12-05 21:15:49	0.07 (+22.03%)	0.06 (+27.86%)	0.06 (+35.49%)	0.05 (+15.00%)	0.01 (+49.69%)	173.30 (-13.00%)	137.94 (-21.32%)	133.60 (-26.19%)	123.40 (-18.06%)	18.40 (+8.53%)
`0a6c11c` — 2025-12-04 00:39:10	0.05 (n/a)	0.05 (n/a)	0.05 (n/a)	0.04 (n/a)	0.00 (n/a)	199.20 (n/a)	175.32 (n/a)	181.00 (n/a)	150.60 (n/a)	16.96 (n/a)

mem_copy_16_cores_2_chans_2048_tile_128_False0

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`1531781` — 2026-01-23 18:49:24	0.10 (+63.55%)	0.08 (+85.69%)	0.09 (+119.34%)	0.04 (+23.15%)	0.02 (+126.16%)	195.30 (-18.79%)	111.62 (-42.82%)	86.60 (-54.42%)	84.60 (-38.87%)	47.65 (+9.68%)
`0f0da0b` — 2026-01-14 16:31:00	0.06 (n/a)	0.04 (n/a)	0.04 (n/a)	0.03 (n/a)	0.01 (n/a)	240.50 (n/a)	195.20 (n/a)	190.00 (n/a)	138.40 (n/a)	43.45 (n/a)

mem_copy_1_cols_1_channels_2048_tile_2048

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`003b49d` — 2025-11-14 19:30:54	0.05 (-16.99%)	0.05 (-12.47%)	0.05 (-7.32%)	0.04 (-11.17%)	0.00 (-29.61%)	197.00 (+12.57%)	171.00 (+13.70%)	164.00 (+7.89%)	153.00 (+20.47%)	17.49 (-4.32%)
`0999a9a` — 2025-11-13 19:24:08	0.06 (n/a)	0.05 (n/a)	0.05 (n/a)	0.04 (n/a)	0.01 (n/a)	175.00 (n/a)	150.40 (n/a)	152.00 (n/a)	127.00 (n/a)	18.28 (n/a)

mem_copy_1_cols_2_channels_2048_tile_1024

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`003b49d` — 2025-11-14 19:30:54	0.06 (+5.56%)	0.05 (-1.59%)	0.05 (-1.39%)	0.03 (-14.07%)	0.01 (+25.03%)	263.00 (+16.37%)	166.00 (+4.53%)	144.00 (+1.41%)	126.00 (-5.26%)	49.38 (+43.17%)
`0999a9a` — 2025-11-13 19:24:08	0.06 (n/a)	0.05 (n/a)	0.05 (n/a)	0.03 (n/a)	0.01 (n/a)	226.00 (n/a)	158.80 (n/a)	142.00 (n/a)	133.00 (n/a)	34.49 (n/a)

mem_copy_1_cores_1_chans_2048_tile_2048_False

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`130b6ea` — 2025-12-05 21:15:49	0.08 (+10.93%)	0.07 (+7.79%)	0.07 (+3.55%)	0.05 (-1.73%)	0.01 (+36.73%)	150.20 (+1.76%)	121.68 (-6.36%)	125.60 (-3.38%)	97.90 (-9.85%)	18.56 (+21.98%)
`0a6c11c` — 2025-12-04 00:39:10	0.08 (n/a)	0.06 (n/a)	0.06 (n/a)	0.06 (n/a)	0.01 (n/a)	147.60 (n/a)	129.94 (n/a)	130.00 (n/a)	108.60 (n/a)	15.21 (n/a)

mem_copy_1_cores_1_chans_2048_tile_2048_False0

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`1531781` — 2026-01-23 18:49:24	0.10 (+31.19%)	0.09 (+26.03%)	0.10 (+33.11%)	0.07 (+25.01%)	0.02 (+78.79%)	118.10 (-20.04%)	96.52 (-19.61%)	84.40 (-24.91%)	82.40 (-23.77%)	18.29 (+9.09%)
`0f0da0b` — 2026-01-14 16:31:00	0.08 (n/a)	0.07 (n/a)	0.07 (n/a)	0.06 (n/a)	0.01 (n/a)	147.70 (n/a)	120.06 (n/a)	112.40 (n/a)	108.10 (n/a)	16.76 (n/a)

mem_copy_2_cols_1_channels_2048_tile_1024

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`003b49d` — 2025-11-14 19:30:54	0.06 (-15.32%)	0.05 (-17.09%)	0.05 (-23.81%)	0.04 (-15.54%)	0.01 (-18.18%)	193.00 (+18.40%)	160.80 (+20.36%)	168.00 (+31.25%)	124.00 (+18.10%)	22.92 (+11.39%)
`0999a9a` — 2025-11-13 19:24:08	0.07 (n/a)	0.06 (n/a)	0.06 (n/a)	0.05 (n/a)	0.01 (n/a)	163.00 (n/a)	133.60 (n/a)	128.00 (n/a)	105.00 (n/a)	20.58 (n/a)

mem_copy_2_cols_2_channels_2048_tile_512

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`003b49d` — 2025-11-14 19:30:54	0.06 (-23.02%)	0.05 (-6.98%)	0.05 (-2.13%)	0.04 (-15.31%)	0.01 (-33.29%)	196.00 (+18.07%)	147.00 (+6.52%)	141.00 (+2.17%)	126.00 (+29.90%)	25.27 (+8.13%)
`0999a9a` — 2025-11-13 19:24:08	0.08 (n/a)	0.06 (n/a)	0.06 (n/a)	0.05 (n/a)	0.01 (n/a)	166.00 (n/a)	138.00 (n/a)	138.00 (n/a)	97.00 (n/a)	23.37 (n/a)

mem_copy_2_cores_1_chans_2048_tile_1024_False

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`130b6ea` — 2025-12-05 21:15:49	0.10 (+31.37%)	0.08 (+29.12%)	0.08 (+29.06%)	0.07 (+41.80%)	0.01 (+0.23%)	109.90 (-29.46%)	98.30 (-23.38%)	100.80 (-22.52%)	79.40 (-23.87%)	10.54 (-46.17%)
`0a6c11c` — 2025-12-04 00:39:10	0.08 (n/a)	0.07 (n/a)	0.06 (n/a)	0.05 (n/a)	0.01 (n/a)	155.80 (n/a)	128.30 (n/a)	130.10 (n/a)	104.30 (n/a)	19.58 (n/a)

mem_copy_2_cores_1_chans_2048_tile_1024_False0

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`1531781` — 2026-01-23 18:49:24	0.12 (+103.55%)	0.09 (+81.99%)	0.09 (+69.86%)	0.07 (+83.95%)	0.02 (+187.42%)	117.10 (-45.66%)	92.20 (-43.53%)	86.90 (-41.12%)	70.40 (-50.84%)	23.32 (-23.43%)
`0f0da0b` — 2026-01-14 16:31:00	0.06 (n/a)	0.05 (n/a)	0.06 (n/a)	0.04 (n/a)	0.01 (n/a)	215.50 (n/a)	163.26 (n/a)	147.60 (n/a)	143.20 (n/a)	30.45 (n/a)

mem_copy_2_cores_2_chans_2048_tile_1024_False

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`130b6ea` — 2025-12-05 21:15:49	0.10 (+25.02%)	0.07 (+10.01%)	0.07 (+7.53%)	0.05 (-1.99%)	0.01 (+77.08%)	150.10 (+2.04%)	114.60 (-7.22%)	111.70 (-6.99%)	83.90 (-20.02%)	21.23 (+43.54%)
`0a6c11c` — 2025-12-04 00:39:10	0.08 (n/a)	0.07 (n/a)	0.07 (n/a)	0.06 (n/a)	0.01 (n/a)	147.10 (n/a)	123.52 (n/a)	120.10 (n/a)	104.90 (n/a)	14.79 (n/a)

mem_copy_2_cores_2_chans_2048_tile_1024_False0

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`1531781` — 2026-01-23 18:49:24	0.10 (+35.25%)	0.08 (+37.83%)	0.09 (+48.91%)	0.06 (+9.50%)	0.02 (+125.91%)	140.20 (-8.66%)	101.66 (-25.39%)	92.10 (-32.82%)	82.90 (-26.11%)	24.23 (+51.17%)
`0f0da0b` — 2026-01-14 16:31:00	0.07 (n/a)	0.06 (n/a)	0.06 (n/a)	0.05 (n/a)	0.01 (n/a)	153.50 (n/a)	136.26 (n/a)	137.10 (n/a)	112.20 (n/a)	16.03 (n/a)

mem_copy_4_cols_1_channels_2048_tile_512

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`003b49d` — 2025-11-14 19:30:54	0.07 (+11.97%)	0.05 (+3.09%)	0.05 (+2.10%)	0.03 (-11.35%)	0.01 (+41.65%)	229.00 (+12.81%)	156.00 (-0.51%)	143.00 (-2.05%)	117.00 (-10.69%)	38.21 (+48.71%)
`0999a9a` — 2025-11-13 19:24:08	0.06 (n/a)	0.05 (n/a)	0.05 (n/a)	0.04 (n/a)	0.01 (n/a)	203.00 (n/a)	156.80 (n/a)	146.00 (n/a)	131.00 (n/a)	25.69 (n/a)

mem_copy_4_cols_2_channels_2048_tile_256

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`003b49d` — 2025-11-14 19:30:54	0.06 (+9.02%)	0.05 (+5.59%)	0.05 (+0.00%)	0.04 (+3.63%)	0.01 (+25.48%)	193.00 (-3.50%)	147.40 (-4.66%)	147.00 (+0.00%)	122.00 (-8.27%)	25.24 (+7.48%)
`0999a9a` — 2025-11-13 19:24:08	0.06 (n/a)	0.05 (n/a)	0.05 (n/a)	0.04 (n/a)	0.01 (n/a)	200.00 (n/a)	154.60 (n/a)	147.00 (n/a)	133.00 (n/a)	23.48 (n/a)

mem_copy_4_cores_1_chans_2048_tile_512_False

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`130b6ea` — 2025-12-05 21:15:49	0.09 (-3.98%)	0.07 (+4.00%)	0.07 (+7.04%)	0.06 (+3.67%)	0.01 (-15.76%)	129.70 (-3.57%)	113.30 (-4.50%)	116.90 (-6.55%)	90.40 (+4.15%)	15.04 (-13.80%)
`0a6c11c` — 2025-12-04 00:39:10	0.09 (n/a)	0.07 (n/a)	0.07 (n/a)	0.06 (n/a)	0.01 (n/a)	134.50 (n/a)	118.64 (n/a)	125.10 (n/a)	86.80 (n/a)	17.45 (n/a)

mem_copy_4_cores_1_chans_2048_tile_512_False0

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`1531781` — 2026-01-23 18:49:24	0.12 (+51.68%)	0.10 (+55.58%)	0.11 (+85.86%)	0.06 (+8.92%)	0.03 (+163.31%)	128.30 (-8.23%)	84.56 (-32.56%)	71.60 (-46.17%)	65.70 (-34.04%)	26.42 (+57.86%)
`0f0da0b` — 2026-01-14 16:31:00	0.08 (n/a)	0.07 (n/a)	0.06 (n/a)	0.06 (n/a)	0.01 (n/a)	139.80 (n/a)	125.38 (n/a)	133.00 (n/a)	99.60 (n/a)	16.73 (n/a)

mem_copy_4_cores_2_chans_2048_tile_512_False

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`130b6ea` — 2025-12-05 21:15:49	0.08 (+9.57%)	0.07 (+16.85%)	0.07 (+19.58%)	0.06 (+24.52%)	0.01 (+1.05%)	134.50 (-19.70%)	116.64 (-14.92%)	116.80 (-16.39%)	98.00 (-8.75%)	15.71 (-25.13%)
`0a6c11c` — 2025-12-04 00:39:10	0.08 (n/a)	0.06 (n/a)	0.06 (n/a)	0.05 (n/a)	0.01 (n/a)	167.50 (n/a)	137.10 (n/a)	139.70 (n/a)	107.40 (n/a)	20.99 (n/a)

mem_copy_4_cores_2_chans_2048_tile_512_False0

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`1531781` — 2026-01-23 18:49:24	0.10 (+21.52%)	0.09 (+51.62%)	0.10 (+79.69%)	0.06 (+23.77%)	0.02 (+25.48%)	133.20 (-19.17%)	92.98 (-33.88%)	83.40 (-44.33%)	82.10 (-17.74%)	22.49 (-15.24%)
`0f0da0b` — 2026-01-14 16:31:00	0.08 (n/a)	0.06 (n/a)	0.05 (n/a)	0.05 (n/a)	0.01 (n/a)	164.80 (n/a)	140.62 (n/a)	149.80 (n/a)	99.80 (n/a)	26.53 (n/a)

mem_copy_8_cols_1_channels_2048_tile_256

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`003b49d` — 2025-11-14 19:30:54	0.05 (-13.48%)	0.05 (-17.79%)	0.05 (-15.33%)	0.03 (-25.09%)	0.01 (+6.67%)	267.00 (+33.50%)	176.80 (+24.51%)	150.00 (+18.11%)	141.00 (+15.57%)	47.41 (+61.41%)
`0999a9a` — 2025-11-13 19:24:08	0.06 (n/a)	0.06 (n/a)	0.06 (n/a)	0.04 (n/a)	0.01 (n/a)	200.00 (n/a)	142.00 (n/a)	127.00 (n/a)	122.00 (n/a)	29.37 (n/a)

mem_copy_8_cols_2_channels_2048_tile_128

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`003b49d` — 2025-11-14 19:30:54	0.05 (+13.38%)	0.05 (+23.85%)	0.05 (+37.91%)	0.05 (+30.72%)	0.00 (-56.85%)	166.00 (-23.50%)	153.40 (-20.60%)	153.00 (-27.49%)	142.00 (-11.80%)	7.63 (-70.60%)
`0999a9a` — 2025-11-13 19:24:08	0.05 (n/a)	0.04 (n/a)	0.04 (n/a)	0.04 (n/a)	0.01 (n/a)	217.00 (n/a)	193.20 (n/a)	211.00 (n/a)	161.00 (n/a)	25.96 (n/a)

mem_copy_8_cores_1_chans_2048_tile_256_False

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`130b6ea` — 2025-12-05 21:15:49	0.09 (+16.31%)	0.08 (+15.34%)	0.08 (+14.60%)	0.07 (+19.30%)	0.01 (-5.41%)	117.30 (-16.15%)	107.80 (-13.61%)	108.50 (-12.78%)	94.10 (-13.99%)	8.29 (-32.01%)
`0a6c11c` — 2025-12-04 00:39:10	0.07 (n/a)	0.07 (n/a)	0.07 (n/a)	0.06 (n/a)	0.01 (n/a)	139.90 (n/a)	124.78 (n/a)	124.40 (n/a)	109.40 (n/a)	12.20 (n/a)

mem_copy_8_cores_1_chans_2048_tile_256_False0

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`1531781` — 2026-01-23 18:49:24	0.09 (+46.69%)	0.08 (+57.18%)	0.09 (+64.02%)	0.06 (+94.45%)	0.01 (+6.58%)	131.30 (-48.59%)	99.78 (-38.77%)	88.70 (-39.00%)	86.90 (-31.84%)	18.90 (-64.18%)
`0f0da0b` — 2026-01-14 16:31:00	0.06 (n/a)	0.05 (n/a)	0.06 (n/a)	0.03 (n/a)	0.01 (n/a)	255.40 (n/a)	162.96 (n/a)	145.40 (n/a)	127.50 (n/a)	52.78 (n/a)

mem_copy_8_cores_2_chans_2048_tile_256_False

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`130b6ea` — 2025-12-05 21:15:49	0.09 (+12.61%)	0.07 (-2.24%)	0.07 (-8.85%)	0.06 (+5.13%)	0.01 (+35.50%)	136.50 (-4.88%)	120.86 (+2.96%)	125.10 (+9.64%)	91.60 (-11.15%)	15.72 (+10.46%)
`0a6c11c` — 2025-12-04 00:39:10	0.08 (n/a)	0.07 (n/a)	0.07 (n/a)	0.06 (n/a)	0.01 (n/a)	143.50 (n/a)	117.38 (n/a)	114.10 (n/a)	103.10 (n/a)	14.23 (n/a)

mem_copy_8_cores_2_chans_2048_tile_256_False0

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`1531781` — 2026-01-23 18:49:24	0.10 (+63.97%)	0.09 (+58.77%)	0.09 (+53.52%)	0.06 (+50.07%)	0.02 (+70.34%)	144.40 (-33.36%)	97.70 (-36.49%)	88.10 (-34.89%)	80.00 (-39.02%)	26.35 (-27.55%)
`0f0da0b` — 2026-01-14 16:31:00	0.06 (n/a)	0.06 (n/a)	0.06 (n/a)	0.04 (n/a)	0.01 (n/a)	216.70 (n/a)	153.84 (n/a)	135.30 (n/a)	131.20 (n/a)	36.36 (n/a)

mha

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`130b6ea` — 2025-12-05 21:15:49	0.18 (+0.16%)	0.18 (+0.19%)	0.18 (+0.12%)	0.18 (+0.32%)	0.00 (-52.24%)	47253.90 (-0.32%)	47213.70 (-0.19%)	47210.90 (-0.12%)	47167.70 (-0.16%)	28.71 (-52.47%)
`0a6c11c` — 2025-12-04 00:39:10	0.18 (n/a)	0.18 (n/a)	0.18 (n/a)	0.18 (n/a)	0.00 (n/a)	47405.00 (n/a)	47301.72 (n/a)	47265.90 (n/a)	47241.80 (n/a)	60.40 (n/a)

mha0

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`1531781` — 2026-01-23 18:49:24	0.18 (-0.08%)	0.18 (+0.22%)	0.18 (+0.32%)	0.18 (+0.35%)	0.00 (-42.58%)	47367.50 (-0.34%)	47232.68 (-0.22%)	47232.00 (-0.32%)	47073.00 (+0.08%)	105.45 (-42.71%)
`0f0da0b` — 2026-01-14 16:31:00	0.18 (n/a)	0.18 (n/a)	0.18 (n/a)	0.18 (n/a)	0.00 (n/a)	47531.20 (n/a)	47335.18 (n/a)	47384.30 (n/a)	47037.20 (n/a)	184.08 (n/a)

relu_1_cols_1_channels_2048_tile_2048

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`1531781` — 2026-01-23 18:49:24	0.11 (+69.81%)	0.09 (+59.10%)	0.10 (+75.49%)	0.05 (+26.24%)	0.03 (+206.46%)	152.90 (-20.82%)	101.62 (-32.81%)	82.20 (-43.00%)	73.00 (-41.13%)	35.75 (+36.67%)
`0f0da0b` — 2026-01-14 16:31:00	0.07 (n/a)	0.06 (n/a)	0.06 (n/a)	0.04 (n/a)	0.01 (n/a)	193.10 (n/a)	151.24 (n/a)	144.20 (n/a)	124.00 (n/a)	26.16 (n/a)

relu_2_cols_1_channels_2048_tile_1024

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`1531781` — 2026-01-23 18:49:24	0.10 (+60.88%)	0.08 (+49.18%)	0.09 (+55.13%)	0.05 (+8.09%)	0.02 (+316.89%)	150.60 (-7.44%)	101.26 (-30.06%)	91.30 (-35.57%)	83.00 (-37.87%)	28.02 (+146.84%)
`0f0da0b` — 2026-01-14 16:31:00	0.06 (n/a)	0.06 (n/a)	0.06 (n/a)	0.05 (n/a)	0.00 (n/a)	162.70 (n/a)	144.78 (n/a)	141.70 (n/a)	133.60 (n/a)	11.35 (n/a)

relu_4_cols_1_channels_2048_tile_512

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`1531781` — 2026-01-23 18:49:24	0.10 (+36.44%)	0.09 (+44.22%)	0.09 (+49.00%)	0.06 (+25.29%)	0.02 (+65.69%)	135.80 (-20.21%)	96.20 (-29.75%)	86.30 (-32.84%)	83.10 (-26.72%)	22.46 (-2.41%)
`0f0da0b` — 2026-01-14 16:31:00	0.07 (n/a)	0.06 (n/a)	0.06 (n/a)	0.05 (n/a)	0.01 (n/a)	170.20 (n/a)	136.94 (n/a)	128.50 (n/a)	113.40 (n/a)	23.02 (n/a)

relu_8_cols_1_channels_2048_tile_256

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`1531781` — 2026-01-23 18:49:24	0.11 (+58.81%)	0.09 (+71.41%)	0.10 (+74.46%)	0.05 (+55.08%)	0.02 (+79.07%)	154.40 (-35.51%)	93.32 (-40.67%)	78.40 (-42.65%)	75.80 (-37.04%)	34.24 (-28.56%)
`0f0da0b` — 2026-01-14 16:31:00	0.07 (n/a)	0.06 (n/a)	0.06 (n/a)	0.03 (n/a)	0.01 (n/a)	239.40 (n/a)	157.28 (n/a)	136.70 (n/a)	120.40 (n/a)	47.93 (n/a)

rms_norm_1_cols_1_channels_2048_tile_2048

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`1531781` — 2026-01-23 18:49:24	0.09 (+27.92%)	0.08 (+35.45%)	0.09 (+45.35%)	0.06 (+20.87%)	0.01 (+35.02%)	131.50 (-17.30%)	100.70 (-25.92%)	92.40 (-31.20%)	90.90 (-21.77%)	17.35 (-10.74%)
`0f0da0b` — 2026-01-14 16:31:00	0.07 (n/a)	0.06 (n/a)	0.06 (n/a)	0.05 (n/a)	0.01 (n/a)	159.00 (n/a)	135.94 (n/a)	134.30 (n/a)	116.20 (n/a)	19.43 (n/a)

rms_norm_1_cols_2_channels_2048_tile_1024

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`1531781` — 2026-01-23 18:49:24	0.12 (+67.22%)	0.10 (+57.83%)	0.11 (+74.39%)	0.06 (-3.10%)	0.03 (+484.78%)	145.20 (+3.20%)	86.54 (-31.66%)	72.10 (-42.64%)	70.50 (-40.25%)	32.80 (+266.62%)
`0f0da0b` — 2026-01-14 16:31:00	0.07 (n/a)	0.06 (n/a)	0.07 (n/a)	0.06 (n/a)	0.00 (n/a)	140.70 (n/a)	126.64 (n/a)	125.70 (n/a)	118.00 (n/a)	8.95 (n/a)

rms_norm_2_cols_1_channels_2048_tile_1024

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`1531781` — 2026-01-23 18:49:24	0.09 (+26.55%)	0.09 (+40.18%)	0.09 (+47.01%)	0.06 (+19.23%)	0.01 (+54.24%)	129.60 (-16.12%)	98.22 (-28.12%)	90.90 (-31.96%)	89.80 (-20.95%)	17.55 (+0.58%)
`0f0da0b` — 2026-01-14 16:31:00	0.07 (n/a)	0.06 (n/a)	0.06 (n/a)	0.05 (n/a)	0.01 (n/a)	154.50 (n/a)	136.64 (n/a)	133.60 (n/a)	113.60 (n/a)	17.45 (n/a)

rms_norm_2_cols_2_channels_2048_tile_512

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`1531781` — 2026-01-23 18:49:24	0.13 (+67.61%)	0.11 (+72.13%)	0.11 (+66.55%)	0.09 (+90.95%)	0.02 (+33.97%)	93.50 (-47.62%)	74.32 (-42.73%)	73.10 (-39.98%)	62.40 (-40.34%)	11.60 (-59.00%)
`0f0da0b` — 2026-01-14 16:31:00	0.08 (n/a)	0.07 (n/a)	0.07 (n/a)	0.05 (n/a)	0.01 (n/a)	178.50 (n/a)	129.78 (n/a)	121.80 (n/a)	104.60 (n/a)	28.30 (n/a)

rms_norm_4_cols_1_channels_2048_tile_512

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`1531781` — 2026-01-23 18:49:24	0.10 (+60.49%)	0.09 (+76.01%)	0.09 (+72.18%)	0.06 (+64.97%)	0.01 (+31.34%)	130.70 (-39.41%)	97.48 (-43.89%)	90.20 (-41.92%)	86.00 (-37.68%)	18.66 (-50.69%)
`0f0da0b` — 2026-01-14 16:31:00	0.06 (n/a)	0.05 (n/a)	0.05 (n/a)	0.04 (n/a)	0.01 (n/a)	215.70 (n/a)	173.74 (n/a)	155.30 (n/a)	138.00 (n/a)	37.85 (n/a)

rms_norm_4_cols_2_channels_2048_tile_256

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`1531781` — 2026-01-23 18:49:24	0.11 (+60.28%)	0.10 (+80.30%)	0.11 (+105.91%)	0.07 (+52.87%)	0.02 (+90.94%)	125.50 (-34.60%)	83.90 (-43.66%)	73.50 (-51.42%)	72.90 (-37.64%)	23.26 (-20.88%)
`0f0da0b` — 2026-01-14 16:31:00	0.07 (n/a)	0.06 (n/a)	0.05 (n/a)	0.04 (n/a)	0.01 (n/a)	191.90 (n/a)	148.92 (n/a)	151.30 (n/a)	116.90 (n/a)	29.40 (n/a)

rms_norm_8_cols_1_channels_2048_tile_256

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`1531781` — 2026-01-23 18:49:24	0.09 (+27.54%)	0.08 (+37.27%)	0.09 (+53.73%)	0.05 (+11.36%)	0.01 (+87.66%)	154.30 (-10.24%)	106.66 (-25.46%)	94.80 (-34.98%)	94.00 (-21.60%)	26.64 (+33.21%)
`0f0da0b` — 2026-01-14 16:31:00	0.07 (n/a)	0.06 (n/a)	0.06 (n/a)	0.05 (n/a)	0.01 (n/a)	171.90 (n/a)	143.10 (n/a)	145.80 (n/a)	119.90 (n/a)	20.00 (n/a)

rms_norm_8_cols_2_channels_2048_tile_128

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`1531781` — 2026-01-23 18:49:24	0.09 (+38.93%)	0.08 (+46.69%)	0.09 (+73.48%)	0.05 (+5.47%)	0.02 (+140.19%)	167.70 (-5.15%)	109.36 (-28.83%)	91.60 (-42.35%)	88.90 (-28.02%)	33.52 (+68.39%)
`0f0da0b` — 2026-01-14 16:31:00	0.07 (n/a)	0.05 (n/a)	0.05 (n/a)	0.05 (n/a)	0.01 (n/a)	176.80 (n/a)	153.66 (n/a)	158.90 (n/a)	123.50 (n/a)	19.91 (n/a)

rope_1_cols_2_channels_4096_tile_4096_0

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`1b2d230` — 2026-01-13 22:38:03	0.19 (-21.66%)	0.17 (-8.53%)	0.18 (+2.09%)	0.14 (-6.71%)	0.02 (-38.07%)	174.10 (+7.20%)	150.06 (+8.11%)	139.60 (-2.04%)	129.80 (+27.63%)	20.36 (-10.65%)
`d48746f` — 2026-01-12 22:17:10	0.24 (n/a)	0.18 (n/a)	0.17 (n/a)	0.15 (n/a)	0.04 (n/a)	162.40 (n/a)	138.80 (n/a)	142.50 (n/a)	101.70 (n/a)	22.79 (n/a)

rope_1c_32rows_512cols_32arows_0m

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`1531781` — 2026-01-23 18:49:24	1.10 (+30.53%)	0.99 (+42.49%)	1.05 (+62.73%)	0.73 (+21.76%)	0.15 (+50.52%)	133.80 (-17.86%)	101.54 (-29.34%)	93.30 (-38.58%)	89.10 (-23.39%)	18.60 (-3.62%)
`0f0da0b` — 2026-01-14 16:31:00	0.85 (n/a)	0.69 (n/a)	0.65 (n/a)	0.60 (n/a)	0.10 (n/a)	162.90 (n/a)	143.70 (n/a)	151.90 (n/a)	116.30 (n/a)	19.30 (n/a)

rope_1c_32rows_512cols_8arows_0m

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`1531781` — 2026-01-23 18:49:24	0.56 (-21.50%)	0.52 (-8.95%)	0.53 (-6.60%)	0.48 (+7.05%)	0.03 (-70.73%)	153.70 (-6.57%)	140.84 (+7.27%)	140.10 (+7.03%)	132.50 (+27.40%)	8.40 (-65.00%)
`0f0da0b` — 2026-01-14 16:31:00	0.71 (n/a)	0.58 (n/a)	0.56 (n/a)	0.45 (n/a)	0.10 (n/a)	164.50 (n/a)	131.30 (n/a)	130.90 (n/a)	104.00 (n/a)	24.00 (n/a)

rope_2_cols_2_channels_4096_tile_2048_0

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`1b2d230` — 2026-01-13 22:38:03	0.21 (-0.56%)	0.17 (-10.62%)	0.16 (-13.58%)	0.13 (-13.33%)	0.03 (+17.05%)	183.70 (+15.39%)	151.76 (+12.77%)	155.20 (+15.73%)	117.30 (+0.60%)	24.24 (+35.73%)
`d48746f` — 2026-01-12 22:17:10	0.21 (n/a)	0.19 (n/a)	0.18 (n/a)	0.15 (n/a)	0.02 (n/a)	159.20 (n/a)	134.58 (n/a)	134.10 (n/a)	116.60 (n/a)	17.86 (n/a)

rope_2c_32rows_512cols_32arows_0m

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`1531781` — 2026-01-23 18:49:24	1.25 (+43.66%)	0.97 (+36.64%)	0.91 (+36.32%)	0.63 (+2.79%)	0.25 (+150.75%)	157.00 (-2.67%)	108.10 (-23.42%)	108.30 (-26.68%)	78.70 (-30.42%)	31.09 (+71.34%)
`0f0da0b` — 2026-01-14 16:31:00	0.87 (n/a)	0.71 (n/a)	0.67 (n/a)	0.61 (n/a)	0.10 (n/a)	161.30 (n/a)	141.16 (n/a)	147.70 (n/a)	113.10 (n/a)	18.15 (n/a)

rope_2c_32rows_512cols_8arows_0m

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`1531781` — 2026-01-23 18:49:24	0.63 (-10.72%)	0.57 (+15.63%)	0.57 (+20.61%)	0.49 (+35.05%)	0.05 (-60.20%)	150.60 (-25.96%)	130.58 (-17.08%)	130.40 (-17.05%)	116.80 (+12.09%)	12.57 (-65.56%)
`0f0da0b` — 2026-01-14 16:31:00	0.71 (n/a)	0.49 (n/a)	0.47 (n/a)	0.36 (n/a)	0.13 (n/a)	203.40 (n/a)	157.48 (n/a)	157.20 (n/a)	104.20 (n/a)	36.49 (n/a)

rope_4_cols_2_channels_4096_tile_1024_0

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`1b2d230` — 2026-01-13 22:38:03	0.22 (+18.92%)	0.17 (+6.24%)	0.17 (+5.57%)	0.14 (-0.50%)	0.03 (+90.77%)	171.10 (+0.47%)	147.40 (-4.39%)	148.10 (-5.31%)	110.10 (-15.89%)	23.59 (+60.67%)
`d48746f` — 2026-01-12 22:17:10	0.19 (n/a)	0.16 (n/a)	0.16 (n/a)	0.14 (n/a)	0.02 (n/a)	170.30 (n/a)	154.16 (n/a)	156.40 (n/a)	130.90 (n/a)	14.69 (n/a)

rope_8_cols_2_channels_4096_tile_512_0

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`1b2d230` — 2026-01-13 22:38:03	0.17 (-1.30%)	0.15 (+6.63%)	0.15 (-8.78%)	0.14 (+39.68%)	0.01 (-67.35%)	174.90 (-28.41%)	163.42 (-10.36%)	167.40 (+9.63%)	146.50 (+1.31%)	11.19 (-76.17%)
`d48746f` — 2026-01-12 22:17:10	0.17 (n/a)	0.14 (n/a)	0.16 (n/a)	0.10 (n/a)	0.03 (n/a)	244.30 (n/a)	182.30 (n/a)	152.70 (n/a)	144.60 (n/a)	46.98 (n/a)

rope_8c_32rows_512cols_32arows_0m

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`1531781` — 2026-01-23 18:49:24	1.34 (+62.86%)	0.89 (+47.92%)	0.85 (+53.70%)	0.60 (+49.43%)	0.27 (+64.12%)	163.50 (-33.07%)	117.38 (-32.11%)	115.50 (-34.93%)	73.60 (-38.56%)	32.08 (-33.78%)
`0f0da0b` — 2026-01-14 16:31:00	0.82 (n/a)	0.60 (n/a)	0.55 (n/a)	0.40 (n/a)	0.16 (n/a)	244.30 (n/a)	172.90 (n/a)	177.50 (n/a)	119.80 (n/a)	48.44 (n/a)

rope_8c_32rows_512cols_8arows_0m

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`1531781` — 2026-01-23 18:49:24	0.54 (-10.87%)	0.52 (+6.74%)	0.53 (+12.23%)	0.48 (+19.79%)	0.03 (-67.26%)	152.90 (-16.49%)	143.32 (-8.02%)	139.70 (-10.91%)	135.50 (+12.17%)	7.46 (-68.88%)
`0f0da0b` — 2026-01-14 16:31:00	0.61 (n/a)	0.48 (n/a)	0.47 (n/a)	0.40 (n/a)	0.08 (n/a)	183.10 (n/a)	155.82 (n/a)	156.80 (n/a)	120.80 (n/a)	23.99 (n/a)

sigmoid_1_cols_1_channels_2048_tile_2048

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`1531781` — 2026-01-23 18:49:24	0.12 (+53.53%)	0.10 (+54.49%)	0.12 (+106.52%)	0.05 (-0.16%)	0.03 (+144.85%)	153.20 (+0.13%)	90.92 (-30.39%)	68.60 (-51.55%)	68.30 (-34.89%)	36.80 (+58.03%)
`0f0da0b` — 2026-01-14 16:31:00	0.08 (n/a)	0.06 (n/a)	0.06 (n/a)	0.05 (n/a)	0.01 (n/a)	153.00 (n/a)	130.62 (n/a)	141.60 (n/a)	104.90 (n/a)	23.29 (n/a)

sigmoid_2_cols_1_channels_2048_tile_1024

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`1531781` — 2026-01-23 18:49:24	0.12 (+73.69%)	0.10 (+60.60%)	0.12 (+98.01%)	0.06 (+2.86%)	0.03 (+394.96%)	140.80 (-2.83%)	89.84 (-32.71%)	68.90 (-49.52%)	68.10 (-42.43%)	32.04 (+163.16%)
`0f0da0b` — 2026-01-14 16:31:00	0.07 (n/a)	0.06 (n/a)	0.06 (n/a)	0.06 (n/a)	0.01 (n/a)	144.90 (n/a)	133.52 (n/a)	136.50 (n/a)	118.30 (n/a)	12.18 (n/a)

sigmoid_4_cols_1_channels_2048_tile_512

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`1531781` — 2026-01-23 18:49:24	0.10 (+30.15%)	0.08 (+37.20%)	0.09 (+57.41%)	0.05 (+2.94%)	0.02 (+110.00%)	169.40 (-2.87%)	106.58 (-23.47%)	87.40 (-36.48%)	84.00 (-23.15%)	36.37 (+56.68%)
`0f0da0b` — 2026-01-14 16:31:00	0.07 (n/a)	0.06 (n/a)	0.06 (n/a)	0.05 (n/a)	0.01 (n/a)	174.40 (n/a)	139.26 (n/a)	137.60 (n/a)	109.30 (n/a)	23.21 (n/a)

sigmoid_8_cols_1_channels_2048_tile_256

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`1531781` — 2026-01-23 18:49:24	0.12 (+109.66%)	0.10 (+84.35%)	0.12 (+117.89%)	0.05 (+3.89%)	0.03 (+649.09%)	157.30 (-3.73%)	89.74 (-39.88%)	68.70 (-54.11%)	65.90 (-52.35%)	38.85 (+251.55%)
`0f0da0b` — 2026-01-14 16:31:00	0.06 (n/a)	0.06 (n/a)	0.05 (n/a)	0.05 (n/a)	0.00 (n/a)	163.40 (n/a)	149.28 (n/a)	149.70 (n/a)	138.30 (n/a)	11.05 (n/a)

silu_1_cols_1_channels_2048_tile_2048

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`1531781` — 2026-01-23 18:49:24	0.10 (+44.80%)	0.08 (+33.22%)	0.10 (+66.93%)	0.06 (-1.08%)	0.02 (+301.40%)	143.30 (+1.06%)	104.88 (-20.01%)	82.20 (-40.13%)	80.60 (-30.93%)	32.24 (+177.47%)
`0f0da0b` — 2026-01-14 16:31:00	0.07 (n/a)	0.06 (n/a)	0.06 (n/a)	0.06 (n/a)	0.01 (n/a)	141.80 (n/a)	131.12 (n/a)	137.30 (n/a)	116.70 (n/a)	11.62 (n/a)

silu_2_cols_1_channels_2048_tile_1024

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`1531781` — 2026-01-23 18:49:24	0.10 (+34.12%)	0.08 (+47.27%)	0.10 (+69.15%)	0.06 (+39.46%)	0.02 (+73.50%)	138.90 (-28.29%)	103.10 (-30.67%)	83.20 (-40.91%)	81.60 (-25.41%)	28.30 (-8.88%)
`0f0da0b` — 2026-01-14 16:31:00	0.07 (n/a)	0.06 (n/a)	0.06 (n/a)	0.04 (n/a)	0.01 (n/a)	193.70 (n/a)	148.70 (n/a)	140.80 (n/a)	109.40 (n/a)	31.06 (n/a)

silu_4_cols_1_channels_2048_tile_512

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`1531781` — 2026-01-23 18:49:24	0.12 (+42.12%)	0.11 (+61.07%)	0.12 (+92.74%)	0.06 (+8.38%)	0.03 (+104.00%)	137.40 (-7.72%)	83.36 (-34.98%)	70.00 (-48.07%)	68.10 (-29.65%)	30.27 (+32.68%)
`0f0da0b` — 2026-01-14 16:31:00	0.08 (n/a)	0.07 (n/a)	0.06 (n/a)	0.06 (n/a)	0.01 (n/a)	148.90 (n/a)	128.20 (n/a)	134.80 (n/a)	96.80 (n/a)	22.82 (n/a)

silu_8_cols_1_channels_2048_tile_256

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`1531781` — 2026-01-23 18:49:24	0.12 (+35.33%)	0.09 (+59.58%)	0.09 (+78.47%)	0.07 (+60.20%)	0.02 (-1.76%)	109.80 (-37.58%)	89.20 (-38.82%)	87.90 (-43.98%)	71.10 (-26.09%)	14.50 (-53.17%)
`0f0da0b` — 2026-01-14 16:31:00	0.09 (n/a)	0.06 (n/a)	0.05 (n/a)	0.05 (n/a)	0.02 (n/a)	175.90 (n/a)	145.80 (n/a)	156.90 (n/a)	96.20 (n/a)	30.97 (n/a)

softmax_1_cols_2_channels_4096_tile_2048

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`1531781` — 2026-01-23 18:49:24	0.21 (+20.19%)	0.16 (+13.84%)	0.16 (+8.19%)	0.12 (+26.61%)	0.04 (+17.18%)	135.50 (-20.99%)	104.68 (-12.65%)	102.70 (-7.56%)	76.50 (-16.85%)	22.55 (-26.21%)
`0f0da0b` — 2026-01-14 16:31:00	0.18 (n/a)	0.14 (n/a)	0.15 (n/a)	0.10 (n/a)	0.03 (n/a)	171.50 (n/a)	119.84 (n/a)	111.10 (n/a)	92.00 (n/a)	30.56 (n/a)

softmax_2_cols_2_channels_4096_tile_1024

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`1531781` — 2026-01-23 18:49:24	0.22 (+40.26%)	0.18 (+43.33%)	0.20 (+64.08%)	0.13 (+24.34%)	0.04 (+68.98%)	127.80 (-19.62%)	94.74 (-28.95%)	83.40 (-39.08%)	73.90 (-28.67%)	24.46 (-5.31%)
`0f0da0b` — 2026-01-14 16:31:00	0.16 (n/a)	0.13 (n/a)	0.12 (n/a)	0.10 (n/a)	0.03 (n/a)	159.00 (n/a)	133.34 (n/a)	136.90 (n/a)	103.60 (n/a)	25.83 (n/a)

softmax_2_cols_2_channels_4096_tile_512

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`1531781` — 2026-01-23 18:49:24	0.24 (+85.21%)	0.17 (+69.97%)	0.15 (+54.16%)	0.11 (+37.21%)	0.05 (+203.82%)	148.10 (-27.12%)	102.78 (-37.88%)	106.70 (-35.10%)	69.60 (-45.96%)	31.81 (+15.98%)
`0f0da0b` — 2026-01-14 16:31:00	0.13 (n/a)	0.10 (n/a)	0.10 (n/a)	0.08 (n/a)	0.02 (n/a)	203.20 (n/a)	165.46 (n/a)	164.40 (n/a)	128.80 (n/a)	27.43 (n/a)

swiglu

No metrics available.

swiglu_decode_1x2048x2048

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`130b6ea` — 2025-12-05 21:15:49	0.00 (-15.46%)	0.00 (-27.20%)	0.00 (-28.17%)	0.00 (-33.90%)	0.00 (+1834.83%)	5187.20 (+51.28%)	4720.84 (+38.38%)	4750.90 (+39.22%)	4015.10 (+18.29%)	390.23 (+3298.45%)
`0a6c11c` — 2025-12-04 00:39:10	0.00 (n/a)	0.00 (n/a)	0.00 (n/a)	0.00 (n/a)	0.00 (n/a)	3428.80 (n/a)	3411.62 (n/a)	3412.60 (n/a)	3394.20 (n/a)	11.48 (n/a)

swiglu_decode_1x2048x2048_0

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`1531781` — 2026-01-23 18:49:24	0.00 (+28.14%)	0.00 (+20.30%)	0.00 (+10.00%)	0.00 (+11.66%)	0.00 (+124.73%)	5111.40 (-10.44%)	4324.88 (-14.91%)	4736.10 (-9.09%)	3410.70 (-21.96%)	843.97 (+53.51%)
`0f0da0b` — 2026-01-14 16:31:00	0.00 (n/a)	0.00 (n/a)	0.00 (n/a)	0.00 (n/a)	0.00 (n/a)	5707.20 (n/a)	5083.00 (n/a)	5209.60 (n/a)	4370.50 (n/a)	549.78 (n/a)

tanh_1_cols_1_channels_2048_tile_2048

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`1531781` — 2026-01-23 18:49:24	0.10 (+42.65%)	0.08 (+33.21%)	0.07 (+27.20%)	0.06 (+13.46%)	0.02 (+130.64%)	140.80 (-11.89%)	109.14 (-23.00%)	115.90 (-21.37%)	85.60 (-29.89%)	23.52 (+37.00%)
`0f0da0b` — 2026-01-14 16:31:00	0.07 (n/a)	0.06 (n/a)	0.06 (n/a)	0.05 (n/a)	0.01 (n/a)	159.80 (n/a)	141.74 (n/a)	147.40 (n/a)	122.10 (n/a)	17.17 (n/a)

tanh_2_cols_1_channels_2048_tile_1024

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`1531781` — 2026-01-23 18:49:24	0.11 (+41.27%)	0.09 (+40.30%)	0.09 (+45.94%)	0.08 (+28.08%)	0.01 (+83.15%)	103.30 (-21.98%)	88.02 (-28.44%)	87.80 (-31.46%)	77.80 (-29.21%)	9.64 (+2.40%)
`0f0da0b` — 2026-01-14 16:31:00	0.07 (n/a)	0.07 (n/a)	0.06 (n/a)	0.06 (n/a)	0.01 (n/a)	132.40 (n/a)	123.00 (n/a)	128.10 (n/a)	109.90 (n/a)	9.41 (n/a)

tanh_4_cols_1_channels_2048_tile_512

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`1531781` — 2026-01-23 18:49:24	0.10 (+48.49%)	0.09 (+56.51%)	0.09 (+68.30%)	0.07 (+54.99%)	0.01 (+44.32%)	121.40 (-35.46%)	93.40 (-36.24%)	86.30 (-40.61%)	84.60 (-32.70%)	15.80 (-36.89%)
`0f0da0b` — 2026-01-14 16:31:00	0.07 (n/a)	0.06 (n/a)	0.06 (n/a)	0.04 (n/a)	0.01 (n/a)	188.10 (n/a)	146.48 (n/a)	145.30 (n/a)	125.70 (n/a)	25.04 (n/a)

tanh_8_cols_1_channels_2048_tile_256

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`1531781` — 2026-01-23 18:49:24	0.11 (+64.36%)	0.09 (+41.14%)	0.10 (+56.21%)	0.05 (-7.33%)	0.03 (+346.69%)	163.80 (+7.91%)	102.90 (-22.05%)	80.40 (-35.94%)	71.70 (-39.13%)	39.99 (+185.25%)
`0f0da0b` — 2026-01-14 16:31:00	0.07 (n/a)	0.06 (n/a)	0.07 (n/a)	0.05 (n/a)	0.01 (n/a)	151.80 (n/a)	132.00 (n/a)	125.50 (n/a)	117.80 (n/a)	14.02 (n/a)

transpose_2048_M_64_N_1_cols_1_channels_64_m_64_n_8_s

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`130b6ea` — 2025-12-05 21:15:49	3.93 (-7.02%)	3.68 (+2.48%)	3.66 (+1.82%)	3.45 (+15.62%)	0.15 (-63.64%)	151.80 (-13.50%)	142.70 (-3.59%)	143.40 (-1.78%)	133.50 (+7.49%)	5.90 (-66.22%)
`0a6c11c` — 2025-12-04 00:39:10	4.22 (n/a)	3.59 (n/a)	3.59 (n/a)	2.99 (n/a)	0.42 (n/a)	175.50 (n/a)	148.02 (n/a)	146.00 (n/a)	124.20 (n/a)	17.46 (n/a)

transpose_2048_M_64_N_1_cols_1_channels_64_m_64_n_8_s0

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`1531781` — 2026-01-23 18:49:24	4.95 (+35.28%)	4.34 (+33.32%)	4.37 (+25.58%)	3.74 (+35.73%)	0.50 (+11.62%)	140.30 (-26.35%)	121.98 (-25.39%)	119.90 (-20.33%)	106.00 (-26.08%)	14.15 (-39.67%)
`0f0da0b` — 2026-01-14 16:31:00	3.66 (n/a)	3.26 (n/a)	3.48 (n/a)	2.75 (n/a)	0.45 (n/a)	190.50 (n/a)	163.48 (n/a)	150.50 (n/a)	143.40 (n/a)	23.46 (n/a)

transpose_2048_M_64_N_1_cols_2_channels_64_m_64_n_8_s

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`130b6ea` — 2025-12-05 21:15:49	5.11 (+15.57%)	3.85 (-3.32%)	3.40 (-18.64%)	3.29 (+7.09%)	0.71 (+49.14%)	159.60 (-6.61%)	140.18 (+4.80%)	154.20 (+22.97%)	102.60 (-13.49%)	22.70 (+19.77%)
`0a6c11c` — 2025-12-04 00:39:10	4.42 (n/a)	3.99 (n/a)	4.18 (n/a)	3.07 (n/a)	0.48 (n/a)	170.90 (n/a)	133.76 (n/a)	125.40 (n/a)	118.60 (n/a)	18.95 (n/a)

transpose_2048_M_64_N_1_cols_2_channels_64_m_64_n_8_s0

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`1531781` — 2026-01-23 18:49:24	6.07 (+44.96%)	4.99 (+45.77%)	5.09 (+52.05%)	4.09 (+35.80%)	0.74 (+53.13%)	128.10 (-26.34%)	106.86 (-31.23%)	103.00 (-34.27%)	86.30 (-31.02%)	15.67 (-22.94%)
`0f0da0b` — 2026-01-14 16:31:00	4.19 (n/a)	3.42 (n/a)	3.35 (n/a)	3.01 (n/a)	0.48 (n/a)	173.90 (n/a)	155.38 (n/a)	156.70 (n/a)	125.10 (n/a)	20.34 (n/a)

weighted_rms_norm_1_cols_2_channels_2048_weights_2048

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`1531781` — 2026-01-23 18:49:24	0.11 (+41.27%)	0.09 (+39.11%)	0.09 (+47.90%)	0.05 (+1.24%)	0.02 (+132.19%)	155.00 (-1.21%)	99.92 (-24.84%)	89.60 (-32.43%)	76.50 (-29.17%)	31.36 (+74.29%)
`0f0da0b` — 2026-01-14 16:31:00	0.08 (n/a)	0.06 (n/a)	0.06 (n/a)	0.05 (n/a)	0.01 (n/a)	156.90 (n/a)	132.94 (n/a)	132.60 (n/a)	108.00 (n/a)	17.99 (n/a)

weighted_rms_norm_2_cols_2_channels_2048_weights_1024

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`1531781` — 2026-01-23 18:49:24	0.10 (+48.28%)	0.09 (+48.11%)	0.09 (+60.52%)	0.05 (+2.82%)	0.02 (+134.54%)	166.70 (-2.74%)	102.72 (-28.96%)	88.50 (-37.68%)	83.10 (-32.55%)	35.87 (+61.58%)
`0f0da0b` — 2026-01-14 16:31:00	0.07 (n/a)	0.06 (n/a)	0.06 (n/a)	0.05 (n/a)	0.01 (n/a)	171.40 (n/a)	144.60 (n/a)	142.00 (n/a)	123.20 (n/a)	22.20 (n/a)

weighted_rms_norm_4_cols_2_channels_2048_weights_512

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`1531781` — 2026-01-23 18:49:24	0.09 (+21.03%)	0.09 (+31.88%)	0.09 (+37.19%)	0.07 (+25.91%)	0.01 (+25.73%)	119.00 (-20.61%)	96.48 (-24.14%)	89.90 (-27.09%)	89.30 (-17.31%)	12.81 (-17.86%)
`0f0da0b` — 2026-01-14 16:31:00	0.08 (n/a)	0.07 (n/a)	0.07 (n/a)	0.05 (n/a)	0.01 (n/a)	149.90 (n/a)	127.18 (n/a)	123.30 (n/a)	108.00 (n/a)	15.59 (n/a)

weighted_rms_norm_8_cols_2_channels_2048_weights_256

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`1531781` — 2026-01-23 18:49:24	0.11 (+67.16%)	0.10 (+80.58%)	0.10 (+80.44%)	0.06 (+74.63%)	0.02 (+72.16%)	135.80 (-42.72%)	89.92 (-44.57%)	80.00 (-44.60%)	74.30 (-40.23%)	25.76 (-41.27%)
`0f0da0b` — 2026-01-14 16:31:00	0.07 (n/a)	0.05 (n/a)	0.06 (n/a)	0.03 (n/a)	0.01 (n/a)	237.10 (n/a)	162.22 (n/a)	144.40 (n/a)	124.30 (n/a)	43.86 (n/a)

github-actions · 2026-02-06T16:28:59Z

📊 Test Results for Test Example Applications

b0ad401 (2026_02_06_16_28_30)

IRONCLAD

Tested on 2026_02_06_16_28_30 at commit b0ad401.

Test	Checks	TTFT (mean)	TPS (mean)
llama_3.2_1b_prompt_13_tokens_1	❌ 0/1	n/a	n/a
llama_3.2_1b_prompt_13_tokens_40	❌ 0/1	n/a	n/a
llama_3.2_1b_prompt_2048_tokens_1	❌ 0/1	n/a	n/a
llama_3.2_1b_prompt_2048_tokens_40	❌ 0/1	n/a	n/a

📈 Trends (vs main branch) for Test Example Applications

b0ad401 (2026_02_06_16_28_30)

IRONCLAD Trends

llama_3.2_1b

Commit/Date	Num Tokens (max)	Num Tokens (mean)	Num Tokens (median)	Num Tokens (min)	Num Tokens (stddev)	TPS (max)	TPS (mean)	TPS (median)	TPS (min)	TPS (stddev)	TTFT (max)	TTFT (mean)	TTFT (median)	TTFT (min)	TTFT (stddev)	Total (max)	Total (mean)	Total (median)	Total (min)	Total (stddev)
`130b6ea` — 2025-12-05 21:33:12	40.00 (+0.00%)	40.00 (+0.00%)	40.00 (+0.00%)	40.00 (+0.00%)	0.00 (n/a)	4.71 (-0.42%)	4.64 (-0.09%)	4.64 (+0.65%)	4.55 (-0.22%)	0.05 (-17.66%)	4.41 (-0.34%)	4.39 (-0.19%)	4.38 (-0.33%)	4.37 (-0.15%)	0.01 (-25.90%)	12.96 (-0.00%)	12.80 (+0.07%)	12.80 (-0.23%)	12.67 (+0.44%)	0.09 (-21.12%)
`0a6c11c` — 2025-12-03 23:35:15	40.00 (n/a)	40.00 (n/a)	40.00 (n/a)	40.00 (n/a)	0.00 (n/a)	4.73 (n/a)	4.64 (n/a)	4.61 (n/a)	4.56 (n/a)	0.06 (n/a)	4.42 (n/a)	4.40 (n/a)	4.40 (n/a)	4.37 (n/a)	0.02 (n/a)	12.96 (n/a)	12.79 (n/a)	12.83 (n/a)	12.62 (n/a)	0.12 (n/a)

llama_3.2_1b_prompt_13_tokens_1

Commit/Date	Num_Tokens (max)	Num_Tokens (mean)	Num_Tokens (median)	Num_Tokens (min)	Num_Tokens (stddev)	TPS (max)	TPS (mean)	TPS (median)	TPS (min)	TPS (stddev)	TTFT (max)	TTFT (mean)	TTFT (median)	TTFT (min)	TTFT (stddev)
`b0ad401` — 2026-02-06 16:28:25	n/a (n/a)	n/a (n/a)	n/a (n/a)	n/a (n/a)	n/a (n/a)	n/a (n/a)	n/a (n/a)	n/a (n/a)	n/a (n/a)	n/a (n/a)	n/a (n/a)	n/a (n/a)	n/a (n/a)	n/a (n/a)	n/a (n/a)
`1531781` — 2026-01-23 18:40:34	1.00 (n/a)	1.00 (n/a)	1.00 (n/a)	1.00 (n/a)	0.00 (n/a)	0.00 (n/a)	0.00 (n/a)	0.00 (n/a)	0.00 (n/a)	0.00 (n/a)	0.61 (n/a)	0.61 (n/a)	0.61 (n/a)	0.61 (n/a)	0.00 (n/a)

llama_3.2_1b_prompt_13_tokens_40

Commit/Date	Num_Tokens (max)	Num_Tokens (mean)	Num_Tokens (median)	Num_Tokens (min)	Num_Tokens (stddev)	TPS (max)	TPS (mean)	TPS (median)	TPS (min)	TPS (stddev)	TTFT (max)	TTFT (mean)	TTFT (median)	TTFT (min)	TTFT (stddev)
`b0ad401` — 2026-02-06 16:28:25	n/a (n/a)	n/a (n/a)	n/a (n/a)	n/a (n/a)	n/a (n/a)	n/a (n/a)	n/a (n/a)	n/a (n/a)	n/a (n/a)	n/a (n/a)	n/a (n/a)	n/a (n/a)	n/a (n/a)	n/a (n/a)	n/a (n/a)
`1531781` — 2026-01-23 18:40:34	40.00 (n/a)	40.00 (n/a)	40.00 (n/a)	40.00 (n/a)	0.00 (n/a)	4.34 (n/a)	4.34 (n/a)	4.34 (n/a)	4.34 (n/a)	0.00 (n/a)	0.61 (n/a)	0.61 (n/a)	0.61 (n/a)	0.61 (n/a)	0.00 (n/a)

llama_3.2_1b_prompt_2048_tokens_1

Commit/Date	Num_Tokens (max)	Num_Tokens (mean)	Num_Tokens (median)	Num_Tokens (min)	Num_Tokens (stddev)	TPS (max)	TPS (mean)	TPS (median)	TPS (min)	TPS (stddev)	TTFT (max)	TTFT (mean)	TTFT (median)	TTFT (min)	TTFT (stddev)
`b0ad401` — 2026-02-06 16:28:25	n/a (n/a)	n/a (n/a)	n/a (n/a)	n/a (n/a)	n/a (n/a)	n/a (n/a)	n/a (n/a)	n/a (n/a)	n/a (n/a)	n/a (n/a)	n/a (n/a)	n/a (n/a)	n/a (n/a)	n/a (n/a)	n/a (n/a)
`1531781` — 2026-01-23 18:40:34	1.00 (n/a)	1.00 (n/a)	1.00 (n/a)	1.00 (n/a)	0.00 (n/a)	0.00 (n/a)	0.00 (n/a)	0.00 (n/a)	0.00 (n/a)	0.00 (n/a)	2.70 (n/a)	2.70 (n/a)	2.70 (n/a)	2.70 (n/a)	0.00 (n/a)

llama_3.2_1b_prompt_2048_tokens_40

Commit/Date	Num_Tokens (max)	Num_Tokens (mean)	Num_Tokens (median)	Num_Tokens (min)	Num_Tokens (stddev)	TPS (max)	TPS (mean)	TPS (median)	TPS (min)	TPS (stddev)	TTFT (max)	TTFT (mean)	TTFT (median)	TTFT (min)	TTFT (stddev)
`b0ad401` — 2026-02-06 16:28:25	n/a (n/a)	n/a (n/a)	n/a (n/a)	n/a (n/a)	n/a (n/a)	n/a (n/a)	n/a (n/a)	n/a (n/a)	n/a (n/a)	n/a (n/a)	n/a (n/a)	n/a (n/a)	n/a (n/a)	n/a (n/a)	n/a (n/a)
`1531781` — 2026-01-23 18:40:34	40.00 (n/a)	40.00 (n/a)	40.00 (n/a)	40.00 (n/a)	0.00 (n/a)	4.22 (n/a)	4.22 (n/a)	4.22 (n/a)	4.22 (n/a)	0.00 (n/a)	2.70 (n/a)	2.70 (n/a)	2.70 (n/a)	2.70 (n/a)	0.00 (n/a)

github-actions · 2026-02-06T17:29:57Z

📊 Test Results for Small Benchmark/Test Suite

dfee1e0 (2026_02_06_17_29_36)

IRONCLAD

Tested on 2026_02_06_17_29_36 at commit dfee1e0.

Test	Checks	Latency (mean)	Bandwidth (mean)	GFLOP/s (mean)

📈 Trends (vs main branch) for Small Benchmark/Test Suite

dfee1e0 (2026_02_06_17_29_36)

IRONCLAD Trends

axpy_1_cols_2_channels_2048_tile_2048_3.0

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`130b6ea` — 2025-12-05 21:15:49	0.13 (-1.96%)	0.11 (+2.52%)	0.10 (+1.86%)	0.08 (-1.91%)	0.02 (-1.72%)	162.40 (+1.95%)	120.96 (-2.36%)	117.00 (-1.85%)	94.40 (+2.05%)	23.11 (+4.09%)
`0a6c11c` — 2025-12-04 00:39:10	0.13 (n/a)	0.10 (n/a)	0.10 (n/a)	0.08 (n/a)	0.02 (n/a)	159.30 (n/a)	123.88 (n/a)	119.20 (n/a)	92.50 (n/a)	22.21 (n/a)

axpy_1_cols_2_channels_2048_tile_2048_3.0_0

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`1531781` — 2026-01-23 18:49:24	0.16 (+80.10%)	0.13 (+54.74%)	0.11 (+34.32%)	0.10 (+39.85%)	0.03 (+432.10%)	118.80 (-28.48%)	100.30 (-33.22%)	110.40 (-25.56%)	78.00 (-44.48%)	20.49 (+105.98%)
`0f0da0b` — 2026-01-14 16:31:00	0.09 (n/a)	0.08 (n/a)	0.08 (n/a)	0.07 (n/a)	0.01 (n/a)	166.10 (n/a)	150.20 (n/a)	148.30 (n/a)	140.50 (n/a)	9.95 (n/a)

axpy_2_cols_2_channels_2048_tile_1024_3.0

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`130b6ea` — 2025-12-05 21:15:49	0.11 (-26.77%)	0.09 (-9.09%)	0.09 (-10.81%)	0.08 (+8.25%)	0.01 (-55.42%)	152.00 (-7.60%)	135.46 (+6.28%)	142.80 (+12.18%)	115.00 (+36.58%)	15.03 (-41.36%)
`0a6c11c` — 2025-12-04 00:39:10	0.15 (n/a)	0.10 (n/a)	0.10 (n/a)	0.07 (n/a)	0.02 (n/a)	164.50 (n/a)	127.46 (n/a)	127.30 (n/a)	84.20 (n/a)	25.64 (n/a)

axpy_2_cols_2_channels_2048_tile_1024_3.0_0

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`1531781` — 2026-01-23 18:49:24	0.14 (+54.25%)	0.12 (+45.65%)	0.14 (+61.31%)	0.08 (-2.79%)	0.03 (+557.87%)	156.90 (+2.89%)	104.46 (-28.09%)	90.30 (-37.98%)	87.90 (-35.22%)	29.52 (+351.33%)
`0f0da0b` — 2026-01-14 16:31:00	0.09 (n/a)	0.08 (n/a)	0.08 (n/a)	0.08 (n/a)	0.00 (n/a)	152.50 (n/a)	145.26 (n/a)	145.60 (n/a)	135.70 (n/a)	6.54 (n/a)

axpy_4_cols_2_channels_2048_tile_512_3.0

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`130b6ea` — 2025-12-05 21:15:49	0.12 (-10.21%)	0.10 (-8.41%)	0.10 (-6.20%)	0.07 (+3.52%)	0.02 (-23.48%)	166.70 (-3.36%)	128.78 (+7.33%)	121.80 (+6.65%)	103.90 (+11.36%)	22.95 (-18.82%)
`0a6c11c` — 2025-12-04 00:39:10	0.13 (n/a)	0.11 (n/a)	0.11 (n/a)	0.07 (n/a)	0.02 (n/a)	172.50 (n/a)	119.98 (n/a)	114.20 (n/a)	93.30 (n/a)	28.27 (n/a)

axpy_4_cols_2_channels_2048_tile_512_3.0_0

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`1531781` — 2026-01-23 18:49:24	0.16 (+65.19%)	0.14 (+57.55%)	0.15 (+63.48%)	0.09 (+24.56%)	0.03 (+220.28%)	134.10 (-19.70%)	92.44 (-34.51%)	83.40 (-38.81%)	78.60 (-39.49%)	23.38 (+58.01%)
`0f0da0b` — 2026-01-14 16:31:00	0.09 (n/a)	0.09 (n/a)	0.09 (n/a)	0.07 (n/a)	0.01 (n/a)	167.00 (n/a)	141.16 (n/a)	136.30 (n/a)	129.90 (n/a)	14.80 (n/a)

axpy_8_cols_2_channels_2048_tile_256_3.0

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`130b6ea` — 2025-12-05 21:15:49	0.10 (-16.19%)	0.08 (-10.40%)	0.09 (-2.09%)	0.07 (-17.38%)	0.01 (-10.61%)	184.40 (+21.00%)	153.18 (+11.99%)	144.10 (+2.20%)	127.10 (+19.34%)	21.26 (+34.76%)
`0a6c11c` — 2025-12-04 00:39:10	0.12 (n/a)	0.09 (n/a)	0.09 (n/a)	0.08 (n/a)	0.01 (n/a)	152.40 (n/a)	136.78 (n/a)	141.00 (n/a)	106.50 (n/a)	15.78 (n/a)

axpy_8_cols_2_channels_2048_tile_256_3.0_0

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`1531781` — 2026-01-23 18:49:24	0.17 (+67.03%)	0.14 (+66.86%)	0.15 (+90.72%)	0.08 (+20.53%)	0.03 (+162.97%)	146.80 (-17.06%)	92.24 (-37.11%)	80.00 (-47.54%)	72.30 (-40.15%)	31.07 (+36.92%)
`0f0da0b` — 2026-01-14 16:31:00	0.10 (n/a)	0.09 (n/a)	0.08 (n/a)	0.07 (n/a)	0.01 (n/a)	177.00 (n/a)	146.68 (n/a)	152.50 (n/a)	120.80 (n/a)	22.69 (n/a)

dequant_1_cols_1_channels_2048_tile_2048

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`130b6ea` — 2025-12-05 21:15:49	0.05 (-18.83%)	0.04 (-5.21%)	0.04 (-2.69%)	0.03 (+37.10%)	0.00 (-53.98%)	160.40 (-27.06%)	141.12 (-0.91%)	136.40 (+2.79%)	114.30 (+23.17%)	17.37 (-59.06%)
`0a6c11c` — 2025-12-04 00:39:10	0.06 (n/a)	0.04 (n/a)	0.04 (n/a)	0.02 (n/a)	0.01 (n/a)	219.90 (n/a)	142.42 (n/a)	132.70 (n/a)	92.80 (n/a)	42.43 (n/a)

dequant_1_cols_1_channels_2048_tile_2048_0

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`1531781` — 2026-01-23 18:49:24	0.06 (+13.11%)	0.05 (+1.55%)	0.04 (-0.75%)	0.04 (+7.05%)	0.01 (+25.85%)	125.50 (-6.55%)	115.08 (-1.12%)	121.60 (+0.75%)	88.10 (-11.63%)	15.40 (+4.09%)
`0f0da0b` — 2026-01-14 16:31:00	0.05 (n/a)	0.05 (n/a)	0.04 (n/a)	0.04 (n/a)	0.01 (n/a)	134.30 (n/a)	116.38 (n/a)	120.70 (n/a)	99.70 (n/a)	14.79 (n/a)

dequant_1_cols_2_channels_2048_tile_1024

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`130b6ea` — 2025-12-05 21:15:49	0.06 (-13.28%)	0.05 (-8.93%)	0.06 (-13.59%)	0.04 (+4.63%)	0.01 (-39.10%)	133.60 (-4.43%)	106.08 (+5.85%)	95.10 (+15.69%)	87.90 (+15.35%)	17.90 (-31.97%)
`0a6c11c` — 2025-12-04 00:39:10	0.07 (n/a)	0.06 (n/a)	0.06 (n/a)	0.04 (n/a)	0.01 (n/a)	139.80 (n/a)	100.22 (n/a)	82.20 (n/a)	76.20 (n/a)	26.31 (n/a)

dequant_1_cols_2_channels_2048_tile_1024_0

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`1531781` — 2026-01-23 18:49:24	0.07 (+24.45%)	0.05 (+9.74%)	0.05 (-2.94%)	0.03 (+16.88%)	0.01 (+14.24%)	163.70 (-14.47%)	114.66 (-9.57%)	111.30 (+3.06%)	76.50 (-19.64%)	31.63 (-21.04%)
`0f0da0b` — 2026-01-14 16:31:00	0.06 (n/a)	0.04 (n/a)	0.05 (n/a)	0.03 (n/a)	0.01 (n/a)	191.40 (n/a)	126.80 (n/a)	108.00 (n/a)	95.20 (n/a)	40.06 (n/a)

dequant_2_cols_1_channels_2048_tile_1024

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`130b6ea` — 2025-12-05 21:15:49	0.07 (+5.66%)	0.05 (+1.21%)	0.05 (+0.61%)	0.04 (-15.02%)	0.01 (+58.42%)	143.40 (+17.64%)	101.02 (+1.49%)	97.60 (-0.61%)	78.90 (-5.40%)	22.56 (+78.52%)
`0a6c11c` — 2025-12-04 00:39:10	0.06 (n/a)	0.05 (n/a)	0.05 (n/a)	0.04 (n/a)	0.01 (n/a)	121.90 (n/a)	99.54 (n/a)	98.20 (n/a)	83.40 (n/a)	12.64 (n/a)

dequant_2_cols_1_channels_2048_tile_1024_0

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`1531781` — 2026-01-23 18:49:24	0.05 (+2.33%)	0.05 (+22.41%)	0.05 (+32.19%)	0.04 (+7.46%)	0.01 (-3.28%)	147.80 (-6.93%)	113.14 (-18.51%)	106.50 (-24.31%)	101.40 (-2.31%)	19.61 (-8.79%)
`0f0da0b` — 2026-01-14 16:31:00	0.05 (n/a)	0.04 (n/a)	0.04 (n/a)	0.03 (n/a)	0.01 (n/a)	158.80 (n/a)	138.84 (n/a)	140.70 (n/a)	103.80 (n/a)	21.51 (n/a)

dequant_2_cols_2_channels_2048_tile_512

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`130b6ea` — 2025-12-05 21:15:49	0.05 (+17.06%)	0.04 (+8.98%)	0.05 (+21.37%)	0.03 (-11.65%)	0.01 (+301.89%)	159.70 (+13.18%)	126.32 (-5.68%)	109.00 (-17.67%)	106.80 (-14.56%)	22.48 (+280.15%)
`0a6c11c` — 2025-12-04 00:39:10	0.04 (n/a)	0.04 (n/a)	0.04 (n/a)	0.04 (n/a)	0.00 (n/a)	141.10 (n/a)	133.92 (n/a)	132.40 (n/a)	125.00 (n/a)	5.91 (n/a)

dequant_2_cols_2_channels_2048_tile_512_0

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`1531781` — 2026-01-23 18:49:24	0.05 (+14.83%)	0.04 (+10.04%)	0.05 (+21.96%)	0.03 (-12.01%)	0.01 (+143.96%)	158.30 (+13.64%)	121.38 (-6.49%)	106.20 (-17.99%)	98.20 (-12.94%)	26.13 (+141.52%)
`0f0da0b` — 2026-01-14 16:31:00	0.05 (n/a)	0.04 (n/a)	0.04 (n/a)	0.04 (n/a)	0.00 (n/a)	139.30 (n/a)	129.80 (n/a)	129.50 (n/a)	112.80 (n/a)	10.82 (n/a)

dequant_4_cols_1_channels_2048_tile_512

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`130b6ea` — 2025-12-05 21:15:49	0.06 (+20.24%)	0.05 (-3.19%)	0.05 (-6.73%)	0.03 (-20.69%)	0.01 (+194.26%)	156.10 (+26.09%)	120.86 (+7.15%)	114.90 (+7.18%)	88.30 (-16.85%)	24.40 (+214.82%)
`0a6c11c` — 2025-12-04 00:39:10	0.05 (n/a)	0.05 (n/a)	0.05 (n/a)	0.04 (n/a)	0.00 (n/a)	123.80 (n/a)	112.80 (n/a)	107.20 (n/a)	106.20 (n/a)	7.75 (n/a)

dequant_4_cols_1_channels_2048_tile_512_0

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`1531781` — 2026-01-23 18:49:24	0.06 (+28.85%)	0.05 (+16.79%)	0.05 (+23.60%)	0.03 (-12.66%)	0.01 (+157.83%)	164.80 (+14.52%)	113.62 (-10.20%)	103.20 (-19.12%)	84.80 (-22.42%)	32.55 (+130.11%)
`0f0da0b` — 2026-01-14 16:31:00	0.05 (n/a)	0.04 (n/a)	0.04 (n/a)	0.04 (n/a)	0.00 (n/a)	143.90 (n/a)	126.52 (n/a)	127.60 (n/a)	109.30 (n/a)	14.15 (n/a)

dequant_4_cols_2_channels_2048_tile_256

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`130b6ea` — 2025-12-05 21:15:49	0.05 (-6.09%)	0.04 (-8.04%)	0.04 (-12.46%)	0.03 (-0.41%)	0.01 (-20.19%)	168.10 (+0.42%)	135.74 (+7.44%)	132.20 (+14.26%)	107.40 (+6.44%)	22.91 (-12.67%)
`0a6c11c` — 2025-12-04 00:39:10	0.05 (n/a)	0.04 (n/a)	0.05 (n/a)	0.03 (n/a)	0.01 (n/a)	167.40 (n/a)	126.34 (n/a)	115.70 (n/a)	100.90 (n/a)	26.23 (n/a)

dequant_4_cols_2_channels_2048_tile_256_0

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`1531781` — 2026-01-23 18:49:24	0.07 (+42.90%)	0.05 (+60.94%)	0.06 (+84.55%)	0.03 (+50.15%)	0.01 (+27.03%)	188.40 (-33.40%)	112.72 (-39.08%)	93.80 (-45.84%)	79.70 (-29.96%)	43.58 (-36.67%)
`0f0da0b` — 2026-01-14 16:31:00	0.05 (n/a)	0.03 (n/a)	0.03 (n/a)	0.02 (n/a)	0.01 (n/a)	282.90 (n/a)	185.04 (n/a)	173.20 (n/a)	113.80 (n/a)	68.82 (n/a)

dequant_8_cols_1_channels_2048_tile_256

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`130b6ea` — 2025-12-05 21:15:49	0.04 (-13.65%)	0.04 (-13.67%)	0.04 (-18.56%)	0.03 (-13.52%)	0.00 (-23.88%)	170.40 (+15.68%)	140.12 (+15.33%)	135.80 (+22.78%)	119.60 (+15.78%)	17.94 (+1.87%)
`0a6c11c` — 2025-12-04 00:39:10	0.05 (n/a)	0.04 (n/a)	0.05 (n/a)	0.04 (n/a)	0.01 (n/a)	147.30 (n/a)	121.50 (n/a)	110.60 (n/a)	103.30 (n/a)	17.61 (n/a)

dequant_8_cols_1_channels_2048_tile_256_0

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`1531781` — 2026-01-23 18:49:24	0.04 (-5.92%)	0.04 (-11.07%)	0.04 (-12.74%)	0.03 (-9.17%)	0.00 (-4.85%)	166.90 (+10.09%)	138.66 (+12.51%)	135.90 (+14.59%)	117.60 (+6.33%)	17.94 (+10.10%)
`0f0da0b` — 2026-01-14 16:31:00	0.05 (n/a)	0.04 (n/a)	0.04 (n/a)	0.03 (n/a)	0.00 (n/a)	151.60 (n/a)	123.24 (n/a)	118.60 (n/a)	110.60 (n/a)	16.30 (n/a)

dequant_8_cols_2_channels_2048_tile_128

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`130b6ea` — 2025-12-05 21:15:49	0.05 (+17.21%)	0.03 (-1.70%)	0.03 (-3.04%)	0.03 (-16.20%)	0.01 (+95.59%)	198.20 (+19.33%)	160.00 (+4.92%)	166.20 (+3.17%)	106.50 (-14.66%)	29.86 (+87.19%)
`0a6c11c` — 2025-12-04 00:39:10	0.04 (n/a)	0.03 (n/a)	0.03 (n/a)	0.03 (n/a)	0.00 (n/a)	166.10 (n/a)	152.50 (n/a)	161.10 (n/a)	124.80 (n/a)	15.95 (n/a)

dequant_8_cols_2_channels_2048_tile_128_0

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`1531781` — 2026-01-23 18:49:24	0.05 (+10.58%)	0.04 (+11.15%)	0.04 (+32.75%)	0.02 (-15.30%)	0.01 (+23.92%)	229.90 (+18.08%)	145.60 (-7.34%)	131.00 (-24.67%)	106.20 (-9.54%)	49.03 (+42.68%)
`0f0da0b` — 2026-01-14 16:31:00	0.04 (n/a)	0.03 (n/a)	0.03 (n/a)	0.03 (n/a)	0.01 (n/a)	194.70 (n/a)	157.14 (n/a)	173.90 (n/a)	117.40 (n/a)	34.37 (n/a)

eltwise_add_1_cols_2_channels_2048_tile_2048

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`1531781` — 2026-01-23 18:49:24	0.17 (+59.76%)	0.12 (+36.17%)	0.10 (+5.69%)	0.08 (+17.86%)	0.04 (+129.37%)	155.80 (-15.19%)	109.78 (-22.78%)	120.50 (-5.34%)	73.80 (-37.40%)	34.38 (+16.64%)
`0f0da0b` — 2026-01-14 16:31:00	0.10 (n/a)	0.09 (n/a)	0.10 (n/a)	0.07 (n/a)	0.02 (n/a)	183.70 (n/a)	142.16 (n/a)	127.30 (n/a)	117.90 (n/a)	29.48 (n/a)

eltwise_add_2_cols_2_channels_2048_tile_1024

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`1531781` — 2026-01-23 18:49:24	0.15 (+24.10%)	0.13 (+23.91%)	0.13 (+30.32%)	0.08 (+16.80%)	0.03 (+31.09%)	146.70 (-14.36%)	101.20 (-18.68%)	91.70 (-23.26%)	81.40 (-19.41%)	26.00 (-6.71%)
`0f0da0b` — 2026-01-14 16:31:00	0.12 (n/a)	0.10 (n/a)	0.10 (n/a)	0.07 (n/a)	0.02 (n/a)	171.30 (n/a)	124.44 (n/a)	119.50 (n/a)	101.00 (n/a)	27.87 (n/a)

eltwise_add_4_cols_2_channels_2048_tile_512

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`1531781` — 2026-01-23 18:49:24	0.17 (+52.81%)	0.15 (+77.83%)	0.15 (+81.17%)	0.11 (+106.70%)	0.02 (+14.06%)	116.00 (-51.65%)	86.46 (-45.77%)	80.90 (-44.82%)	72.70 (-34.56%)	17.22 (-64.39%)
`0f0da0b` — 2026-01-14 16:31:00	0.11 (n/a)	0.08 (n/a)	0.08 (n/a)	0.05 (n/a)	0.02 (n/a)	239.90 (n/a)	159.44 (n/a)	146.60 (n/a)	111.10 (n/a)	48.37 (n/a)

eltwise_add_8_cols_2_channels_2048_tile_256

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`1531781` — 2026-01-23 18:49:24	0.16 (+58.53%)	0.14 (+73.02%)	0.14 (+70.95%)	0.11 (+188.94%)	0.02 (-23.24%)	109.50 (-65.38%)	90.00 (-48.14%)	86.00 (-41.50%)	75.30 (-36.93%)	13.44 (-83.58%)
`0f0da0b` — 2026-01-14 16:31:00	0.10 (n/a)	0.08 (n/a)	0.08 (n/a)	0.04 (n/a)	0.03 (n/a)	316.30 (n/a)	173.56 (n/a)	147.00 (n/a)	119.40 (n/a)	81.81 (n/a)

eltwise_mul_1_cols_2_channels_2048_tile_2048

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`1531781` — 2026-01-23 18:49:24	0.14 (+42.19%)	0.13 (+64.04%)	0.13 (+74.51%)	0.10 (+75.35%)	0.01 (+0.96%)	118.10 (-42.97%)	97.88 (-39.87%)	92.60 (-42.70%)	90.40 (-29.70%)	11.55 (-59.47%)
`0f0da0b` — 2026-01-14 16:31:00	0.10 (n/a)	0.08 (n/a)	0.08 (n/a)	0.06 (n/a)	0.01 (n/a)	207.10 (n/a)	162.78 (n/a)	161.60 (n/a)	128.60 (n/a)	28.50 (n/a)

eltwise_mul_2_cols_2_channels_2048_tile_1024

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`1531781` — 2026-01-23 18:49:24	0.13 (+44.68%)	0.12 (+43.52%)	0.13 (+47.71%)	0.09 (+35.56%)	0.02 (+80.76%)	130.00 (-26.26%)	102.10 (-29.85%)	93.30 (-32.34%)	91.10 (-30.88%)	16.53 (-8.94%)
`0f0da0b` — 2026-01-14 16:31:00	0.09 (n/a)	0.09 (n/a)	0.09 (n/a)	0.07 (n/a)	0.01 (n/a)	176.30 (n/a)	145.54 (n/a)	137.90 (n/a)	131.80 (n/a)	18.15 (n/a)

eltwise_mul_4_cols_2_channels_2048_tile_512

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`1531781` — 2026-01-23 18:49:24	0.17 (+63.62%)	0.15 (+73.87%)	0.15 (+72.82%)	0.11 (+104.66%)	0.02 (+10.66%)	114.00 (-51.16%)	86.18 (-44.46%)	81.50 (-42.12%)	74.20 (-38.88%)	15.85 (-65.75%)
`0f0da0b` — 2026-01-14 16:31:00	0.10 (n/a)	0.08 (n/a)	0.09 (n/a)	0.05 (n/a)	0.02 (n/a)	233.40 (n/a)	155.18 (n/a)	140.80 (n/a)	121.40 (n/a)	46.29 (n/a)

eltwise_mul_8_cols_2_channels_2048_tile_256

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`1531781` — 2026-01-23 18:49:24	0.13 (+46.12%)	0.11 (+59.06%)	0.12 (+66.26%)	0.08 (+48.55%)	0.02 (+55.56%)	154.80 (-32.70%)	112.28 (-36.89%)	104.60 (-39.85%)	95.50 (-31.59%)	24.14 (-26.88%)
`0f0da0b` — 2026-01-14 16:31:00	0.09 (n/a)	0.07 (n/a)	0.07 (n/a)	0.05 (n/a)	0.01 (n/a)	230.00 (n/a)	177.92 (n/a)	173.90 (n/a)	139.60 (n/a)	33.01 (n/a)

gelu_1_cols_1_channels_2048_tile_2048

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`1531781` — 2026-01-23 18:49:24	0.11 (+53.19%)	0.10 (+56.21%)	0.11 (+82.01%)	0.08 (+32.44%)	0.02 (+220.41%)	101.60 (-24.52%)	83.80 (-34.61%)	72.40 (-45.11%)	72.20 (-34.72%)	15.75 (+58.78%)
`0f0da0b` — 2026-01-14 16:31:00	0.07 (n/a)	0.06 (n/a)	0.06 (n/a)	0.06 (n/a)	0.01 (n/a)	134.60 (n/a)	128.16 (n/a)	131.90 (n/a)	110.60 (n/a)	9.92 (n/a)

gelu_1_cols_2_channels_2048_tile_1024

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`1531781` — 2026-01-23 18:49:24	0.12 (+63.09%)	0.11 (+76.43%)	0.12 (+100.14%)	0.07 (+62.57%)	0.02 (+62.52%)	110.00 (-38.48%)	80.26 (-43.34%)	69.10 (-50.00%)	66.40 (-38.69%)	18.54 (-39.34%)
`0f0da0b` — 2026-01-14 16:31:00	0.08 (n/a)	0.06 (n/a)	0.06 (n/a)	0.05 (n/a)	0.01 (n/a)	178.80 (n/a)	141.66 (n/a)	138.20 (n/a)	108.30 (n/a)	30.55 (n/a)

gelu_2_cols_1_channels_2048_tile_1024

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`1531781` — 2026-01-23 18:49:24	0.12 (+51.06%)	0.10 (+80.20%)	0.12 (+120.49%)	0.06 (+32.94%)	0.03 (+101.39%)	137.80 (-24.74%)	85.02 (-42.29%)	68.20 (-54.65%)	67.80 (-33.79%)	30.24 (+4.82%)
`0f0da0b` — 2026-01-14 16:31:00	0.08 (n/a)	0.06 (n/a)	0.05 (n/a)	0.04 (n/a)	0.01 (n/a)	183.10 (n/a)	147.32 (n/a)	150.40 (n/a)	102.40 (n/a)	28.85 (n/a)

gelu_2_cols_2_channels_2048_tile_512

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`1531781` — 2026-01-23 18:49:24	0.12 (+49.86%)	0.11 (+81.45%)	0.12 (+78.91%)	0.08 (+119.08%)	0.02 (+8.63%)	107.80 (-54.38%)	76.80 (-47.80%)	69.50 (-44.13%)	67.10 (-33.23%)	17.40 (-67.51%)
`0f0da0b` — 2026-01-14 16:31:00	0.08 (n/a)	0.06 (n/a)	0.07 (n/a)	0.03 (n/a)	0.02 (n/a)	236.30 (n/a)	147.12 (n/a)	124.40 (n/a)	100.50 (n/a)	53.56 (n/a)

gelu_4_cols_1_channels_2048_tile_512

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`1531781` — 2026-01-23 18:49:24	0.13 (+73.67%)	0.11 (+92.11%)	0.12 (+121.10%)	0.05 (+47.52%)	0.03 (+98.64%)	153.10 (-32.20%)	85.56 (-45.77%)	69.30 (-54.77%)	63.00 (-42.47%)	37.94 (-17.18%)
`0f0da0b` — 2026-01-14 16:31:00	0.07 (n/a)	0.06 (n/a)	0.05 (n/a)	0.04 (n/a)	0.02 (n/a)	225.80 (n/a)	157.78 (n/a)	153.20 (n/a)	109.50 (n/a)	45.81 (n/a)

gelu_4_cols_2_channels_2048_tile_256

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`1531781` — 2026-01-23 18:49:24	0.12 (+87.02%)	0.11 (+117.65%)	0.12 (+144.58%)	0.07 (+103.45%)	0.02 (+63.92%)	110.50 (-50.82%)	79.80 (-54.61%)	70.50 (-59.08%)	70.00 (-46.56%)	17.48 (-56.79%)
`0f0da0b` — 2026-01-14 16:31:00	0.06 (n/a)	0.05 (n/a)	0.05 (n/a)	0.04 (n/a)	0.01 (n/a)	224.70 (n/a)	175.80 (n/a)	172.30 (n/a)	131.00 (n/a)	40.46 (n/a)

gelu_8_cols_1_channels_2048_tile_256

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`1531781` — 2026-01-23 18:49:24	0.12 (+61.36%)	0.10 (+65.53%)	0.11 (+79.09%)	0.04 (-5.63%)	0.03 (+170.92%)	208.20 (+5.95%)	98.74 (-30.26%)	71.90 (-44.13%)	70.20 (-38.04%)	61.20 (+79.92%)
`0f0da0b` — 2026-01-14 16:31:00	0.07 (n/a)	0.06 (n/a)	0.06 (n/a)	0.04 (n/a)	0.01 (n/a)	196.50 (n/a)	141.58 (n/a)	128.70 (n/a)	113.30 (n/a)	34.01 (n/a)

gelu_8_cols_2_channels_2048_tile_128

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`1531781` — 2026-01-23 18:49:24	0.10 (+64.20%)	0.09 (+72.61%)	0.10 (+77.79%)	0.06 (+39.91%)	0.02 (+119.55%)	133.40 (-28.55%)	93.44 (-40.99%)	82.30 (-43.75%)	81.90 (-39.11%)	22.47 (-5.49%)
`0f0da0b` — 2026-01-14 16:31:00	0.06 (n/a)	0.05 (n/a)	0.06 (n/a)	0.04 (n/a)	0.01 (n/a)	186.70 (n/a)	158.34 (n/a)	146.30 (n/a)	134.50 (n/a)	23.77 (n/a)

gemm_1792x896x1152_64x32x48_8cols_ccolmaj

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)	Throughput (max)	Throughput (mean)	Throughput (median)	Throughput (min)	Throughput (stddev)
`1531781` — 2026-01-23 18:49:24	5.15 (+4.03%)	4.90 (+9.86%)	4.99 (+12.67%)	4.27 (+4.36%)	0.36 (+11.38%)	2201.70 (-4.17%)	1929.94 (-8.91%)	1885.00 (-11.25%)	1826.70 (-3.87%)	154.05 (+4.53%)	2025.19 (+4.03%)	1925.76 (+9.86%)	1962.51 (+12.67%)	1680.20 (+4.36%)	139.97 (+11.38%)
`0f0da0b` — 2026-01-14 16:31:00	4.95 (n/a)	4.46 (n/a)	4.43 (n/a)	4.09 (n/a)	0.32 (n/a)	2297.60 (n/a)	2118.80 (n/a)	2123.90 (n/a)	1900.20 (n/a)	147.37 (n/a)	1946.79 (n/a)	1752.94 (n/a)	1741.75 (n/a)	1610.08 (n/a)	125.67 (n/a)

gemm_192x384x64_48x96x16_4cols

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)	Throughput (max)	Throughput (mean)	Throughput (median)	Throughput (min)	Throughput (stddev)
`1531781` — 2026-01-23 18:49:24	2.36 (+52.61%)	1.85 (+39.10%)	1.78 (+41.59%)	1.32 (+14.31%)	0.39 (+132.56%)	167.80 (-12.51%)	124.24 (-26.22%)	124.40 (-29.36%)	93.70 (-34.52%)	28.11 (+35.82%)	100.67 (+52.61%)	78.94 (+39.10%)	75.88 (+41.59%)	56.25 (+14.31%)	16.73 (+132.56%)
`0f0da0b` — 2026-01-14 16:31:00	1.55 (n/a)	1.33 (n/a)	1.26 (n/a)	1.15 (n/a)	0.17 (n/a)	191.80 (n/a)	168.40 (n/a)	176.10 (n/a)	143.10 (n/a)	20.70 (n/a)	65.96 (n/a)	56.75 (n/a)	53.59 (n/a)	49.20 (n/a)	7.20 (n/a)

gemm_192x384x64_48x96x16_4cols_bcolmaj_ccolmaj

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)	Throughput (max)	Throughput (mean)	Throughput (median)	Throughput (min)	Throughput (stddev)
`1531781` — 2026-01-23 18:49:24	2.22 (+30.89%)	2.06 (+54.21%)	2.20 (+75.49%)	1.60 (+35.74%)	0.26 (+25.57%)	138.20 (-26.33%)	109.20 (-35.27%)	100.50 (-43.03%)	99.80 (-23.58%)	16.54 (-28.10%)	94.54 (+30.89%)	87.77 (+54.21%)	93.91 (+75.49%)	68.30 (+35.74%)	11.26 (+25.57%)
`0f0da0b` — 2026-01-14 16:31:00	1.69 (n/a)	1.33 (n/a)	1.25 (n/a)	1.18 (n/a)	0.21 (n/a)	187.60 (n/a)	168.70 (n/a)	176.40 (n/a)	130.60 (n/a)	23.00 (n/a)	72.23 (n/a)	56.92 (n/a)	53.51 (n/a)	50.31 (n/a)	8.97 (n/a)

gemm_2048x2048x2048_64x64x32_8_cols_0_bcolmaj_0_ccolmaj_0

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	GFLOP/s (max)	GFLOP/s (mean)	GFLOP/s (median)	GFLOP/s (min)	GFLOP/s (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`3a053d3` — 2025-11-26 01:12:16	3.72 (-0.11%)	3.55 (-2.75%)	3.56 (-2.75%)	3.36 (-5.33%)	0.12 (+76.10%)	2725.23 (-0.11%)	2602.52 (-2.75%)	2612.11 (-2.75%)	2459.89 (-5.33%)	86.27 (+76.10%)	6984.00 (+5.63%)	6608.60 (+2.91%)	6577.00 (+2.83%)	6304.00 (+0.11%)	221.87 (+87.15%)
`9e98049` — 2025-11-25 20:50:55	3.72 (n/a)	3.65 (n/a)	3.66 (n/a)	3.54 (n/a)	0.07 (n/a)	2728.26 (n/a)	2676.06 (n/a)	2686.03 (n/a)	2598.29 (n/a)	48.99 (n/a)	6612.00 (n/a)	6422.00 (n/a)	6396.00 (n/a)	6297.00 (n/a)	118.55 (n/a)

gemm_2048x2048x2048_64x64x32_8_cols_0_bcolmaj_1_ccolmaj_0

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	GFLOP/s (max)	GFLOP/s (mean)	GFLOP/s (median)	GFLOP/s (min)	GFLOP/s (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`3a053d3` — 2025-11-26 01:12:16	3.40 (-3.14%)	3.38 (-0.82%)	3.38 (-0.82%)	3.35 (+4.93%)	0.02 (-84.51%)	2493.81 (-3.14%)	2475.27 (-0.82%)	2479.06 (-0.82%)	2453.92 (+4.93%)	13.33 (-84.51%)	7001.00 (-4.70%)	6940.80 (+0.71%)	6930.00 (+0.83%)	6889.00 (+3.24%)	37.43 (-84.76%)
`9e98049` — 2025-11-25 20:50:55	3.51 (n/a)	3.40 (n/a)	3.41 (n/a)	3.19 (n/a)	0.12 (n/a)	2574.53 (n/a)	2495.72 (n/a)	2499.62 (n/a)	2338.67 (n/a)	86.02 (n/a)	7346.00 (n/a)	6892.20 (n/a)	6873.00 (n/a)	6673.00 (n/a)	245.57 (n/a)

gemm_2048x2048x2048_64x64x32_8_cols_1_bcolmaj_0_ccolmaj_0

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	GFLOP/s (max)	GFLOP/s (mean)	GFLOP/s (median)	GFLOP/s (min)	GFLOP/s (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`3a053d3` — 2025-11-26 01:12:16	3.02 (+0.52%)	2.87 (-2.05%)	2.86 (-2.24%)	2.70 (-6.16%)	0.11 (+131.66%)	2216.19 (+0.52%)	2104.44 (-2.05%)	2094.08 (-2.24%)	1980.84 (-6.16%)	80.25 (+131.67%)	8673.00 (+6.56%)	8175.60 (+2.21%)	8204.00 (+2.29%)	7752.00 (-0.51%)	314.03 (+145.36%)
`9e98049` — 2025-11-25 20:50:55	3.01 (n/a)	2.93 (n/a)	2.92 (n/a)	2.88 (n/a)	0.05 (n/a)	2204.81 (n/a)	2148.41 (n/a)	2142.13 (n/a)	2110.81 (n/a)	34.64 (n/a)	8139.00 (n/a)	7998.60 (n/a)	8020.00 (n/a)	7792.00 (n/a)	127.99 (n/a)

gemm_2048x2048x2048_64x64x64_1cols

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)	Throughput (max)	Throughput (mean)	Throughput (median)	Throughput (min)	Throughput (stddev)
`1531781` — 2026-01-23 18:49:24	0.52 (-0.75%)	0.52 (-0.07%)	0.52 (+0.01%)	0.52 (+0.22%)	0.00 (-86.34%)	48415.10 (-0.22%)	48370.52 (+0.07%)	48362.20 (-0.01%)	48344.40 (+0.76%)	29.40 (-86.25%)	355.36 (-0.75%)	355.17 (-0.07%)	355.23 (+0.01%)	354.85 (+0.22%)	0.22 (-86.34%)
`0f0da0b` — 2026-01-14 16:31:00	0.52 (n/a)	0.52 (n/a)	0.52 (n/a)	0.52 (n/a)	0.00 (n/a)	48520.80 (n/a)	48335.84 (n/a)	48365.40 (n/a)	47981.00 (n/a)	213.79 (n/a)	358.06 (n/a)	355.43 (n/a)	355.21 (n/a)	354.07 (n/a)	1.58 (n/a)

gemm_2048x2048x2048_64x64x64_2_cols_0_bcolmaj_0_ccolmaj_0

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`130b6ea` — 2025-12-05 21:15:49	1.04 (+0.94%)	1.03 (+0.23%)	1.03 (+0.09%)	1.02 (+0.07%)	0.00 (+197.51%)	24576.00 (-0.07%)	24485.82 (-0.23%)	24530.40 (-0.09%)	24266.40 (-0.94%)	111.26 (+194.35%)
`0a6c11c` — 2025-12-04 00:39:10	1.03 (n/a)	1.03 (n/a)	1.03 (n/a)	1.02 (n/a)	0.00 (n/a)	24594.30 (n/a)	24542.14 (n/a)	24551.90 (n/a)	24495.50 (n/a)	37.80 (n/a)

gemm_2048x2048x2048_64x64x64_2_cols_0_bcolmaj_0_ccolmaj_0_0

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)	Throughput (max)	Throughput (mean)	Throughput (median)	Throughput (min)	Throughput (stddev)
`a4b6ffe` — 2026-01-05 19:24:05	1.04 (+0.28%)	1.03 (-0.15%)	1.03 (+0.05%)	1.02 (-0.52%)	0.01 (+42.72%)	24719.80 (+0.52%)	24512.42 (+0.16%)	24529.20 (-0.05%)	24092.60 (-0.28%)	252.92 (+43.15%)	713.08 (+0.28%)	700.92 (-0.15%)	700.38 (+0.05%)	694.98 (-0.52%)	7.30 (+42.71%)
`50ad37c` — 2025-12-29 20:35:53	1.04 (n/a)	1.03 (n/a)	1.03 (n/a)	1.02 (n/a)	0.01 (n/a)	24592.30 (n/a)	24473.70 (n/a)	24540.90 (n/a)	24160.80 (n/a)	176.68 (n/a)	711.06 (n/a)	702.00 (n/a)	700.05 (n/a)	698.59 (n/a)	5.12 (n/a)

gemm_2048x2048x2048_64x64x64_2_cols_0_bcolmaj_1_ccolmaj_0

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`130b6ea` — 2025-12-05 21:15:49	0.99 (+1.69%)	0.97 (+0.29%)	0.97 (-0.04%)	0.97 (-0.15%)	0.01 (+434.92%)	26025.30 (+0.15%)	25839.34 (-0.29%)	25916.70 (+0.04%)	25446.00 (-1.67%)	202.05 (+425.44%)
`0a6c11c` — 2025-12-04 00:39:10	0.97 (n/a)	0.97 (n/a)	0.97 (n/a)	0.97 (n/a)	0.00 (n/a)	25986.90 (n/a)	25913.52 (n/a)	25905.70 (n/a)	25877.00 (n/a)	38.45 (n/a)

gemm_2048x2048x2048_64x64x64_2_cols_0_bcolmaj_1_ccolmaj_0_0

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)	Throughput (max)	Throughput (mean)	Throughput (median)	Throughput (min)	Throughput (stddev)
`a4b6ffe` — 2026-01-05 19:24:05	0.97 (-0.01%)	0.97 (+0.03%)	0.97 (+0.04%)	0.97 (+0.00%)	0.00 (-6.43%)	25927.90 (-0.00%)	25903.20 (-0.03%)	25898.60 (-0.04%)	25892.80 (+0.01%)	14.05 (-6.56%)	663.50 (-0.01%)	663.23 (+0.03%)	663.35 (+0.04%)	662.60 (+0.00%)	0.36 (-6.43%)
`50ad37c` — 2025-12-29 20:35:53	0.97 (n/a)	0.97 (n/a)	0.97 (n/a)	0.97 (n/a)	0.00 (n/a)	25928.10 (n/a)	25910.14 (n/a)	25908.10 (n/a)	25891.20 (n/a)	15.04 (n/a)	663.54 (n/a)	663.06 (n/a)	663.11 (n/a)	662.60 (n/a)	0.38 (n/a)

gemm_2048x2048x2048_64x64x64_2_cols_1_bcolmaj_0_ccolmaj_0

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`130b6ea` — 2025-12-05 21:15:49	0.92 (-1.02%)	0.92 (-0.66%)	0.92 (-0.10%)	0.91 (-0.78%)	0.01 (-4.54%)	27751.40 (+0.79%)	27480.74 (+0.66%)	27354.80 (+0.10%)	27320.50 (+1.03%)	182.69 (-2.84%)
`0a6c11c` — 2025-12-04 00:39:10	0.93 (n/a)	0.92 (n/a)	0.92 (n/a)	0.91 (n/a)	0.01 (n/a)	27534.80 (n/a)	27299.32 (n/a)	27326.70 (n/a)	27042.30 (n/a)	188.04 (n/a)

gemm_2048x2048x2048_64x64x64_2_cols_1_bcolmaj_0_ccolmaj_0_0

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)	Throughput (max)	Throughput (mean)	Throughput (median)	Throughput (min)	Throughput (stddev)
`a4b6ffe` — 2026-01-05 19:24:05	0.92 (+0.11%)	0.91 (-0.12%)	0.91 (-0.21%)	0.91 (-0.24%)	0.00 (+30.34%)	27723.50 (+0.24%)	27548.94 (+0.12%)	27522.90 (+0.21%)	27403.60 (-0.11%)	124.26 (+30.58%)	626.92 (+0.11%)	623.62 (-0.12%)	624.20 (-0.21%)	619.69 (-0.24%)	2.81 (+30.34%)
`50ad37c` — 2025-12-29 20:35:53	0.92 (n/a)	0.91 (n/a)	0.92 (n/a)	0.91 (n/a)	0.00 (n/a)	27657.10 (n/a)	27515.58 (n/a)	27464.70 (n/a)	27434.60 (n/a)	95.16 (n/a)	626.21 (n/a)	624.37 (n/a)	625.52 (n/a)	621.17 (n/a)	2.16 (n/a)

gemm_2048x2048x2048_64x64x64_2cols_bcolmaj

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)	Throughput (max)	Throughput (mean)	Throughput (median)	Throughput (min)	Throughput (stddev)
`1531781` — 2026-01-23 18:49:24	0.93 (+1.26%)	0.91 (+0.29%)	0.91 (+0.24%)	0.90 (-0.34%)	0.01 (+132.37%)	27826.70 (+0.34%)	27533.84 (-0.28%)	27519.90 (-0.24%)	27105.20 (-1.24%)	272.35 (+129.91%)	633.82 (+1.26%)	624.00 (+0.29%)	624.27 (+0.24%)	617.39 (-0.34%)	6.21 (+132.36%)
`0f0da0b` — 2026-01-14 16:31:00	0.92 (n/a)	0.91 (n/a)	0.91 (n/a)	0.91 (n/a)	0.00 (n/a)	27731.40 (n/a)	27612.36 (n/a)	27585.70 (n/a)	27446.50 (n/a)	118.46 (n/a)	625.94 (n/a)	622.19 (n/a)	622.78 (n/a)	619.51 (n/a)	2.67 (n/a)

gemm_2048x2048x2048_64x64x64_8_cols_0_bcolmaj_0_ccolmaj_0

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	GFLOP/s (max)	GFLOP/s (mean)	GFLOP/s (median)	GFLOP/s (min)	GFLOP/s (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`31ab24e` — 2025-11-22 02:01:56	3.65 (-1.06%)	3.59 (+0.83%)	3.63 (-0.26%)	3.45 (+2.66%)	0.07 (-44.71%)	2678.08 (-1.06%)	2632.13 (+0.83%)	2658.60 (-0.26%)	2528.31 (+2.66%)	53.86 (-44.71%)	6795.00 (-2.59%)	6529.80 (-0.92%)	6462.00 (+0.26%)	6415.00 (+1.07%)	137.14 (-45.27%)
`d9a64c5` — 2025-11-21 20:49:35	3.69 (n/a)	3.56 (n/a)	3.64 (n/a)	3.36 (n/a)	0.13 (n/a)	2706.77 (n/a)	2610.50 (n/a)	2665.61 (n/a)	2462.71 (n/a)	97.41 (n/a)	6976.00 (n/a)	6590.40 (n/a)	6445.00 (n/a)	6347.00 (n/a)	250.58 (n/a)

gemm_2048x2048x2048_64x64x64_8_cols_0_bcolmaj_0_ccolmaj_0_0

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)	Throughput (max)	Throughput (mean)	Throughput (median)	Throughput (min)	Throughput (stddev)
`d48746f` — 2026-01-12 22:17:10	3.98 (+0.20%)	3.84 (-2.95%)	3.76 (-5.27%)	3.72 (-4.97%)	0.13 (+473.97%)	6760.00 (+5.23%)	6559.50 (+3.13%)	6697.10 (+5.56%)	6321.00 (-0.20%)	218.68 (+499.40%)	2717.91 (+0.20%)	2621.44 (-2.95%)	2565.27 (-5.27%)	2541.40 (-4.97%)	88.37 (+473.97%)
`331dcca` — 2026-01-08 18:34:10	3.97 (n/a)	3.96 (n/a)	3.97 (n/a)	3.92 (n/a)	0.02 (n/a)	6423.80 (n/a)	6360.32 (n/a)	6344.30 (n/a)	6333.90 (n/a)	36.48 (n/a)	2712.39 (n/a)	2701.17 (n/a)	2707.91 (n/a)	2674.40 (n/a)	15.40 (n/a)

gemm_2048x2048x2048_64x64x64_8_cols_0_bcolmaj_1_ccolmaj_0_0

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)	Throughput (max)	Throughput (mean)	Throughput (median)	Throughput (min)	Throughput (stddev)
`d48746f` — 2026-01-12 22:17:10	3.75 (+0.28%)	3.70 (+1.25%)	3.71 (-0.21%)	3.65 (+4.77%)	0.04 (-66.90%)	6901.80 (-4.56%)	6795.46 (-1.30%)	6790.20 (+0.21%)	6715.90 (-0.28%)	67.05 (-68.40%)	2558.08 (+0.28%)	2528.33 (+1.25%)	2530.10 (-0.21%)	2489.18 (+4.77%)	24.81 (-66.90%)
`331dcca` — 2026-01-08 18:34:10	3.74 (n/a)	3.66 (n/a)	3.71 (n/a)	3.48 (n/a)	0.11 (n/a)	7231.20 (n/a)	6885.18 (n/a)	6775.70 (n/a)	6734.90 (n/a)	212.21 (n/a)	2550.87 (n/a)	2497.04 (n/a)	2535.50 (n/a)	2375.78 (n/a)	74.96 (n/a)

gemm_2048x2048x2048_64x64x64_8_cols_1_bcolmaj_0_ccolmaj_0_0

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)	Throughput (max)	Throughput (mean)	Throughput (median)	Throughput (min)	Throughput (stddev)
`d48746f` — 2026-01-12 22:17:10	3.62 (-0.77%)	3.54 (+0.14%)	3.60 (-0.04%)	3.44 (+2.17%)	0.09 (-25.77%)	7307.40 (-2.13%)	7103.46 (-0.18%)	6982.50 (+0.04%)	6948.20 (+0.77%)	182.95 (-26.55%)	2472.55 (-0.77%)	2419.80 (+0.14%)	2460.41 (-0.04%)	2351.04 (+2.17%)	61.76 (-25.77%)
`331dcca` — 2026-01-08 18:34:10	3.65 (n/a)	3.54 (n/a)	3.61 (n/a)	3.37 (n/a)	0.12 (n/a)	7466.20 (n/a)	7116.40 (n/a)	6979.60 (n/a)	6895.10 (n/a)	249.09 (n/a)	2491.61 (n/a)	2416.45 (n/a)	2461.45 (n/a)	2301.00 (n/a)	83.20 (n/a)

gemm_2048x2048x2048_64x64x64_8cols_bcolmaj_ccolmaj

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)	Throughput (max)	Throughput (mean)	Throughput (median)	Throughput (min)	Throughput (stddev)
`1531781` — 2026-01-23 18:49:24	3.46 (+1.18%)	3.42 (+2.60%)	3.45 (+2.90%)	3.29 (+2.92%)	0.07 (-25.59%)	7651.20 (-2.83%)	7368.34 (-2.57%)	7300.80 (-2.82%)	7263.70 (-1.17%)	162.53 (-28.22%)	2365.17 (+1.18%)	2332.46 (+2.60%)	2353.14 (+2.90%)	2245.39 (+2.92%)	50.15 (-25.59%)
`0f0da0b` — 2026-01-14 16:31:00	3.42 (n/a)	3.33 (n/a)	3.35 (n/a)	3.20 (n/a)	0.10 (n/a)	7874.40 (n/a)	7562.58 (n/a)	7512.80 (n/a)	7349.50 (n/a)	226.42 (n/a)	2337.56 (n/a)	2273.31 (n/a)	2286.73 (n/a)	2181.74 (n/a)	67.40 (n/a)

gemm_384x1536x1792_32x48x64_4cols_bcolmaj

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)	Throughput (max)	Throughput (mean)	Throughput (median)	Throughput (min)	Throughput (stddev)
`1531781` — 2026-01-23 18:49:24	4.35 (-0.80%)	4.24 (+3.81%)	4.33 (+9.58%)	3.85 (-0.98%)	0.22 (-4.90%)	2094.90 (+0.99%)	1905.64 (-3.69%)	1862.00 (-8.75%)	1852.00 (+0.81%)	105.88 (-3.06%)	1141.45 (-0.80%)	1111.86 (+3.81%)	1135.29 (+9.58%)	1009.08 (-0.98%)	57.51 (-4.90%)
`0f0da0b` — 2026-01-14 16:31:00	4.39 (n/a)	4.08 (n/a)	3.95 (n/a)	3.89 (n/a)	0.23 (n/a)	2074.30 (n/a)	1978.70 (n/a)	2040.50 (n/a)	1837.10 (n/a)	109.22 (n/a)	1150.67 (n/a)	1071.02 (n/a)	1036.00 (n/a)	1019.11 (n/a)	60.48 (n/a)

gemm_896x1792x640_32x64x80_8cols_ccolmaj

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)	Throughput (max)	Throughput (mean)	Throughput (median)	Throughput (min)	Throughput (stddev)
`1531781` — 2026-01-23 18:49:24	6.80 (+27.89%)	6.43 (+24.70%)	6.70 (+30.77%)	5.33 (+5.74%)	0.62 (+474.69%)	1247.10 (-5.44%)	1042.70 (-19.16%)	992.90 (-23.53%)	978.30 (-21.81%)	114.59 (+329.89%)	2100.72 (+27.89%)	1987.76 (+24.70%)	2069.85 (+30.77%)	1647.93 (+5.74%)	190.83 (+474.69%)
`0f0da0b` — 2026-01-14 16:31:00	5.32 (n/a)	5.16 (n/a)	5.12 (n/a)	5.04 (n/a)	0.11 (n/a)	1318.80 (n/a)	1289.76 (n/a)	1298.50 (n/a)	1251.20 (n/a)	26.66 (n/a)	1642.57 (n/a)	1594.05 (n/a)	1582.78 (n/a)	1558.44 (n/a)	33.21 (n/a)

layer_norm_1_cols_1_channels_2048_tile_2048

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`1531781` — 2026-01-23 18:49:24	0.11 (+88.61%)	0.09 (+61.18%)	0.08 (+46.00%)	0.07 (+61.81%)	0.02 (+189.47%)	119.70 (-38.20%)	97.44 (-36.14%)	100.50 (-31.54%)	71.50 (-46.96%)	22.67 (-5.36%)
`0f0da0b` — 2026-01-14 16:31:00	0.06 (n/a)	0.05 (n/a)	0.06 (n/a)	0.04 (n/a)	0.01 (n/a)	193.70 (n/a)	152.58 (n/a)	146.80 (n/a)	134.80 (n/a)	23.96 (n/a)

layer_norm_1_cols_2_channels_2048_tile_1024

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`1531781` — 2026-01-23 18:49:24	0.11 (+73.83%)	0.09 (+50.07%)	0.08 (+46.18%)	0.06 (+7.28%)	0.02 (+683.83%)	138.10 (-6.82%)	98.78 (-30.35%)	97.10 (-31.57%)	76.40 (-42.47%)	24.88 (+313.86%)
`0f0da0b` — 2026-01-14 16:31:00	0.06 (n/a)	0.06 (n/a)	0.06 (n/a)	0.06 (n/a)	0.00 (n/a)	148.20 (n/a)	141.82 (n/a)	141.90 (n/a)	132.80 (n/a)	6.01 (n/a)

layer_norm_2_cols_1_channels_2048_tile_1024

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`1531781` — 2026-01-23 18:49:24	0.12 (+58.00%)	0.09 (+35.21%)	0.09 (+33.88%)	0.06 (+12.88%)	0.02 (+142.40%)	140.10 (-11.44%)	98.38 (-23.32%)	94.40 (-25.26%)	68.10 (-36.71%)	26.11 (+36.70%)
`0f0da0b` — 2026-01-14 16:31:00	0.08 (n/a)	0.06 (n/a)	0.06 (n/a)	0.05 (n/a)	0.01 (n/a)	158.20 (n/a)	128.30 (n/a)	126.30 (n/a)	107.60 (n/a)	19.10 (n/a)

layer_norm_2_cols_2_channels_2048_tile_512

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`1531781` — 2026-01-23 18:49:24	0.13 (+115.82%)	0.09 (+79.97%)	0.09 (+57.15%)	0.06 (+84.94%)	0.03 (+142.79%)	130.40 (-45.94%)	91.90 (-43.54%)	93.50 (-36.39%)	62.00 (-53.66%)	25.59 (-42.10%)
`0f0da0b` — 2026-01-14 16:31:00	0.06 (n/a)	0.05 (n/a)	0.06 (n/a)	0.03 (n/a)	0.01 (n/a)	241.20 (n/a)	162.78 (n/a)	147.00 (n/a)	133.80 (n/a)	44.20 (n/a)

layer_norm_4_cols_1_channels_2048_tile_512

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`1531781` — 2026-01-23 18:49:24	0.11 (+63.95%)	0.09 (+43.09%)	0.11 (+55.71%)	0.06 (-6.72%)	0.02 (+806.48%)	141.00 (+7.22%)	91.54 (-25.97%)	77.90 (-35.78%)	73.00 (-39.07%)	28.43 (+505.85%)
`0f0da0b` — 2026-01-14 16:31:00	0.07 (n/a)	0.07 (n/a)	0.07 (n/a)	0.06 (n/a)	0.00 (n/a)	131.50 (n/a)	123.66 (n/a)	121.30 (n/a)	119.80 (n/a)	4.69 (n/a)

layer_norm_4_cols_2_channels_2048_tile_256

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`1531781` — 2026-01-23 18:49:24	0.11 (+76.69%)	0.09 (+61.27%)	0.09 (+70.26%)	0.06 (+36.41%)	0.02 (+195.74%)	148.40 (-26.72%)	103.14 (-34.74%)	87.80 (-41.27%)	73.70 (-43.44%)	32.30 (+19.79%)
`0f0da0b` — 2026-01-14 16:31:00	0.06 (n/a)	0.05 (n/a)	0.05 (n/a)	0.04 (n/a)	0.01 (n/a)	202.50 (n/a)	158.04 (n/a)	149.50 (n/a)	130.30 (n/a)	26.96 (n/a)

layer_norm_8_cols_1_channels_2048_tile_256

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`1531781` — 2026-01-23 18:49:24	0.12 (+75.09%)	0.10 (+78.44%)	0.11 (+94.76%)	0.05 (+12.92%)	0.03 (+213.89%)	153.90 (-11.45%)	90.36 (-39.81%)	77.40 (-48.67%)	67.30 (-42.87%)	35.86 (+72.31%)
`0f0da0b` — 2026-01-14 16:31:00	0.07 (n/a)	0.06 (n/a)	0.05 (n/a)	0.05 (n/a)	0.01 (n/a)	173.80 (n/a)	150.12 (n/a)	150.80 (n/a)	117.80 (n/a)	20.81 (n/a)

layer_norm_8_cols_2_channels_2048_tile_128

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`1531781` — 2026-01-23 18:49:24	0.09 (+88.05%)	0.08 (+72.40%)	0.08 (+83.22%)	0.06 (+32.78%)	0.01 (+723.36%)	140.30 (-24.69%)	106.50 (-40.67%)	97.60 (-45.44%)	92.40 (-46.80%)	19.84 (+234.33%)
`0f0da0b` — 2026-01-14 16:31:00	0.05 (n/a)	0.05 (n/a)	0.05 (n/a)	0.04 (n/a)	0.00 (n/a)	186.30 (n/a)	179.50 (n/a)	178.90 (n/a)	173.70 (n/a)	5.93 (n/a)

matrix_vector_mul_128x128_32_1col

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)
`130b6ea` — 2025-12-05 21:15:49	0.37 (+17.13%)	0.30 (+24.87%)	0.30 (+38.03%)	0.23 (+12.47%)	0.06 (+35.23%)
`0a6c11c` — 2025-12-04 00:39:10	0.32 (n/a)	0.24 (n/a)	0.22 (n/a)	0.20 (n/a)	0.04 (n/a)

matrix_vector_mul_128x128_32_1col0

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Throughput (max)	Throughput (mean)	Throughput (median)	Throughput (min)	Throughput (stddev)
`331dcca` — 2026-01-08 18:34:10	0.26 (-1.81%)	0.24 (+4.06%)	0.24 (+0.52%)	0.21 (+16.97%)	0.02 (-48.16%)	0.26 (-1.81%)	0.24 (+4.06%)	0.24 (+0.52%)	0.20 (+16.97%)	0.02 (-48.16%)
`a4b6ffe` — 2026-01-05 19:24:05	0.26 (n/a)	0.23 (n/a)	0.24 (n/a)	0.18 (n/a)	0.04 (n/a)	0.26 (n/a)	0.23 (n/a)	0.24 (n/a)	0.17 (n/a)	0.04 (n/a)

matrix_vector_mul_128x128_32tsi_128tso_1col0

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Throughput (max)	Throughput (mean)	Throughput (median)	Throughput (min)	Throughput (stddev)
`1531781` — 2026-01-23 18:49:24	0.45 (+63.20%)	0.35 (+43.38%)	0.35 (+38.29%)	0.22 (+12.34%)	0.08 (+156.17%)	0.44 (+63.20%)	0.34 (+43.38%)	0.34 (+38.29%)	0.22 (+12.34%)	0.08 (+156.17%)
`0f0da0b` — 2026-01-14 16:31:00	0.27 (n/a)	0.24 (n/a)	0.25 (n/a)	0.20 (n/a)	0.03 (n/a)	0.27 (n/a)	0.24 (n/a)	0.25 (n/a)	0.19 (n/a)	0.03 (n/a)

matrix_vector_mul_2048x8192_1_1col

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)
`130b6ea` — 2025-12-05 21:15:49	13.57 (+3.47%)	13.15 (+1.09%)	13.10 (+0.16%)	12.52 (-1.58%)	0.37 (+153.19%)
`0a6c11c` — 2025-12-04 00:39:10	13.11 (n/a)	13.01 (n/a)	13.07 (n/a)	12.72 (n/a)	0.15 (n/a)

matrix_vector_mul_2048x8192_1_1col0

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Throughput (max)	Throughput (mean)	Throughput (median)	Throughput (min)	Throughput (stddev)
`331dcca` — 2026-01-08 18:34:10	13.43 (-0.07%)	13.30 (+4.62%)	13.27 (+1.29%)	13.16 (+16.70%)	0.12 (-86.69%)	13.42 (-0.07%)	13.29 (+4.62%)	13.26 (+1.29%)	13.16 (+16.70%)	0.12 (-86.69%)
`a4b6ffe` — 2026-01-05 19:24:05	13.44 (n/a)	12.71 (n/a)	13.10 (n/a)	11.28 (n/a)	0.87 (n/a)	13.43 (n/a)	12.70 (n/a)	13.09 (n/a)	11.27 (n/a)	0.87 (n/a)

matrix_vector_mul_2048x8192_1_2col

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)
`130b6ea` — 2025-12-05 21:15:49	25.06 (+2.47%)	23.71 (+1.81%)	24.13 (+0.50%)	20.90 (+5.41%)	1.48 (-15.60%)
`0a6c11c` — 2025-12-04 00:39:10	24.46 (n/a)	23.29 (n/a)	24.01 (n/a)	19.83 (n/a)	1.75 (n/a)

matrix_vector_mul_2048x8192_1_2col0

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Throughput (max)	Throughput (mean)	Throughput (median)	Throughput (min)	Throughput (stddev)
`331dcca` — 2026-01-08 18:34:10	24.46 (-0.60%)	21.07 (-8.03%)	19.75 (-17.83%)	18.08 (+1.18%)	3.03 (+7.07%)	24.44 (-0.60%)	21.06 (-8.03%)	19.74 (-17.83%)	18.07 (+1.18%)	3.03 (+7.07%)
`a4b6ffe` — 2026-01-05 19:24:05	24.60 (n/a)	22.91 (n/a)	24.04 (n/a)	17.87 (n/a)	2.83 (n/a)	24.59 (n/a)	22.90 (n/a)	24.02 (n/a)	17.86 (n/a)	2.83 (n/a)

matrix_vector_mul_2048x8192_1_4col

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)
`130b6ea` — 2025-12-05 21:15:49	42.05 (+1.54%)	40.50 (+12.60%)	40.49 (+2.36%)	38.98 (+97.53%)	0.97 (-88.09%)
`0a6c11c` — 2025-12-04 00:39:10	41.41 (n/a)	35.97 (n/a)	39.55 (n/a)	19.73 (n/a)	8.15 (n/a)

matrix_vector_mul_2048x8192_1_4col0

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Throughput (max)	Throughput (mean)	Throughput (median)	Throughput (min)	Throughput (stddev)
`331dcca` — 2026-01-08 18:34:10	42.58 (+1.51%)	41.37 (+14.29%)	41.25 (+4.89%)	40.33 (+91.42%)	0.92 (-89.18%)	42.56 (+1.51%)	41.34 (+14.29%)	41.23 (+4.89%)	40.30 (+91.42%)	0.92 (-89.18%)
`a4b6ffe` — 2026-01-05 19:24:05	41.95 (n/a)	36.20 (n/a)	39.33 (n/a)	21.07 (n/a)	8.54 (n/a)	41.92 (n/a)	36.17 (n/a)	39.31 (n/a)	21.05 (n/a)	8.53 (n/a)

matrix_vector_mul_2048x8192_1_8col

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)
`130b6ea` — 2025-12-05 21:15:49	46.62 (+3.11%)	44.82 (+0.17%)	44.78 (+0.17%)	42.60 (-3.80%)	1.60 (+367.72%)
`0a6c11c` — 2025-12-04 00:39:10	45.21 (n/a)	44.75 (n/a)	44.70 (n/a)	44.29 (n/a)	0.34 (n/a)

matrix_vector_mul_2048x8192_1_8col0

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Throughput (max)	Throughput (mean)	Throughput (median)	Throughput (min)	Throughput (stddev)
`331dcca` — 2026-01-08 18:34:10	46.04 (+4.14%)	45.06 (+3.47%)	44.55 (+2.76%)	44.30 (+2.84%)	0.87 (+66.58%)	46.01 (+4.14%)	45.03 (+3.47%)	44.52 (+2.76%)	44.27 (+2.84%)	0.87 (+66.58%)
`a4b6ffe` — 2026-01-05 19:24:05	44.21 (n/a)	43.55 (n/a)	43.35 (n/a)	43.07 (n/a)	0.52 (n/a)	44.18 (n/a)	43.52 (n/a)	43.32 (n/a)	43.05 (n/a)	0.52 (n/a)

matrix_vector_mul_2048x8192_1tsi_1024tso_2col0

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Throughput (max)	Throughput (mean)	Throughput (median)	Throughput (min)	Throughput (stddev)
`1531781` — 2026-01-23 18:49:24	24.80 (+0.16%)	24.59 (-0.08%)	24.55 (-0.22%)	24.40 (+0.04%)	0.18 (+11.75%)	24.79 (+0.16%)	24.58 (-0.08%)	24.53 (-0.22%)	24.38 (+0.04%)	0.18 (+11.75%)
`0f0da0b` — 2026-01-14 16:31:00	24.76 (n/a)	24.61 (n/a)	24.60 (n/a)	24.39 (n/a)	0.16 (n/a)	24.75 (n/a)	24.60 (n/a)	24.58 (n/a)	24.37 (n/a)	0.16 (n/a)

matrix_vector_mul_2048x8192_1tsi_2048tso_1col0

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Throughput (max)	Throughput (mean)	Throughput (median)	Throughput (min)	Throughput (stddev)
`1531781` — 2026-01-23 18:49:24	13.54 (+0.77%)	13.34 (+3.03%)	13.45 (+0.75%)	13.10 (+15.45%)	0.21 (-77.02%)	13.53 (+0.77%)	13.33 (+3.03%)	13.44 (+0.75%)	13.09 (+15.45%)	0.21 (-77.02%)
`0f0da0b` — 2026-01-14 16:31:00	13.44 (n/a)	12.95 (n/a)	13.35 (n/a)	11.35 (n/a)	0.90 (n/a)	13.43 (n/a)	12.94 (n/a)	13.34 (n/a)	11.34 (n/a)	0.90 (n/a)

matrix_vector_mul_2048x8192_1tsi_256tso_8col0

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Throughput (max)	Throughput (mean)	Throughput (median)	Throughput (min)	Throughput (stddev)
`1531781` — 2026-01-23 18:49:24	46.81 (+1.97%)	45.91 (+2.21%)	46.38 (+2.98%)	44.81 (+2.36%)	0.89 (-7.11%)	46.79 (+1.97%)	45.88 (+2.21%)	46.35 (+2.98%)	44.79 (+2.36%)	0.89 (-7.11%)
`0f0da0b` — 2026-01-14 16:31:00	45.91 (n/a)	44.92 (n/a)	45.04 (n/a)	43.78 (n/a)	0.96 (n/a)	45.88 (n/a)	44.89 (n/a)	45.01 (n/a)	43.75 (n/a)	0.96 (n/a)

matrix_vector_mul_2048x8192_1tsi_512tso_4col0

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Throughput (max)	Throughput (mean)	Throughput (median)	Throughput (min)	Throughput (stddev)
`1531781` — 2026-01-23 18:49:24	41.49 (-0.92%)	40.81 (+0.09%)	41.13 (+0.38%)	39.95 (+0.86%)	0.65 (-24.40%)	41.46 (-0.92%)	40.79 (+0.09%)	41.10 (+0.38%)	39.93 (+0.86%)	0.65 (-24.40%)
`0f0da0b` — 2026-01-14 16:31:00	41.88 (n/a)	40.78 (n/a)	40.97 (n/a)	39.61 (n/a)	0.86 (n/a)	41.85 (n/a)	40.75 (n/a)	40.95 (n/a)	39.59 (n/a)	0.86 (n/a)

matrix_vector_mul_8192x2048_4_1col

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)
`130b6ea` — 2025-12-05 21:15:49	13.29 (+1.63%)	13.19 (+6.08%)	13.19 (+1.17%)	13.09 (+26.86%)	0.08 (-92.94%)
`0a6c11c` — 2025-12-04 00:39:10	13.08 (n/a)	12.44 (n/a)	13.03 (n/a)	10.32 (n/a)	1.07 (n/a)

matrix_vector_mul_8192x2048_4_1col0

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Throughput (max)	Throughput (mean)	Throughput (median)	Throughput (min)	Throughput (stddev)
`331dcca` — 2026-01-08 18:34:10	13.68 (+3.42%)	13.30 (+8.54%)	13.38 (+7.25%)	12.72 (+14.96%)	0.35 (-66.09%)	13.67 (+3.42%)	13.29 (+8.54%)	13.38 (+7.25%)	12.71 (+14.96%)	0.35 (-66.09%)
`a4b6ffe` — 2026-01-05 19:24:05	13.23 (n/a)	12.25 (n/a)	12.48 (n/a)	11.07 (n/a)	1.04 (n/a)	13.22 (n/a)	12.24 (n/a)	12.47 (n/a)	11.06 (n/a)	1.04 (n/a)

matrix_vector_mul_8192x2048_4_2col

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)
`130b6ea` — 2025-12-05 21:15:49	25.58 (+4.24%)	24.23 (+6.59%)	24.03 (+0.29%)	23.58 (+37.29%)	0.70 (-74.97%)
`0a6c11c` — 2025-12-04 00:39:10	24.54 (n/a)	22.73 (n/a)	23.96 (n/a)	17.18 (n/a)	2.79 (n/a)

matrix_vector_mul_8192x2048_4_2col0

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Throughput (max)	Throughput (mean)	Throughput (median)	Throughput (min)	Throughput (stddev)
`331dcca` — 2026-01-08 18:34:10	25.08 (+1.32%)	24.68 (+13.42%)	24.61 (+3.26%)	24.47 (+48.32%)	0.24 (-93.56%)	25.07 (+1.32%)	24.67 (+13.42%)	24.59 (+3.26%)	24.46 (+48.32%)	0.24 (-93.56%)
`a4b6ffe` — 2026-01-05 19:24:05	24.76 (n/a)	21.76 (n/a)	23.83 (n/a)	16.50 (n/a)	3.74 (n/a)	24.74 (n/a)	21.75 (n/a)	23.82 (n/a)	16.49 (n/a)	3.73 (n/a)

matrix_vector_mul_8192x2048_4_4col

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)
`130b6ea` — 2025-12-05 21:15:49	40.94 (-2.19%)	39.74 (-1.10%)	39.76 (-0.60%)	38.17 (-2.14%)	1.04 (+4.39%)
`0a6c11c` — 2025-12-04 00:39:10	41.86 (n/a)	40.18 (n/a)	40.00 (n/a)	39.00 (n/a)	1.00 (n/a)

matrix_vector_mul_8192x2048_4_4col0

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Throughput (max)	Throughput (mean)	Throughput (median)	Throughput (min)	Throughput (stddev)
`331dcca` — 2026-01-08 18:34:10	40.74 (-1.09%)	37.16 (-7.15%)	40.25 (+1.47%)	24.51 (-37.44%)	7.07 (+736.13%)	40.72 (-1.09%)	37.13 (-7.15%)	40.23 (+1.47%)	24.50 (-37.44%)	7.07 (+736.13%)
`a4b6ffe` — 2026-01-05 19:24:05	41.19 (n/a)	40.02 (n/a)	39.67 (n/a)	39.18 (n/a)	0.85 (n/a)	41.17 (n/a)	39.99 (n/a)	39.64 (n/a)	39.16 (n/a)	0.85 (n/a)

matrix_vector_mul_8192x2048_4_8col

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)
`130b6ea` — 2025-12-05 21:15:49	46.71 (-1.04%)	44.91 (-2.34%)	44.71 (-2.98%)	43.80 (-1.44%)	0.96 (+6.93%)
`0a6c11c` — 2025-12-04 00:39:10	47.20 (n/a)	45.98 (n/a)	46.08 (n/a)	44.44 (n/a)	0.90 (n/a)

matrix_vector_mul_8192x2048_4_8col0

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Throughput (max)	Throughput (mean)	Throughput (median)	Throughput (min)	Throughput (stddev)
`331dcca` — 2026-01-08 18:34:10	46.28 (+5.68%)	44.89 (+14.59%)	44.74 (+4.26%)	44.20 (+62.80%)	0.84 (-87.96%)	46.25 (+5.68%)	44.86 (+14.59%)	44.71 (+4.26%)	44.17 (+62.80%)	0.84 (-87.96%)
`a4b6ffe` — 2026-01-05 19:24:05	43.79 (n/a)	39.17 (n/a)	42.91 (n/a)	27.15 (n/a)	6.98 (n/a)	43.77 (n/a)	39.15 (n/a)	42.88 (n/a)	27.13 (n/a)	6.98 (n/a)

matrix_vector_mul_8192x2048_4tsi_1024tso_1col0

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Throughput (max)	Throughput (mean)	Throughput (median)	Throughput (min)	Throughput (stddev)
`1531781` — 2026-01-23 18:49:24	13.46 (+0.53%)	13.33 (+2.85%)	13.32 (+1.17%)	13.20 (+10.91%)	0.11 (-81.73%)	13.45 (+0.53%)	13.32 (+2.85%)	13.31 (+1.17%)	13.19 (+10.91%)	0.11 (-81.73%)
`0f0da0b` — 2026-01-14 16:31:00	13.39 (n/a)	12.96 (n/a)	13.17 (n/a)	11.90 (n/a)	0.62 (n/a)	13.38 (n/a)	12.95 (n/a)	13.16 (n/a)	11.89 (n/a)	0.62 (n/a)

matrix_vector_mul_8192x2048_4tsi_1024tso_2col0

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Throughput (max)	Throughput (mean)	Throughput (median)	Throughput (min)	Throughput (stddev)
`1531781` — 2026-01-23 18:49:24	25.27 (+0.55%)	24.75 (-0.75%)	24.61 (-1.28%)	24.55 (-0.37%)	0.30 (+52.22%)	25.25 (+0.55%)	24.74 (-0.75%)	24.59 (-1.28%)	24.53 (-0.37%)	0.30 (+52.22%)
`0f0da0b` — 2026-01-14 16:31:00	25.13 (n/a)	24.94 (n/a)	24.93 (n/a)	24.64 (n/a)	0.20 (n/a)	25.11 (n/a)	24.92 (n/a)	24.91 (n/a)	24.62 (n/a)	0.20 (n/a)

matrix_vector_mul_8192x2048_4tsi_1024tso_4col0

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Throughput (max)	Throughput (mean)	Throughput (median)	Throughput (min)	Throughput (stddev)
`1531781` — 2026-01-23 18:49:24	40.86 (-4.37%)	40.01 (-4.03%)	40.43 (-3.32%)	38.86 (-2.46%)	0.89 (-24.84%)	40.84 (-4.37%)	39.98 (-4.03%)	40.41 (-3.32%)	38.84 (-2.46%)	0.89 (-24.84%)
`0f0da0b` — 2026-01-14 16:31:00	42.73 (n/a)	41.69 (n/a)	41.82 (n/a)	39.85 (n/a)	1.18 (n/a)	42.70 (n/a)	41.66 (n/a)	41.80 (n/a)	39.82 (n/a)	1.18 (n/a)

matrix_vector_mul_8192x2048_4tsi_1024tso_8col0

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Throughput (max)	Throughput (mean)	Throughput (median)	Throughput (min)	Throughput (stddev)
`1531781` — 2026-01-23 18:49:24	46.08 (-1.82%)	45.11 (-1.49%)	45.49 (-0.25%)	43.89 (-1.05%)	0.97 (-6.46%)	46.05 (-1.82%)	45.08 (-1.49%)	45.46 (-0.25%)	43.87 (-1.05%)	0.97 (-6.46%)
`0f0da0b` — 2026-01-14 16:31:00	46.94 (n/a)	45.79 (n/a)	45.60 (n/a)	44.36 (n/a)	1.03 (n/a)	46.91 (n/a)	45.76 (n/a)	45.58 (n/a)	44.33 (n/a)	1.03 (n/a)

mem_copy_16_cores_2_chans_2048_tile_128_False

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`130b6ea` — 2025-12-05 21:15:49	0.07 (+22.03%)	0.06 (+27.86%)	0.06 (+35.49%)	0.05 (+15.00%)	0.01 (+49.69%)	173.30 (-13.00%)	137.94 (-21.32%)	133.60 (-26.19%)	123.40 (-18.06%)	18.40 (+8.53%)
`0a6c11c` — 2025-12-04 00:39:10	0.05 (n/a)	0.05 (n/a)	0.05 (n/a)	0.04 (n/a)	0.00 (n/a)	199.20 (n/a)	175.32 (n/a)	181.00 (n/a)	150.60 (n/a)	16.96 (n/a)

mem_copy_16_cores_2_chans_2048_tile_128_False0

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`1531781` — 2026-01-23 18:49:24	0.10 (+63.55%)	0.08 (+85.69%)	0.09 (+119.34%)	0.04 (+23.15%)	0.02 (+126.16%)	195.30 (-18.79%)	111.62 (-42.82%)	86.60 (-54.42%)	84.60 (-38.87%)	47.65 (+9.68%)
`0f0da0b` — 2026-01-14 16:31:00	0.06 (n/a)	0.04 (n/a)	0.04 (n/a)	0.03 (n/a)	0.01 (n/a)	240.50 (n/a)	195.20 (n/a)	190.00 (n/a)	138.40 (n/a)	43.45 (n/a)

mem_copy_1_cols_1_channels_2048_tile_2048

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`003b49d` — 2025-11-14 19:30:54	0.05 (-16.99%)	0.05 (-12.47%)	0.05 (-7.32%)	0.04 (-11.17%)	0.00 (-29.61%)	197.00 (+12.57%)	171.00 (+13.70%)	164.00 (+7.89%)	153.00 (+20.47%)	17.49 (-4.32%)
`0999a9a` — 2025-11-13 19:24:08	0.06 (n/a)	0.05 (n/a)	0.05 (n/a)	0.04 (n/a)	0.01 (n/a)	175.00 (n/a)	150.40 (n/a)	152.00 (n/a)	127.00 (n/a)	18.28 (n/a)

mem_copy_1_cols_2_channels_2048_tile_1024

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`003b49d` — 2025-11-14 19:30:54	0.06 (+5.56%)	0.05 (-1.59%)	0.05 (-1.39%)	0.03 (-14.07%)	0.01 (+25.03%)	263.00 (+16.37%)	166.00 (+4.53%)	144.00 (+1.41%)	126.00 (-5.26%)	49.38 (+43.17%)
`0999a9a` — 2025-11-13 19:24:08	0.06 (n/a)	0.05 (n/a)	0.05 (n/a)	0.03 (n/a)	0.01 (n/a)	226.00 (n/a)	158.80 (n/a)	142.00 (n/a)	133.00 (n/a)	34.49 (n/a)

mem_copy_1_cores_1_chans_2048_tile_2048_False

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`130b6ea` — 2025-12-05 21:15:49	0.08 (+10.93%)	0.07 (+7.79%)	0.07 (+3.55%)	0.05 (-1.73%)	0.01 (+36.73%)	150.20 (+1.76%)	121.68 (-6.36%)	125.60 (-3.38%)	97.90 (-9.85%)	18.56 (+21.98%)
`0a6c11c` — 2025-12-04 00:39:10	0.08 (n/a)	0.06 (n/a)	0.06 (n/a)	0.06 (n/a)	0.01 (n/a)	147.60 (n/a)	129.94 (n/a)	130.00 (n/a)	108.60 (n/a)	15.21 (n/a)

mem_copy_1_cores_1_chans_2048_tile_2048_False0

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`1531781` — 2026-01-23 18:49:24	0.10 (+31.19%)	0.09 (+26.03%)	0.10 (+33.11%)	0.07 (+25.01%)	0.02 (+78.79%)	118.10 (-20.04%)	96.52 (-19.61%)	84.40 (-24.91%)	82.40 (-23.77%)	18.29 (+9.09%)
`0f0da0b` — 2026-01-14 16:31:00	0.08 (n/a)	0.07 (n/a)	0.07 (n/a)	0.06 (n/a)	0.01 (n/a)	147.70 (n/a)	120.06 (n/a)	112.40 (n/a)	108.10 (n/a)	16.76 (n/a)

mem_copy_2_cols_1_channels_2048_tile_1024

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`003b49d` — 2025-11-14 19:30:54	0.06 (-15.32%)	0.05 (-17.09%)	0.05 (-23.81%)	0.04 (-15.54%)	0.01 (-18.18%)	193.00 (+18.40%)	160.80 (+20.36%)	168.00 (+31.25%)	124.00 (+18.10%)	22.92 (+11.39%)
`0999a9a` — 2025-11-13 19:24:08	0.07 (n/a)	0.06 (n/a)	0.06 (n/a)	0.05 (n/a)	0.01 (n/a)	163.00 (n/a)	133.60 (n/a)	128.00 (n/a)	105.00 (n/a)	20.58 (n/a)

mem_copy_2_cols_2_channels_2048_tile_512

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`003b49d` — 2025-11-14 19:30:54	0.06 (-23.02%)	0.05 (-6.98%)	0.05 (-2.13%)	0.04 (-15.31%)	0.01 (-33.29%)	196.00 (+18.07%)	147.00 (+6.52%)	141.00 (+2.17%)	126.00 (+29.90%)	25.27 (+8.13%)
`0999a9a` — 2025-11-13 19:24:08	0.08 (n/a)	0.06 (n/a)	0.06 (n/a)	0.05 (n/a)	0.01 (n/a)	166.00 (n/a)	138.00 (n/a)	138.00 (n/a)	97.00 (n/a)	23.37 (n/a)

mem_copy_2_cores_1_chans_2048_tile_1024_False

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`130b6ea` — 2025-12-05 21:15:49	0.10 (+31.37%)	0.08 (+29.12%)	0.08 (+29.06%)	0.07 (+41.80%)	0.01 (+0.23%)	109.90 (-29.46%)	98.30 (-23.38%)	100.80 (-22.52%)	79.40 (-23.87%)	10.54 (-46.17%)
`0a6c11c` — 2025-12-04 00:39:10	0.08 (n/a)	0.07 (n/a)	0.06 (n/a)	0.05 (n/a)	0.01 (n/a)	155.80 (n/a)	128.30 (n/a)	130.10 (n/a)	104.30 (n/a)	19.58 (n/a)

mem_copy_2_cores_1_chans_2048_tile_1024_False0

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`1531781` — 2026-01-23 18:49:24	0.12 (+103.55%)	0.09 (+81.99%)	0.09 (+69.86%)	0.07 (+83.95%)	0.02 (+187.42%)	117.10 (-45.66%)	92.20 (-43.53%)	86.90 (-41.12%)	70.40 (-50.84%)	23.32 (-23.43%)
`0f0da0b` — 2026-01-14 16:31:00	0.06 (n/a)	0.05 (n/a)	0.06 (n/a)	0.04 (n/a)	0.01 (n/a)	215.50 (n/a)	163.26 (n/a)	147.60 (n/a)	143.20 (n/a)	30.45 (n/a)

mem_copy_2_cores_2_chans_2048_tile_1024_False

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`130b6ea` — 2025-12-05 21:15:49	0.10 (+25.02%)	0.07 (+10.01%)	0.07 (+7.53%)	0.05 (-1.99%)	0.01 (+77.08%)	150.10 (+2.04%)	114.60 (-7.22%)	111.70 (-6.99%)	83.90 (-20.02%)	21.23 (+43.54%)
`0a6c11c` — 2025-12-04 00:39:10	0.08 (n/a)	0.07 (n/a)	0.07 (n/a)	0.06 (n/a)	0.01 (n/a)	147.10 (n/a)	123.52 (n/a)	120.10 (n/a)	104.90 (n/a)	14.79 (n/a)

mem_copy_2_cores_2_chans_2048_tile_1024_False0

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`1531781` — 2026-01-23 18:49:24	0.10 (+35.25%)	0.08 (+37.83%)	0.09 (+48.91%)	0.06 (+9.50%)	0.02 (+125.91%)	140.20 (-8.66%)	101.66 (-25.39%)	92.10 (-32.82%)	82.90 (-26.11%)	24.23 (+51.17%)
`0f0da0b` — 2026-01-14 16:31:00	0.07 (n/a)	0.06 (n/a)	0.06 (n/a)	0.05 (n/a)	0.01 (n/a)	153.50 (n/a)	136.26 (n/a)	137.10 (n/a)	112.20 (n/a)	16.03 (n/a)

mem_copy_4_cols_1_channels_2048_tile_512

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`003b49d` — 2025-11-14 19:30:54	0.07 (+11.97%)	0.05 (+3.09%)	0.05 (+2.10%)	0.03 (-11.35%)	0.01 (+41.65%)	229.00 (+12.81%)	156.00 (-0.51%)	143.00 (-2.05%)	117.00 (-10.69%)	38.21 (+48.71%)
`0999a9a` — 2025-11-13 19:24:08	0.06 (n/a)	0.05 (n/a)	0.05 (n/a)	0.04 (n/a)	0.01 (n/a)	203.00 (n/a)	156.80 (n/a)	146.00 (n/a)	131.00 (n/a)	25.69 (n/a)

mem_copy_4_cols_2_channels_2048_tile_256

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`003b49d` — 2025-11-14 19:30:54	0.06 (+9.02%)	0.05 (+5.59%)	0.05 (+0.00%)	0.04 (+3.63%)	0.01 (+25.48%)	193.00 (-3.50%)	147.40 (-4.66%)	147.00 (+0.00%)	122.00 (-8.27%)	25.24 (+7.48%)
`0999a9a` — 2025-11-13 19:24:08	0.06 (n/a)	0.05 (n/a)	0.05 (n/a)	0.04 (n/a)	0.01 (n/a)	200.00 (n/a)	154.60 (n/a)	147.00 (n/a)	133.00 (n/a)	23.48 (n/a)

mem_copy_4_cores_1_chans_2048_tile_512_False

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`130b6ea` — 2025-12-05 21:15:49	0.09 (-3.98%)	0.07 (+4.00%)	0.07 (+7.04%)	0.06 (+3.67%)	0.01 (-15.76%)	129.70 (-3.57%)	113.30 (-4.50%)	116.90 (-6.55%)	90.40 (+4.15%)	15.04 (-13.80%)
`0a6c11c` — 2025-12-04 00:39:10	0.09 (n/a)	0.07 (n/a)	0.07 (n/a)	0.06 (n/a)	0.01 (n/a)	134.50 (n/a)	118.64 (n/a)	125.10 (n/a)	86.80 (n/a)	17.45 (n/a)

mem_copy_4_cores_1_chans_2048_tile_512_False0

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`1531781` — 2026-01-23 18:49:24	0.12 (+51.68%)	0.10 (+55.58%)	0.11 (+85.86%)	0.06 (+8.92%)	0.03 (+163.31%)	128.30 (-8.23%)	84.56 (-32.56%)	71.60 (-46.17%)	65.70 (-34.04%)	26.42 (+57.86%)
`0f0da0b` — 2026-01-14 16:31:00	0.08 (n/a)	0.07 (n/a)	0.06 (n/a)	0.06 (n/a)	0.01 (n/a)	139.80 (n/a)	125.38 (n/a)	133.00 (n/a)	99.60 (n/a)	16.73 (n/a)

mem_copy_4_cores_2_chans_2048_tile_512_False

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`130b6ea` — 2025-12-05 21:15:49	0.08 (+9.57%)	0.07 (+16.85%)	0.07 (+19.58%)	0.06 (+24.52%)	0.01 (+1.05%)	134.50 (-19.70%)	116.64 (-14.92%)	116.80 (-16.39%)	98.00 (-8.75%)	15.71 (-25.13%)
`0a6c11c` — 2025-12-04 00:39:10	0.08 (n/a)	0.06 (n/a)	0.06 (n/a)	0.05 (n/a)	0.01 (n/a)	167.50 (n/a)	137.10 (n/a)	139.70 (n/a)	107.40 (n/a)	20.99 (n/a)

mem_copy_4_cores_2_chans_2048_tile_512_False0

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`1531781` — 2026-01-23 18:49:24	0.10 (+21.52%)	0.09 (+51.62%)	0.10 (+79.69%)	0.06 (+23.77%)	0.02 (+25.48%)	133.20 (-19.17%)	92.98 (-33.88%)	83.40 (-44.33%)	82.10 (-17.74%)	22.49 (-15.24%)
`0f0da0b` — 2026-01-14 16:31:00	0.08 (n/a)	0.06 (n/a)	0.05 (n/a)	0.05 (n/a)	0.01 (n/a)	164.80 (n/a)	140.62 (n/a)	149.80 (n/a)	99.80 (n/a)	26.53 (n/a)

mem_copy_8_cols_1_channels_2048_tile_256

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`003b49d` — 2025-11-14 19:30:54	0.05 (-13.48%)	0.05 (-17.79%)	0.05 (-15.33%)	0.03 (-25.09%)	0.01 (+6.67%)	267.00 (+33.50%)	176.80 (+24.51%)	150.00 (+18.11%)	141.00 (+15.57%)	47.41 (+61.41%)
`0999a9a` — 2025-11-13 19:24:08	0.06 (n/a)	0.06 (n/a)	0.06 (n/a)	0.04 (n/a)	0.01 (n/a)	200.00 (n/a)	142.00 (n/a)	127.00 (n/a)	122.00 (n/a)	29.37 (n/a)

mem_copy_8_cols_2_channels_2048_tile_128

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`003b49d` — 2025-11-14 19:30:54	0.05 (+13.38%)	0.05 (+23.85%)	0.05 (+37.91%)	0.05 (+30.72%)	0.00 (-56.85%)	166.00 (-23.50%)	153.40 (-20.60%)	153.00 (-27.49%)	142.00 (-11.80%)	7.63 (-70.60%)
`0999a9a` — 2025-11-13 19:24:08	0.05 (n/a)	0.04 (n/a)	0.04 (n/a)	0.04 (n/a)	0.01 (n/a)	217.00 (n/a)	193.20 (n/a)	211.00 (n/a)	161.00 (n/a)	25.96 (n/a)

mem_copy_8_cores_1_chans_2048_tile_256_False

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`130b6ea` — 2025-12-05 21:15:49	0.09 (+16.31%)	0.08 (+15.34%)	0.08 (+14.60%)	0.07 (+19.30%)	0.01 (-5.41%)	117.30 (-16.15%)	107.80 (-13.61%)	108.50 (-12.78%)	94.10 (-13.99%)	8.29 (-32.01%)
`0a6c11c` — 2025-12-04 00:39:10	0.07 (n/a)	0.07 (n/a)	0.07 (n/a)	0.06 (n/a)	0.01 (n/a)	139.90 (n/a)	124.78 (n/a)	124.40 (n/a)	109.40 (n/a)	12.20 (n/a)

mem_copy_8_cores_1_chans_2048_tile_256_False0

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`1531781` — 2026-01-23 18:49:24	0.09 (+46.69%)	0.08 (+57.18%)	0.09 (+64.02%)	0.06 (+94.45%)	0.01 (+6.58%)	131.30 (-48.59%)	99.78 (-38.77%)	88.70 (-39.00%)	86.90 (-31.84%)	18.90 (-64.18%)
`0f0da0b` — 2026-01-14 16:31:00	0.06 (n/a)	0.05 (n/a)	0.06 (n/a)	0.03 (n/a)	0.01 (n/a)	255.40 (n/a)	162.96 (n/a)	145.40 (n/a)	127.50 (n/a)	52.78 (n/a)

mem_copy_8_cores_2_chans_2048_tile_256_False

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`130b6ea` — 2025-12-05 21:15:49	0.09 (+12.61%)	0.07 (-2.24%)	0.07 (-8.85%)	0.06 (+5.13%)	0.01 (+35.50%)	136.50 (-4.88%)	120.86 (+2.96%)	125.10 (+9.64%)	91.60 (-11.15%)	15.72 (+10.46%)
`0a6c11c` — 2025-12-04 00:39:10	0.08 (n/a)	0.07 (n/a)	0.07 (n/a)	0.06 (n/a)	0.01 (n/a)	143.50 (n/a)	117.38 (n/a)	114.10 (n/a)	103.10 (n/a)	14.23 (n/a)

mem_copy_8_cores_2_chans_2048_tile_256_False0

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`1531781` — 2026-01-23 18:49:24	0.10 (+63.97%)	0.09 (+58.77%)	0.09 (+53.52%)	0.06 (+50.07%)	0.02 (+70.34%)	144.40 (-33.36%)	97.70 (-36.49%)	88.10 (-34.89%)	80.00 (-39.02%)	26.35 (-27.55%)
`0f0da0b` — 2026-01-14 16:31:00	0.06 (n/a)	0.06 (n/a)	0.06 (n/a)	0.04 (n/a)	0.01 (n/a)	216.70 (n/a)	153.84 (n/a)	135.30 (n/a)	131.20 (n/a)	36.36 (n/a)

mha

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`130b6ea` — 2025-12-05 21:15:49	0.18 (+0.16%)	0.18 (+0.19%)	0.18 (+0.12%)	0.18 (+0.32%)	0.00 (-52.24%)	47253.90 (-0.32%)	47213.70 (-0.19%)	47210.90 (-0.12%)	47167.70 (-0.16%)	28.71 (-52.47%)
`0a6c11c` — 2025-12-04 00:39:10	0.18 (n/a)	0.18 (n/a)	0.18 (n/a)	0.18 (n/a)	0.00 (n/a)	47405.00 (n/a)	47301.72 (n/a)	47265.90 (n/a)	47241.80 (n/a)	60.40 (n/a)

mha0

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`1531781` — 2026-01-23 18:49:24	0.18 (-0.08%)	0.18 (+0.22%)	0.18 (+0.32%)	0.18 (+0.35%)	0.00 (-42.58%)	47367.50 (-0.34%)	47232.68 (-0.22%)	47232.00 (-0.32%)	47073.00 (+0.08%)	105.45 (-42.71%)
`0f0da0b` — 2026-01-14 16:31:00	0.18 (n/a)	0.18 (n/a)	0.18 (n/a)	0.18 (n/a)	0.00 (n/a)	47531.20 (n/a)	47335.18 (n/a)	47384.30 (n/a)	47037.20 (n/a)	184.08 (n/a)

relu_1_cols_1_channels_2048_tile_2048

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`1531781` — 2026-01-23 18:49:24	0.11 (+69.81%)	0.09 (+59.10%)	0.10 (+75.49%)	0.05 (+26.24%)	0.03 (+206.46%)	152.90 (-20.82%)	101.62 (-32.81%)	82.20 (-43.00%)	73.00 (-41.13%)	35.75 (+36.67%)
`0f0da0b` — 2026-01-14 16:31:00	0.07 (n/a)	0.06 (n/a)	0.06 (n/a)	0.04 (n/a)	0.01 (n/a)	193.10 (n/a)	151.24 (n/a)	144.20 (n/a)	124.00 (n/a)	26.16 (n/a)

relu_2_cols_1_channels_2048_tile_1024

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`1531781` — 2026-01-23 18:49:24	0.10 (+60.88%)	0.08 (+49.18%)	0.09 (+55.13%)	0.05 (+8.09%)	0.02 (+316.89%)	150.60 (-7.44%)	101.26 (-30.06%)	91.30 (-35.57%)	83.00 (-37.87%)	28.02 (+146.84%)
`0f0da0b` — 2026-01-14 16:31:00	0.06 (n/a)	0.06 (n/a)	0.06 (n/a)	0.05 (n/a)	0.00 (n/a)	162.70 (n/a)	144.78 (n/a)	141.70 (n/a)	133.60 (n/a)	11.35 (n/a)

relu_4_cols_1_channels_2048_tile_512

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`1531781` — 2026-01-23 18:49:24	0.10 (+36.44%)	0.09 (+44.22%)	0.09 (+49.00%)	0.06 (+25.29%)	0.02 (+65.69%)	135.80 (-20.21%)	96.20 (-29.75%)	86.30 (-32.84%)	83.10 (-26.72%)	22.46 (-2.41%)
`0f0da0b` — 2026-01-14 16:31:00	0.07 (n/a)	0.06 (n/a)	0.06 (n/a)	0.05 (n/a)	0.01 (n/a)	170.20 (n/a)	136.94 (n/a)	128.50 (n/a)	113.40 (n/a)	23.02 (n/a)

relu_8_cols_1_channels_2048_tile_256

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`1531781` — 2026-01-23 18:49:24	0.11 (+58.81%)	0.09 (+71.41%)	0.10 (+74.46%)	0.05 (+55.08%)	0.02 (+79.07%)	154.40 (-35.51%)	93.32 (-40.67%)	78.40 (-42.65%)	75.80 (-37.04%)	34.24 (-28.56%)
`0f0da0b` — 2026-01-14 16:31:00	0.07 (n/a)	0.06 (n/a)	0.06 (n/a)	0.03 (n/a)	0.01 (n/a)	239.40 (n/a)	157.28 (n/a)	136.70 (n/a)	120.40 (n/a)	47.93 (n/a)

rms_norm_1_cols_1_channels_2048_tile_2048

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`1531781` — 2026-01-23 18:49:24	0.09 (+27.92%)	0.08 (+35.45%)	0.09 (+45.35%)	0.06 (+20.87%)	0.01 (+35.02%)	131.50 (-17.30%)	100.70 (-25.92%)	92.40 (-31.20%)	90.90 (-21.77%)	17.35 (-10.74%)
`0f0da0b` — 2026-01-14 16:31:00	0.07 (n/a)	0.06 (n/a)	0.06 (n/a)	0.05 (n/a)	0.01 (n/a)	159.00 (n/a)	135.94 (n/a)	134.30 (n/a)	116.20 (n/a)	19.43 (n/a)

rms_norm_1_cols_2_channels_2048_tile_1024

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`1531781` — 2026-01-23 18:49:24	0.12 (+67.22%)	0.10 (+57.83%)	0.11 (+74.39%)	0.06 (-3.10%)	0.03 (+484.78%)	145.20 (+3.20%)	86.54 (-31.66%)	72.10 (-42.64%)	70.50 (-40.25%)	32.80 (+266.62%)
`0f0da0b` — 2026-01-14 16:31:00	0.07 (n/a)	0.06 (n/a)	0.07 (n/a)	0.06 (n/a)	0.00 (n/a)	140.70 (n/a)	126.64 (n/a)	125.70 (n/a)	118.00 (n/a)	8.95 (n/a)

rms_norm_2_cols_1_channels_2048_tile_1024

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`1531781` — 2026-01-23 18:49:24	0.09 (+26.55%)	0.09 (+40.18%)	0.09 (+47.01%)	0.06 (+19.23%)	0.01 (+54.24%)	129.60 (-16.12%)	98.22 (-28.12%)	90.90 (-31.96%)	89.80 (-20.95%)	17.55 (+0.58%)
`0f0da0b` — 2026-01-14 16:31:00	0.07 (n/a)	0.06 (n/a)	0.06 (n/a)	0.05 (n/a)	0.01 (n/a)	154.50 (n/a)	136.64 (n/a)	133.60 (n/a)	113.60 (n/a)	17.45 (n/a)

rms_norm_2_cols_2_channels_2048_tile_512

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`1531781` — 2026-01-23 18:49:24	0.13 (+67.61%)	0.11 (+72.13%)	0.11 (+66.55%)	0.09 (+90.95%)	0.02 (+33.97%)	93.50 (-47.62%)	74.32 (-42.73%)	73.10 (-39.98%)	62.40 (-40.34%)	11.60 (-59.00%)
`0f0da0b` — 2026-01-14 16:31:00	0.08 (n/a)	0.07 (n/a)	0.07 (n/a)	0.05 (n/a)	0.01 (n/a)	178.50 (n/a)	129.78 (n/a)	121.80 (n/a)	104.60 (n/a)	28.30 (n/a)

rms_norm_4_cols_1_channels_2048_tile_512

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`1531781` — 2026-01-23 18:49:24	0.10 (+60.49%)	0.09 (+76.01%)	0.09 (+72.18%)	0.06 (+64.97%)	0.01 (+31.34%)	130.70 (-39.41%)	97.48 (-43.89%)	90.20 (-41.92%)	86.00 (-37.68%)	18.66 (-50.69%)
`0f0da0b` — 2026-01-14 16:31:00	0.06 (n/a)	0.05 (n/a)	0.05 (n/a)	0.04 (n/a)	0.01 (n/a)	215.70 (n/a)	173.74 (n/a)	155.30 (n/a)	138.00 (n/a)	37.85 (n/a)

rms_norm_4_cols_2_channels_2048_tile_256

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`1531781` — 2026-01-23 18:49:24	0.11 (+60.28%)	0.10 (+80.30%)	0.11 (+105.91%)	0.07 (+52.87%)	0.02 (+90.94%)	125.50 (-34.60%)	83.90 (-43.66%)	73.50 (-51.42%)	72.90 (-37.64%)	23.26 (-20.88%)
`0f0da0b` — 2026-01-14 16:31:00	0.07 (n/a)	0.06 (n/a)	0.05 (n/a)	0.04 (n/a)	0.01 (n/a)	191.90 (n/a)	148.92 (n/a)	151.30 (n/a)	116.90 (n/a)	29.40 (n/a)

rms_norm_8_cols_1_channels_2048_tile_256

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`1531781` — 2026-01-23 18:49:24	0.09 (+27.54%)	0.08 (+37.27%)	0.09 (+53.73%)	0.05 (+11.36%)	0.01 (+87.66%)	154.30 (-10.24%)	106.66 (-25.46%)	94.80 (-34.98%)	94.00 (-21.60%)	26.64 (+33.21%)
`0f0da0b` — 2026-01-14 16:31:00	0.07 (n/a)	0.06 (n/a)	0.06 (n/a)	0.05 (n/a)	0.01 (n/a)	171.90 (n/a)	143.10 (n/a)	145.80 (n/a)	119.90 (n/a)	20.00 (n/a)

rms_norm_8_cols_2_channels_2048_tile_128

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`1531781` — 2026-01-23 18:49:24	0.09 (+38.93%)	0.08 (+46.69%)	0.09 (+73.48%)	0.05 (+5.47%)	0.02 (+140.19%)	167.70 (-5.15%)	109.36 (-28.83%)	91.60 (-42.35%)	88.90 (-28.02%)	33.52 (+68.39%)
`0f0da0b` — 2026-01-14 16:31:00	0.07 (n/a)	0.05 (n/a)	0.05 (n/a)	0.05 (n/a)	0.01 (n/a)	176.80 (n/a)	153.66 (n/a)	158.90 (n/a)	123.50 (n/a)	19.91 (n/a)

rope_1_cols_2_channels_4096_tile_4096_0

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`1b2d230` — 2026-01-13 22:38:03	0.19 (-21.66%)	0.17 (-8.53%)	0.18 (+2.09%)	0.14 (-6.71%)	0.02 (-38.07%)	174.10 (+7.20%)	150.06 (+8.11%)	139.60 (-2.04%)	129.80 (+27.63%)	20.36 (-10.65%)
`d48746f` — 2026-01-12 22:17:10	0.24 (n/a)	0.18 (n/a)	0.17 (n/a)	0.15 (n/a)	0.04 (n/a)	162.40 (n/a)	138.80 (n/a)	142.50 (n/a)	101.70 (n/a)	22.79 (n/a)

rope_1c_32rows_512cols_32arows_0m

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`1531781` — 2026-01-23 18:49:24	1.10 (+30.53%)	0.99 (+42.49%)	1.05 (+62.73%)	0.73 (+21.76%)	0.15 (+50.52%)	133.80 (-17.86%)	101.54 (-29.34%)	93.30 (-38.58%)	89.10 (-23.39%)	18.60 (-3.62%)
`0f0da0b` — 2026-01-14 16:31:00	0.85 (n/a)	0.69 (n/a)	0.65 (n/a)	0.60 (n/a)	0.10 (n/a)	162.90 (n/a)	143.70 (n/a)	151.90 (n/a)	116.30 (n/a)	19.30 (n/a)

rope_1c_32rows_512cols_8arows_0m

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`1531781` — 2026-01-23 18:49:24	0.56 (-21.50%)	0.52 (-8.95%)	0.53 (-6.60%)	0.48 (+7.05%)	0.03 (-70.73%)	153.70 (-6.57%)	140.84 (+7.27%)	140.10 (+7.03%)	132.50 (+27.40%)	8.40 (-65.00%)
`0f0da0b` — 2026-01-14 16:31:00	0.71 (n/a)	0.58 (n/a)	0.56 (n/a)	0.45 (n/a)	0.10 (n/a)	164.50 (n/a)	131.30 (n/a)	130.90 (n/a)	104.00 (n/a)	24.00 (n/a)

rope_2_cols_2_channels_4096_tile_2048_0

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`1b2d230` — 2026-01-13 22:38:03	0.21 (-0.56%)	0.17 (-10.62%)	0.16 (-13.58%)	0.13 (-13.33%)	0.03 (+17.05%)	183.70 (+15.39%)	151.76 (+12.77%)	155.20 (+15.73%)	117.30 (+0.60%)	24.24 (+35.73%)
`d48746f` — 2026-01-12 22:17:10	0.21 (n/a)	0.19 (n/a)	0.18 (n/a)	0.15 (n/a)	0.02 (n/a)	159.20 (n/a)	134.58 (n/a)	134.10 (n/a)	116.60 (n/a)	17.86 (n/a)

rope_2c_32rows_512cols_32arows_0m

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`1531781` — 2026-01-23 18:49:24	1.25 (+43.66%)	0.97 (+36.64%)	0.91 (+36.32%)	0.63 (+2.79%)	0.25 (+150.75%)	157.00 (-2.67%)	108.10 (-23.42%)	108.30 (-26.68%)	78.70 (-30.42%)	31.09 (+71.34%)
`0f0da0b` — 2026-01-14 16:31:00	0.87 (n/a)	0.71 (n/a)	0.67 (n/a)	0.61 (n/a)	0.10 (n/a)	161.30 (n/a)	141.16 (n/a)	147.70 (n/a)	113.10 (n/a)	18.15 (n/a)

rope_2c_32rows_512cols_8arows_0m

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`1531781` — 2026-01-23 18:49:24	0.63 (-10.72%)	0.57 (+15.63%)	0.57 (+20.61%)	0.49 (+35.05%)	0.05 (-60.20%)	150.60 (-25.96%)	130.58 (-17.08%)	130.40 (-17.05%)	116.80 (+12.09%)	12.57 (-65.56%)
`0f0da0b` — 2026-01-14 16:31:00	0.71 (n/a)	0.49 (n/a)	0.47 (n/a)	0.36 (n/a)	0.13 (n/a)	203.40 (n/a)	157.48 (n/a)	157.20 (n/a)	104.20 (n/a)	36.49 (n/a)

rope_4_cols_2_channels_4096_tile_1024_0

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`1b2d230` — 2026-01-13 22:38:03	0.22 (+18.92%)	0.17 (+6.24%)	0.17 (+5.57%)	0.14 (-0.50%)	0.03 (+90.77%)	171.10 (+0.47%)	147.40 (-4.39%)	148.10 (-5.31%)	110.10 (-15.89%)	23.59 (+60.67%)
`d48746f` — 2026-01-12 22:17:10	0.19 (n/a)	0.16 (n/a)	0.16 (n/a)	0.14 (n/a)	0.02 (n/a)	170.30 (n/a)	154.16 (n/a)	156.40 (n/a)	130.90 (n/a)	14.69 (n/a)

rope_8_cols_2_channels_4096_tile_512_0

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`1b2d230` — 2026-01-13 22:38:03	0.17 (-1.30%)	0.15 (+6.63%)	0.15 (-8.78%)	0.14 (+39.68%)	0.01 (-67.35%)	174.90 (-28.41%)	163.42 (-10.36%)	167.40 (+9.63%)	146.50 (+1.31%)	11.19 (-76.17%)
`d48746f` — 2026-01-12 22:17:10	0.17 (n/a)	0.14 (n/a)	0.16 (n/a)	0.10 (n/a)	0.03 (n/a)	244.30 (n/a)	182.30 (n/a)	152.70 (n/a)	144.60 (n/a)	46.98 (n/a)

rope_8c_32rows_512cols_32arows_0m

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`1531781` — 2026-01-23 18:49:24	1.34 (+62.86%)	0.89 (+47.92%)	0.85 (+53.70%)	0.60 (+49.43%)	0.27 (+64.12%)	163.50 (-33.07%)	117.38 (-32.11%)	115.50 (-34.93%)	73.60 (-38.56%)	32.08 (-33.78%)
`0f0da0b` — 2026-01-14 16:31:00	0.82 (n/a)	0.60 (n/a)	0.55 (n/a)	0.40 (n/a)	0.16 (n/a)	244.30 (n/a)	172.90 (n/a)	177.50 (n/a)	119.80 (n/a)	48.44 (n/a)

rope_8c_32rows_512cols_8arows_0m

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`1531781` — 2026-01-23 18:49:24	0.54 (-10.87%)	0.52 (+6.74%)	0.53 (+12.23%)	0.48 (+19.79%)	0.03 (-67.26%)	152.90 (-16.49%)	143.32 (-8.02%)	139.70 (-10.91%)	135.50 (+12.17%)	7.46 (-68.88%)
`0f0da0b` — 2026-01-14 16:31:00	0.61 (n/a)	0.48 (n/a)	0.47 (n/a)	0.40 (n/a)	0.08 (n/a)	183.10 (n/a)	155.82 (n/a)	156.80 (n/a)	120.80 (n/a)	23.99 (n/a)

sigmoid_1_cols_1_channels_2048_tile_2048

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`1531781` — 2026-01-23 18:49:24	0.12 (+53.53%)	0.10 (+54.49%)	0.12 (+106.52%)	0.05 (-0.16%)	0.03 (+144.85%)	153.20 (+0.13%)	90.92 (-30.39%)	68.60 (-51.55%)	68.30 (-34.89%)	36.80 (+58.03%)
`0f0da0b` — 2026-01-14 16:31:00	0.08 (n/a)	0.06 (n/a)	0.06 (n/a)	0.05 (n/a)	0.01 (n/a)	153.00 (n/a)	130.62 (n/a)	141.60 (n/a)	104.90 (n/a)	23.29 (n/a)

sigmoid_2_cols_1_channels_2048_tile_1024

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`1531781` — 2026-01-23 18:49:24	0.12 (+73.69%)	0.10 (+60.60%)	0.12 (+98.01%)	0.06 (+2.86%)	0.03 (+394.96%)	140.80 (-2.83%)	89.84 (-32.71%)	68.90 (-49.52%)	68.10 (-42.43%)	32.04 (+163.16%)
`0f0da0b` — 2026-01-14 16:31:00	0.07 (n/a)	0.06 (n/a)	0.06 (n/a)	0.06 (n/a)	0.01 (n/a)	144.90 (n/a)	133.52 (n/a)	136.50 (n/a)	118.30 (n/a)	12.18 (n/a)

sigmoid_4_cols_1_channels_2048_tile_512

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`1531781` — 2026-01-23 18:49:24	0.10 (+30.15%)	0.08 (+37.20%)	0.09 (+57.41%)	0.05 (+2.94%)	0.02 (+110.00%)	169.40 (-2.87%)	106.58 (-23.47%)	87.40 (-36.48%)	84.00 (-23.15%)	36.37 (+56.68%)
`0f0da0b` — 2026-01-14 16:31:00	0.07 (n/a)	0.06 (n/a)	0.06 (n/a)	0.05 (n/a)	0.01 (n/a)	174.40 (n/a)	139.26 (n/a)	137.60 (n/a)	109.30 (n/a)	23.21 (n/a)

sigmoid_8_cols_1_channels_2048_tile_256

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`1531781` — 2026-01-23 18:49:24	0.12 (+109.66%)	0.10 (+84.35%)	0.12 (+117.89%)	0.05 (+3.89%)	0.03 (+649.09%)	157.30 (-3.73%)	89.74 (-39.88%)	68.70 (-54.11%)	65.90 (-52.35%)	38.85 (+251.55%)
`0f0da0b` — 2026-01-14 16:31:00	0.06 (n/a)	0.06 (n/a)	0.05 (n/a)	0.05 (n/a)	0.00 (n/a)	163.40 (n/a)	149.28 (n/a)	149.70 (n/a)	138.30 (n/a)	11.05 (n/a)

silu_1_cols_1_channels_2048_tile_2048

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`1531781` — 2026-01-23 18:49:24	0.10 (+44.80%)	0.08 (+33.22%)	0.10 (+66.93%)	0.06 (-1.08%)	0.02 (+301.40%)	143.30 (+1.06%)	104.88 (-20.01%)	82.20 (-40.13%)	80.60 (-30.93%)	32.24 (+177.47%)
`0f0da0b` — 2026-01-14 16:31:00	0.07 (n/a)	0.06 (n/a)	0.06 (n/a)	0.06 (n/a)	0.01 (n/a)	141.80 (n/a)	131.12 (n/a)	137.30 (n/a)	116.70 (n/a)	11.62 (n/a)

silu_2_cols_1_channels_2048_tile_1024

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`1531781` — 2026-01-23 18:49:24	0.10 (+34.12%)	0.08 (+47.27%)	0.10 (+69.15%)	0.06 (+39.46%)	0.02 (+73.50%)	138.90 (-28.29%)	103.10 (-30.67%)	83.20 (-40.91%)	81.60 (-25.41%)	28.30 (-8.88%)
`0f0da0b` — 2026-01-14 16:31:00	0.07 (n/a)	0.06 (n/a)	0.06 (n/a)	0.04 (n/a)	0.01 (n/a)	193.70 (n/a)	148.70 (n/a)	140.80 (n/a)	109.40 (n/a)	31.06 (n/a)

silu_4_cols_1_channels_2048_tile_512

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`1531781` — 2026-01-23 18:49:24	0.12 (+42.12%)	0.11 (+61.07%)	0.12 (+92.74%)	0.06 (+8.38%)	0.03 (+104.00%)	137.40 (-7.72%)	83.36 (-34.98%)	70.00 (-48.07%)	68.10 (-29.65%)	30.27 (+32.68%)
`0f0da0b` — 2026-01-14 16:31:00	0.08 (n/a)	0.07 (n/a)	0.06 (n/a)	0.06 (n/a)	0.01 (n/a)	148.90 (n/a)	128.20 (n/a)	134.80 (n/a)	96.80 (n/a)	22.82 (n/a)

silu_8_cols_1_channels_2048_tile_256

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`1531781` — 2026-01-23 18:49:24	0.12 (+35.33%)	0.09 (+59.58%)	0.09 (+78.47%)	0.07 (+60.20%)	0.02 (-1.76%)	109.80 (-37.58%)	89.20 (-38.82%)	87.90 (-43.98%)	71.10 (-26.09%)	14.50 (-53.17%)
`0f0da0b` — 2026-01-14 16:31:00	0.09 (n/a)	0.06 (n/a)	0.05 (n/a)	0.05 (n/a)	0.02 (n/a)	175.90 (n/a)	145.80 (n/a)	156.90 (n/a)	96.20 (n/a)	30.97 (n/a)

softmax_1_cols_2_channels_4096_tile_2048

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`1531781` — 2026-01-23 18:49:24	0.21 (+20.19%)	0.16 (+13.84%)	0.16 (+8.19%)	0.12 (+26.61%)	0.04 (+17.18%)	135.50 (-20.99%)	104.68 (-12.65%)	102.70 (-7.56%)	76.50 (-16.85%)	22.55 (-26.21%)
`0f0da0b` — 2026-01-14 16:31:00	0.18 (n/a)	0.14 (n/a)	0.15 (n/a)	0.10 (n/a)	0.03 (n/a)	171.50 (n/a)	119.84 (n/a)	111.10 (n/a)	92.00 (n/a)	30.56 (n/a)

softmax_2_cols_2_channels_4096_tile_1024

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`1531781` — 2026-01-23 18:49:24	0.22 (+40.26%)	0.18 (+43.33%)	0.20 (+64.08%)	0.13 (+24.34%)	0.04 (+68.98%)	127.80 (-19.62%)	94.74 (-28.95%)	83.40 (-39.08%)	73.90 (-28.67%)	24.46 (-5.31%)
`0f0da0b` — 2026-01-14 16:31:00	0.16 (n/a)	0.13 (n/a)	0.12 (n/a)	0.10 (n/a)	0.03 (n/a)	159.00 (n/a)	133.34 (n/a)	136.90 (n/a)	103.60 (n/a)	25.83 (n/a)

softmax_2_cols_2_channels_4096_tile_512

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`1531781` — 2026-01-23 18:49:24	0.24 (+85.21%)	0.17 (+69.97%)	0.15 (+54.16%)	0.11 (+37.21%)	0.05 (+203.82%)	148.10 (-27.12%)	102.78 (-37.88%)	106.70 (-35.10%)	69.60 (-45.96%)	31.81 (+15.98%)
`0f0da0b` — 2026-01-14 16:31:00	0.13 (n/a)	0.10 (n/a)	0.10 (n/a)	0.08 (n/a)	0.02 (n/a)	203.20 (n/a)	165.46 (n/a)	164.40 (n/a)	128.80 (n/a)	27.43 (n/a)

swiglu

No metrics available.

swiglu_decode_1x2048x2048

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`130b6ea` — 2025-12-05 21:15:49	0.00 (-15.46%)	0.00 (-27.20%)	0.00 (-28.17%)	0.00 (-33.90%)	0.00 (+1834.83%)	5187.20 (+51.28%)	4720.84 (+38.38%)	4750.90 (+39.22%)	4015.10 (+18.29%)	390.23 (+3298.45%)
`0a6c11c` — 2025-12-04 00:39:10	0.00 (n/a)	0.00 (n/a)	0.00 (n/a)	0.00 (n/a)	0.00 (n/a)	3428.80 (n/a)	3411.62 (n/a)	3412.60 (n/a)	3394.20 (n/a)	11.48 (n/a)

swiglu_decode_1x2048x2048_0

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`1531781` — 2026-01-23 18:49:24	0.00 (+28.14%)	0.00 (+20.30%)	0.00 (+10.00%)	0.00 (+11.66%)	0.00 (+124.73%)	5111.40 (-10.44%)	4324.88 (-14.91%)	4736.10 (-9.09%)	3410.70 (-21.96%)	843.97 (+53.51%)
`0f0da0b` — 2026-01-14 16:31:00	0.00 (n/a)	0.00 (n/a)	0.00 (n/a)	0.00 (n/a)	0.00 (n/a)	5707.20 (n/a)	5083.00 (n/a)	5209.60 (n/a)	4370.50 (n/a)	549.78 (n/a)

tanh_1_cols_1_channels_2048_tile_2048

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`1531781` — 2026-01-23 18:49:24	0.10 (+42.65%)	0.08 (+33.21%)	0.07 (+27.20%)	0.06 (+13.46%)	0.02 (+130.64%)	140.80 (-11.89%)	109.14 (-23.00%)	115.90 (-21.37%)	85.60 (-29.89%)	23.52 (+37.00%)
`0f0da0b` — 2026-01-14 16:31:00	0.07 (n/a)	0.06 (n/a)	0.06 (n/a)	0.05 (n/a)	0.01 (n/a)	159.80 (n/a)	141.74 (n/a)	147.40 (n/a)	122.10 (n/a)	17.17 (n/a)

tanh_2_cols_1_channels_2048_tile_1024

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`1531781` — 2026-01-23 18:49:24	0.11 (+41.27%)	0.09 (+40.30%)	0.09 (+45.94%)	0.08 (+28.08%)	0.01 (+83.15%)	103.30 (-21.98%)	88.02 (-28.44%)	87.80 (-31.46%)	77.80 (-29.21%)	9.64 (+2.40%)
`0f0da0b` — 2026-01-14 16:31:00	0.07 (n/a)	0.07 (n/a)	0.06 (n/a)	0.06 (n/a)	0.01 (n/a)	132.40 (n/a)	123.00 (n/a)	128.10 (n/a)	109.90 (n/a)	9.41 (n/a)

tanh_4_cols_1_channels_2048_tile_512

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`1531781` — 2026-01-23 18:49:24	0.10 (+48.49%)	0.09 (+56.51%)	0.09 (+68.30%)	0.07 (+54.99%)	0.01 (+44.32%)	121.40 (-35.46%)	93.40 (-36.24%)	86.30 (-40.61%)	84.60 (-32.70%)	15.80 (-36.89%)
`0f0da0b` — 2026-01-14 16:31:00	0.07 (n/a)	0.06 (n/a)	0.06 (n/a)	0.04 (n/a)	0.01 (n/a)	188.10 (n/a)	146.48 (n/a)	145.30 (n/a)	125.70 (n/a)	25.04 (n/a)

tanh_8_cols_1_channels_2048_tile_256

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`1531781` — 2026-01-23 18:49:24	0.11 (+64.36%)	0.09 (+41.14%)	0.10 (+56.21%)	0.05 (-7.33%)	0.03 (+346.69%)	163.80 (+7.91%)	102.90 (-22.05%)	80.40 (-35.94%)	71.70 (-39.13%)	39.99 (+185.25%)
`0f0da0b` — 2026-01-14 16:31:00	0.07 (n/a)	0.06 (n/a)	0.07 (n/a)	0.05 (n/a)	0.01 (n/a)	151.80 (n/a)	132.00 (n/a)	125.50 (n/a)	117.80 (n/a)	14.02 (n/a)

transpose_2048_M_64_N_1_cols_1_channels_64_m_64_n_8_s

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`130b6ea` — 2025-12-05 21:15:49	3.93 (-7.02%)	3.68 (+2.48%)	3.66 (+1.82%)	3.45 (+15.62%)	0.15 (-63.64%)	151.80 (-13.50%)	142.70 (-3.59%)	143.40 (-1.78%)	133.50 (+7.49%)	5.90 (-66.22%)
`0a6c11c` — 2025-12-04 00:39:10	4.22 (n/a)	3.59 (n/a)	3.59 (n/a)	2.99 (n/a)	0.42 (n/a)	175.50 (n/a)	148.02 (n/a)	146.00 (n/a)	124.20 (n/a)	17.46 (n/a)

transpose_2048_M_64_N_1_cols_1_channels_64_m_64_n_8_s0

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`1531781` — 2026-01-23 18:49:24	4.95 (+35.28%)	4.34 (+33.32%)	4.37 (+25.58%)	3.74 (+35.73%)	0.50 (+11.62%)	140.30 (-26.35%)	121.98 (-25.39%)	119.90 (-20.33%)	106.00 (-26.08%)	14.15 (-39.67%)
`0f0da0b` — 2026-01-14 16:31:00	3.66 (n/a)	3.26 (n/a)	3.48 (n/a)	2.75 (n/a)	0.45 (n/a)	190.50 (n/a)	163.48 (n/a)	150.50 (n/a)	143.40 (n/a)	23.46 (n/a)

transpose_2048_M_64_N_1_cols_2_channels_64_m_64_n_8_s

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`130b6ea` — 2025-12-05 21:15:49	5.11 (+15.57%)	3.85 (-3.32%)	3.40 (-18.64%)	3.29 (+7.09%)	0.71 (+49.14%)	159.60 (-6.61%)	140.18 (+4.80%)	154.20 (+22.97%)	102.60 (-13.49%)	22.70 (+19.77%)
`0a6c11c` — 2025-12-04 00:39:10	4.42 (n/a)	3.99 (n/a)	4.18 (n/a)	3.07 (n/a)	0.48 (n/a)	170.90 (n/a)	133.76 (n/a)	125.40 (n/a)	118.60 (n/a)	18.95 (n/a)

transpose_2048_M_64_N_1_cols_2_channels_64_m_64_n_8_s0

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`1531781` — 2026-01-23 18:49:24	6.07 (+44.96%)	4.99 (+45.77%)	5.09 (+52.05%)	4.09 (+35.80%)	0.74 (+53.13%)	128.10 (-26.34%)	106.86 (-31.23%)	103.00 (-34.27%)	86.30 (-31.02%)	15.67 (-22.94%)
`0f0da0b` — 2026-01-14 16:31:00	4.19 (n/a)	3.42 (n/a)	3.35 (n/a)	3.01 (n/a)	0.48 (n/a)	173.90 (n/a)	155.38 (n/a)	156.70 (n/a)	125.10 (n/a)	20.34 (n/a)

weighted_rms_norm_1_cols_2_channels_2048_weights_2048

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`1531781` — 2026-01-23 18:49:24	0.11 (+41.27%)	0.09 (+39.11%)	0.09 (+47.90%)	0.05 (+1.24%)	0.02 (+132.19%)	155.00 (-1.21%)	99.92 (-24.84%)	89.60 (-32.43%)	76.50 (-29.17%)	31.36 (+74.29%)
`0f0da0b` — 2026-01-14 16:31:00	0.08 (n/a)	0.06 (n/a)	0.06 (n/a)	0.05 (n/a)	0.01 (n/a)	156.90 (n/a)	132.94 (n/a)	132.60 (n/a)	108.00 (n/a)	17.99 (n/a)

weighted_rms_norm_2_cols_2_channels_2048_weights_1024

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`1531781` — 2026-01-23 18:49:24	0.10 (+48.28%)	0.09 (+48.11%)	0.09 (+60.52%)	0.05 (+2.82%)	0.02 (+134.54%)	166.70 (-2.74%)	102.72 (-28.96%)	88.50 (-37.68%)	83.10 (-32.55%)	35.87 (+61.58%)
`0f0da0b` — 2026-01-14 16:31:00	0.07 (n/a)	0.06 (n/a)	0.06 (n/a)	0.05 (n/a)	0.01 (n/a)	171.40 (n/a)	144.60 (n/a)	142.00 (n/a)	123.20 (n/a)	22.20 (n/a)

weighted_rms_norm_4_cols_2_channels_2048_weights_512

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`1531781` — 2026-01-23 18:49:24	0.09 (+21.03%)	0.09 (+31.88%)	0.09 (+37.19%)	0.07 (+25.91%)	0.01 (+25.73%)	119.00 (-20.61%)	96.48 (-24.14%)	89.90 (-27.09%)	89.30 (-17.31%)	12.81 (-17.86%)
`0f0da0b` — 2026-01-14 16:31:00	0.08 (n/a)	0.07 (n/a)	0.07 (n/a)	0.05 (n/a)	0.01 (n/a)	149.90 (n/a)	127.18 (n/a)	123.30 (n/a)	108.00 (n/a)	15.59 (n/a)

weighted_rms_norm_8_cols_2_channels_2048_weights_256

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`1531781` — 2026-01-23 18:49:24	0.11 (+67.16%)	0.10 (+80.58%)	0.10 (+80.44%)	0.06 (+74.63%)	0.02 (+72.16%)	135.80 (-42.72%)	89.92 (-44.57%)	80.00 (-44.60%)	74.30 (-40.23%)	25.76 (-41.27%)
`0f0da0b` — 2026-01-14 16:31:00	0.07 (n/a)	0.05 (n/a)	0.06 (n/a)	0.03 (n/a)	0.01 (n/a)	237.10 (n/a)	162.22 (n/a)	144.40 (n/a)	124.30 (n/a)	43.86 (n/a)

github-actions · 2026-02-06T17:43:14Z

📊 Test Results for Small Benchmark/Test Suite

e199ef2 (2026_02_06_17_42_34)

IRONCLAD

Tested on 2026_02_06_17_42_34 at commit e199ef2.

Test	Checks	Latency (mean)	Bandwidth (mean)	GFLOP/s (mean)
axpy_1_cols_2_channels_2048_tile_2048_3.0_0	✅ 5/5	139.08	0.09	n/a
axpy_2_cols_2_channels_2048_tile_1024_3.0_0	✅ 5/5	139.30	0.09	n/a
axpy_4_cols_2_channels_2048_tile_512_3.0_0	✅ 5/5	175.72	0.07	n/a
axpy_8_cols_2_channels_2048_tile_256_3.0_0	✅ 5/5	189.72	0.07	n/a
dequant_1_cols_1_channels_2048_tile_2048_0	✅ 5/5	138.34	0.04	n/a
dequant_1_cols_2_channels_2048_tile_1024_0	✅ 5/5	151.46	0.04	n/a
dequant_2_cols_1_channels_2048_tile_1024_0	✅ 5/5	165.80	0.03	n/a
dequant_2_cols_2_channels_2048_tile_512_0	✅ 5/5	166.16	0.03	n/a
dequant_4_cols_1_channels_2048_tile_512_0	✅ 5/5	178.30	0.03	n/a
dequant_4_cols_2_channels_2048_tile_256_0	✅ 5/5	179.14	0.03	n/a
dequant_8_cols_1_channels_2048_tile_256_0	✅ 5/5	169.86	0.03	n/a
dequant_8_cols_2_channels_2048_tile_128_0	✅ 5/5	188.50	0.03	n/a
eltwise_add_1_cols_2_channels_2048_tile_2048	✅ 5/5	141.20	0.09	n/a
eltwise_add_2_cols_2_channels_2048_tile_1024	✅ 5/5	170.76	0.08	n/a
eltwise_add_4_cols_2_channels_2048_tile_512	✅ 5/5	161.76	0.08	n/a
eltwise_add_8_cols_2_channels_2048_tile_256	✅ 5/5	169.00	0.07	n/a
eltwise_mul_1_cols_2_channels_2048_tile_2048	✅ 5/5	146.70	0.09	n/a
eltwise_mul_2_cols_2_channels_2048_tile_1024	✅ 5/5	139.72	0.10	n/a
eltwise_mul_4_cols_2_channels_2048_tile_512	✅ 5/5	156.88	0.08	n/a
eltwise_mul_8_cols_2_channels_2048_tile_256	✅ 5/5	224.16	0.06	n/a
gelu_1_cols_1_channels_2048_tile_2048	✅ 5/5	182.90	0.05	n/a
gelu_1_cols_2_channels_2048_tile_1024	✅ 5/5	177.56	0.05	n/a
gelu_2_cols_1_channels_2048_tile_1024	✅ 5/5	173.62	0.05	n/a
gelu_2_cols_2_channels_2048_tile_512	✅ 5/5	175.14	0.05	n/a
gelu_4_cols_1_channels_2048_tile_512	✅ 5/5	153.24	0.06	n/a
gelu_4_cols_2_channels_2048_tile_256	✅ 5/5	187.40	0.05	n/a
gelu_8_cols_1_channels_2048_tile_256	✅ 5/5	172.66	0.05	n/a
gelu_8_cols_2_channels_2048_tile_128	✅ 5/5	217.74	0.04	n/a
gemm_1792x896x1152_64x32x48_8cols_ccolmaj	❌ 0/5	1899.16	5.01	n/a
gemm_192x384x64_48x96x16_4cols	❌ 0/5	247.04	1.02	n/a
gemm_192x384x64_48x96x16_4cols_bcolmaj_ccolmaj	❌ 0/5	259.52	0.96	n/a
gemm_2048x2048x2048_64x64x64_1cols	❌ 0/5	46538.78	0.54	n/a
gemm_2048x2048x2048_64x64x64_2cols_bcolmaj	❌ 0/5	26806.12	0.94	n/a
gemm_2048x2048x2048_64x64x64_8cols_bcolmaj_ccolmaj	❌ 0/5	7272.60	3.46	n/a
gemm_384x1536x1792_32x48x64_4cols_bcolmaj	❌ 0/5	2057.74	3.98	n/a
gemm_896x1792x640_32x64x80_8cols_ccolmaj	❌ 0/5	1246.94	5.40	n/a
layer_norm_1_cols_1_channels_2048_tile_2048	✅ 5/5	149.54	0.06	n/a
layer_norm_1_cols_2_channels_2048_tile_1024	✅ 5/5	150.06	0.06	n/a
layer_norm_2_cols_1_channels_2048_tile_1024	✅ 5/5	152.52	0.06	n/a
layer_norm_2_cols_2_channels_2048_tile_512	✅ 5/5	172.52	0.05	n/a
layer_norm_4_cols_1_channels_2048_tile_512	✅ 5/5	149.52	0.06	n/a
layer_norm_4_cols_2_channels_2048_tile_256	✅ 5/5	178.44	0.05	n/a
layer_norm_8_cols_1_channels_2048_tile_256	✅ 5/5	186.48	0.04	n/a
layer_norm_8_cols_2_channels_2048_tile_128	✅ 5/5	218.34	0.04	n/a
matrix_vector_mul_128x128_32tsi_128tso_1col0	✅ 5/5	n/a	0.22	n/a
matrix_vector_mul_2048x8192_1tsi_1024tso_2col0	✅ 5/5	n/a	23.78	n/a
matrix_vector_mul_2048x8192_1tsi_2048tso_1col0	✅ 5/5	n/a	12.60	n/a
matrix_vector_mul_2048x8192_1tsi_256tso_8col0	✅ 5/5	n/a	42.19	n/a
matrix_vector_mul_2048x8192_1tsi_512tso_4col0	✅ 5/5	n/a	39.55	n/a
matrix_vector_mul_8192x2048_4tsi_1024tso_1col0	✅ 5/5	n/a	12.91	n/a
matrix_vector_mul_8192x2048_4tsi_1024tso_2col0	✅ 5/5	n/a	23.84	n/a
matrix_vector_mul_8192x2048_4tsi_1024tso_4col0	✅ 5/5	n/a	37.97	n/a
matrix_vector_mul_8192x2048_4tsi_1024tso_8col0	✅ 5/5	n/a	42.30	n/a
mem_copy_16_cores_2_chans_2048_tile_128_False0	✅ 5/5	238.94	0.04	n/a
mem_copy_1_cores_1_chans_2048_tile_2048_False0	✅ 5/5	154.62	0.05	n/a
mem_copy_2_cores_1_chans_2048_tile_1024_False0	✅ 5/5	168.64	0.05	n/a
mem_copy_2_cores_2_chans_2048_tile_1024_False0	✅ 5/5	169.56	0.05	n/a
mem_copy_4_cores_1_chans_2048_tile_512_False0	✅ 5/5	175.72	0.05	n/a
mem_copy_4_cores_2_chans_2048_tile_512_False0	✅ 5/5	179.02	0.05	n/a
mem_copy_8_cores_1_chans_2048_tile_256_False0	✅ 5/5	238.42	0.04	n/a
mem_copy_8_cores_2_chans_2048_tile_256_False0	✅ 5/5	234.80	0.04	n/a
mha0	✅ 5/5	47625.06	0.18	n/a
relu_1_cols_1_channels_2048_tile_2048	✅ 5/5	164.34	0.05	n/a
relu_2_cols_1_channels_2048_tile_1024	✅ 5/5	177.30	0.05	n/a
relu_4_cols_1_channels_2048_tile_512	✅ 5/5	163.50	0.05	n/a
relu_8_cols_1_channels_2048_tile_256	✅ 5/5	197.58	0.04	n/a
rms_norm_1_cols_1_channels_2048_tile_2048	✅ 5/5	174.62	0.05	n/a
rms_norm_1_cols_2_channels_2048_tile_1024	✅ 5/5	184.54	0.05	n/a
rms_norm_2_cols_1_channels_2048_tile_1024	✅ 5/5	177.88	0.05	n/a
rms_norm_2_cols_2_channels_2048_tile_512	✅ 5/5	213.72	0.04	n/a
rms_norm_4_cols_1_channels_2048_tile_512	✅ 5/5	193.18	0.04	n/a
rms_norm_4_cols_2_channels_2048_tile_256	✅ 5/5	231.42	0.04	n/a
rms_norm_8_cols_1_channels_2048_tile_256	✅ 5/5	200.30	0.04	n/a
rms_norm_8_cols_2_channels_2048_tile_128	✅ 5/5	214.88	0.04	n/a
rope_1c_32rows_512cols_32arows_0m	✅ 5/5	187.58	0.55	n/a
rope_1c_32rows_512cols_8arows_0m	✅ 5/5	150.18	0.50	n/a
rope_2c_32rows_512cols_32arows_0m	✅ 5/5	173.08	0.58	n/a
rope_2c_32rows_512cols_8arows_0m	✅ 5/5	186.20	0.41	n/a
rope_8c_32rows_512cols_32arows_0m	✅ 5/5	181.56	0.55	n/a
rope_8c_32rows_512cols_8arows_0m	✅ 5/5	216.00	0.36	n/a
sigmoid_1_cols_1_channels_2048_tile_2048	✅ 5/5	148.74	0.06	n/a
sigmoid_2_cols_1_channels_2048_tile_1024	✅ 5/5	148.86	0.06	n/a
sigmoid_4_cols_1_channels_2048_tile_512	✅ 5/5	152.16	0.06	n/a
sigmoid_8_cols_1_channels_2048_tile_256	✅ 5/5	158.88	0.05	n/a
silu_1_cols_1_channels_2048_tile_2048	✅ 5/5	181.20	0.05	n/a
silu_2_cols_1_channels_2048_tile_1024	✅ 5/5	152.36	0.05	n/a
silu_4_cols_1_channels_2048_tile_512	✅ 5/5	163.08	0.05	n/a
silu_8_cols_1_channels_2048_tile_256	✅ 5/5	174.10	0.05	n/a
softmax_2_cols_2_channels_32768_tile_1024	✅ 5/5	147.78	0.90	n/a
softmax_2_cols_2_channels_32768_tile_512	✅ 5/5	177.08	0.75	n/a
softmax_4_cols_4_channels_32768_tile_2048	✅ 5/5	230.78	0.58	n/a
swiglu_decode_1x2048x2048_0	❌ 0/5	n/a	n/a	n/a
swiglu_prefill_256x2048x2048_0	❌ 0/5	n/a	n/a	n/a
tanh_1_cols_1_channels_2048_tile_2048	✅ 5/5	148.62	0.06	n/a
tanh_2_cols_1_channels_2048_tile_1024	✅ 5/5	153.96	0.06	n/a
tanh_4_cols_1_channels_2048_tile_512	✅ 5/5	144.96	0.06	n/a
tanh_8_cols_1_channels_2048_tile_256	✅ 5/5	179.84	0.05	n/a
transpose_2048_M_64_N_1_cols_1_channels_64_m_64_n_8_s0	✅ 5/5	187.54	2.81	n/a
transpose_2048_M_64_N_1_cols_2_channels_64_m_64_n_8_s0	✅ 5/5	187.26	2.95	n/a
weighted_rms_norm_1_cols_2_channels_2048_weights_2048	❌ 0/5	n/a	n/a	n/a
weighted_rms_norm_2_cols_2_channels_2048_weights_1024	❌ 0/5	n/a	n/a	n/a
weighted_rms_norm_4_cols_2_channels_2048_weights_512	❌ 0/5	n/a	n/a	n/a
weighted_rms_norm_8_cols_2_channels_2048_weights_256	❌ 0/5	n/a	n/a	n/a

📈 Trends (vs main branch) for Small Benchmark/Test Suite

e199ef2 (2026_02_06_17_42_34)

IRONCLAD Trends

axpy_1_cols_2_channels_2048_tile_2048_3.0

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`130b6ea` — 2025-12-05 21:15:49	0.13 (-1.96%)	0.11 (+2.52%)	0.10 (+1.86%)	0.08 (-1.91%)	0.02 (-1.72%)	162.40 (+1.95%)	120.96 (-2.36%)	117.00 (-1.85%)	94.40 (+2.05%)	23.11 (+4.09%)
`0a6c11c` — 2025-12-04 00:39:10	0.13 (n/a)	0.10 (n/a)	0.10 (n/a)	0.08 (n/a)	0.02 (n/a)	159.30 (n/a)	123.88 (n/a)	119.20 (n/a)	92.50 (n/a)	22.21 (n/a)

axpy_1_cols_2_channels_2048_tile_2048_3.0_0

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`e199ef2` — 2026-02-06 17:34:56	0.12 (-23.02%)	0.09 (-27.67%)	0.08 (-24.68%)	0.07 (-33.92%)	0.02 (-25.64%)	179.80 (+51.35%)	139.08 (+38.66%)	146.60 (+32.79%)	101.30 (+29.87%)	30.23 (+47.53%)
`1531781` — 2026-01-23 18:49:24	0.16 (n/a)	0.13 (n/a)	0.11 (n/a)	0.10 (n/a)	0.03 (n/a)	118.80 (n/a)	100.30 (n/a)	110.40 (n/a)	78.00 (n/a)	20.49 (n/a)

axpy_2_cols_2_channels_2048_tile_1024_3.0

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`130b6ea` — 2025-12-05 21:15:49	0.11 (-26.77%)	0.09 (-9.09%)	0.09 (-10.81%)	0.08 (+8.25%)	0.01 (-55.42%)	152.00 (-7.60%)	135.46 (+6.28%)	142.80 (+12.18%)	115.00 (+36.58%)	15.03 (-41.36%)
`0a6c11c` — 2025-12-04 00:39:10	0.15 (n/a)	0.10 (n/a)	0.10 (n/a)	0.07 (n/a)	0.02 (n/a)	164.50 (n/a)	127.46 (n/a)	127.30 (n/a)	84.20 (n/a)	25.64 (n/a)

axpy_2_cols_2_channels_2048_tile_1024_3.0_0

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`e199ef2` — 2026-02-06 17:34:56	0.10 (-26.08%)	0.09 (-27.37%)	0.09 (-32.99%)	0.07 (-6.26%)	0.01 (-51.78%)	167.40 (+6.69%)	139.30 (+33.35%)	134.70 (+49.17%)	119.00 (+35.38%)	20.02 (-32.17%)
`1531781` — 2026-01-23 18:49:24	0.14 (n/a)	0.12 (n/a)	0.14 (n/a)	0.08 (n/a)	0.03 (n/a)	156.90 (n/a)	104.46 (n/a)	90.30 (n/a)	87.90 (n/a)	29.52 (n/a)

axpy_4_cols_2_channels_2048_tile_512_3.0

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`130b6ea` — 2025-12-05 21:15:49	0.12 (-10.21%)	0.10 (-8.41%)	0.10 (-6.20%)	0.07 (+3.52%)	0.02 (-23.48%)	166.70 (-3.36%)	128.78 (+7.33%)	121.80 (+6.65%)	103.90 (+11.36%)	22.95 (-18.82%)
`0a6c11c` — 2025-12-04 00:39:10	0.13 (n/a)	0.11 (n/a)	0.11 (n/a)	0.07 (n/a)	0.02 (n/a)	172.50 (n/a)	119.98 (n/a)	114.20 (n/a)	93.30 (n/a)	28.27 (n/a)

axpy_4_cols_2_channels_2048_tile_512_3.0_0

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`e199ef2` — 2026-02-06 17:34:56	0.11 (-32.15%)	0.07 (-46.71%)	0.07 (-53.73%)	0.05 (-40.43%)	0.02 (-23.77%)	225.10 (+67.86%)	175.72 (+90.09%)	180.20 (+116.07%)	115.90 (+47.46%)	41.72 (+78.42%)
`1531781` — 2026-01-23 18:49:24	0.16 (n/a)	0.14 (n/a)	0.15 (n/a)	0.09 (n/a)	0.03 (n/a)	134.10 (n/a)	92.44 (n/a)	83.40 (n/a)	78.60 (n/a)	23.38 (n/a)

axpy_8_cols_2_channels_2048_tile_256_3.0

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`130b6ea` — 2025-12-05 21:15:49	0.10 (-16.19%)	0.08 (-10.40%)	0.09 (-2.09%)	0.07 (-17.38%)	0.01 (-10.61%)	184.40 (+21.00%)	153.18 (+11.99%)	144.10 (+2.20%)	127.10 (+19.34%)	21.26 (+34.76%)
`0a6c11c` — 2025-12-04 00:39:10	0.12 (n/a)	0.09 (n/a)	0.09 (n/a)	0.08 (n/a)	0.01 (n/a)	152.40 (n/a)	136.78 (n/a)	141.00 (n/a)	106.50 (n/a)	15.78 (n/a)

axpy_8_cols_2_channels_2048_tile_256_3.0_0

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`e199ef2` — 2026-02-06 17:34:56	0.09 (-45.52%)	0.07 (-52.24%)	0.06 (-60.45%)	0.05 (-40.32%)	0.02 (-50.05%)	246.00 (+67.57%)	189.72 (+105.68%)	202.20 (+152.75%)	132.80 (+83.68%)	45.47 (+46.36%)
`1531781` — 2026-01-23 18:49:24	0.17 (n/a)	0.14 (n/a)	0.15 (n/a)	0.08 (n/a)	0.03 (n/a)	146.80 (n/a)	92.24 (n/a)	80.00 (n/a)	72.30 (n/a)	31.07 (n/a)

dequant_1_cols_1_channels_2048_tile_2048

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`130b6ea` — 2025-12-05 21:15:49	0.05 (-18.83%)	0.04 (-5.21%)	0.04 (-2.69%)	0.03 (+37.10%)	0.00 (-53.98%)	160.40 (-27.06%)	141.12 (-0.91%)	136.40 (+2.79%)	114.30 (+23.17%)	17.37 (-59.06%)
`0a6c11c` — 2025-12-04 00:39:10	0.06 (n/a)	0.04 (n/a)	0.04 (n/a)	0.02 (n/a)	0.01 (n/a)	219.90 (n/a)	142.42 (n/a)	132.70 (n/a)	92.80 (n/a)	42.43 (n/a)

dequant_1_cols_1_channels_2048_tile_2048_0

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`e199ef2` — 2026-02-06 17:34:56	0.05 (-14.04%)	0.04 (-13.10%)	0.04 (+2.83%)	0.03 (-39.58%)	0.01 (+34.89%)	207.70 (+65.50%)	138.34 (+20.21%)	118.30 (-2.71%)	102.50 (+16.35%)	41.91 (+172.15%)
`1531781` — 2026-01-23 18:49:24	0.06 (n/a)	0.05 (n/a)	0.04 (n/a)	0.04 (n/a)	0.01 (n/a)	125.50 (n/a)	115.08 (n/a)	121.60 (n/a)	88.10 (n/a)	15.40 (n/a)

dequant_1_cols_2_channels_2048_tile_1024

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`130b6ea` — 2025-12-05 21:15:49	0.06 (-13.28%)	0.05 (-8.93%)	0.06 (-13.59%)	0.04 (+4.63%)	0.01 (-39.10%)	133.60 (-4.43%)	106.08 (+5.85%)	95.10 (+15.69%)	87.90 (+15.35%)	17.90 (-31.97%)
`0a6c11c` — 2025-12-04 00:39:10	0.07 (n/a)	0.06 (n/a)	0.06 (n/a)	0.04 (n/a)	0.01 (n/a)	139.80 (n/a)	100.22 (n/a)	82.20 (n/a)	76.20 (n/a)	26.31 (n/a)

dequant_1_cols_2_channels_2048_tile_1024_0

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`e199ef2` — 2026-02-06 17:34:56	0.04 (-41.44%)	0.04 (-26.98%)	0.04 (-20.73%)	0.03 (-17.45%)	0.01 (-57.86%)	198.40 (+21.20%)	151.46 (+32.09%)	140.40 (+26.15%)	130.70 (+70.85%)	27.78 (-12.16%)
`1531781` — 2026-01-23 18:49:24	0.07 (n/a)	0.05 (n/a)	0.05 (n/a)	0.03 (n/a)	0.01 (n/a)	163.70 (n/a)	114.66 (n/a)	111.30 (n/a)	76.50 (n/a)	31.63 (n/a)

dequant_2_cols_1_channels_2048_tile_1024

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`130b6ea` — 2025-12-05 21:15:49	0.07 (+5.66%)	0.05 (+1.21%)	0.05 (+0.61%)	0.04 (-15.02%)	0.01 (+58.42%)	143.40 (+17.64%)	101.02 (+1.49%)	97.60 (-0.61%)	78.90 (-5.40%)	22.56 (+78.52%)
`0a6c11c` — 2025-12-04 00:39:10	0.06 (n/a)	0.05 (n/a)	0.05 (n/a)	0.04 (n/a)	0.01 (n/a)	121.90 (n/a)	99.54 (n/a)	98.20 (n/a)	83.40 (n/a)	12.64 (n/a)

dequant_2_cols_1_channels_2048_tile_1024_0

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`e199ef2` — 2026-02-06 17:34:56	0.04 (-18.33%)	0.03 (-28.42%)	0.04 (-22.33%)	0.02 (-34.56%)	0.01 (+36.85%)	225.80 (+52.77%)	165.80 (+46.54%)	137.10 (+28.73%)	124.20 (+22.49%)	49.83 (+154.04%)
`1531781` — 2026-01-23 18:49:24	0.05 (n/a)	0.05 (n/a)	0.05 (n/a)	0.04 (n/a)	0.01 (n/a)	147.80 (n/a)	113.14 (n/a)	106.50 (n/a)	101.40 (n/a)	19.61 (n/a)

dequant_2_cols_2_channels_2048_tile_512

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`130b6ea` — 2025-12-05 21:15:49	0.05 (+17.06%)	0.04 (+8.98%)	0.05 (+21.37%)	0.03 (-11.65%)	0.01 (+301.89%)	159.70 (+13.18%)	126.32 (-5.68%)	109.00 (-17.67%)	106.80 (-14.56%)	22.48 (+280.15%)
`0a6c11c` — 2025-12-04 00:39:10	0.04 (n/a)	0.04 (n/a)	0.04 (n/a)	0.04 (n/a)	0.00 (n/a)	141.10 (n/a)	133.92 (n/a)	132.40 (n/a)	125.00 (n/a)	5.91 (n/a)

dequant_2_cols_2_channels_2048_tile_512_0

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`e199ef2` — 2026-02-06 17:34:56	0.04 (-30.30%)	0.03 (-28.16%)	0.03 (-36.64%)	0.03 (-17.80%)	0.00 (-45.33%)	192.60 (+21.67%)	166.16 (+36.89%)	167.60 (+57.82%)	140.90 (+43.48%)	24.62 (-5.78%)
`1531781` — 2026-01-23 18:49:24	0.05 (n/a)	0.04 (n/a)	0.05 (n/a)	0.03 (n/a)	0.01 (n/a)	158.30 (n/a)	121.38 (n/a)	106.20 (n/a)	98.20 (n/a)	26.13 (n/a)

dequant_4_cols_1_channels_2048_tile_512

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`130b6ea` — 2025-12-05 21:15:49	0.06 (+20.24%)	0.05 (-3.19%)	0.05 (-6.73%)	0.03 (-20.69%)	0.01 (+194.26%)	156.10 (+26.09%)	120.86 (+7.15%)	114.90 (+7.18%)	88.30 (-16.85%)	24.40 (+214.82%)
`0a6c11c` — 2025-12-04 00:39:10	0.05 (n/a)	0.05 (n/a)	0.05 (n/a)	0.04 (n/a)	0.00 (n/a)	123.80 (n/a)	112.80 (n/a)	107.20 (n/a)	106.20 (n/a)	7.75 (n/a)

dequant_4_cols_1_channels_2048_tile_512_0

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`e199ef2` — 2026-02-06 17:34:56	0.05 (-26.04%)	0.03 (-36.96%)	0.03 (-47.45%)	0.03 (-19.27%)	0.01 (-30.41%)	204.10 (+23.85%)	178.30 (+56.93%)	196.50 (+90.41%)	114.70 (+35.26%)	37.25 (+14.44%)
`1531781` — 2026-01-23 18:49:24	0.06 (n/a)	0.05 (n/a)	0.05 (n/a)	0.03 (n/a)	0.01 (n/a)	164.80 (n/a)	113.62 (n/a)	103.20 (n/a)	84.80 (n/a)	32.55 (n/a)

dequant_4_cols_2_channels_2048_tile_256

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`130b6ea` — 2025-12-05 21:15:49	0.05 (-6.09%)	0.04 (-8.04%)	0.04 (-12.46%)	0.03 (-0.41%)	0.01 (-20.19%)	168.10 (+0.42%)	135.74 (+7.44%)	132.20 (+14.26%)	107.40 (+6.44%)	22.91 (-12.67%)
`0a6c11c` — 2025-12-04 00:39:10	0.05 (n/a)	0.04 (n/a)	0.05 (n/a)	0.03 (n/a)	0.01 (n/a)	167.40 (n/a)	126.34 (n/a)	115.70 (n/a)	100.90 (n/a)	26.23 (n/a)

dequant_4_cols_2_channels_2048_tile_256_0

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`e199ef2` — 2026-02-06 17:34:56	0.05 (-17.51%)	0.03 (-36.60%)	0.03 (-48.34%)	0.02 (-21.46%)	0.01 (-10.51%)	239.80 (+27.28%)	179.14 (+58.92%)	181.70 (+93.71%)	96.60 (+21.20%)	53.55 (+22.87%)
`1531781` — 2026-01-23 18:49:24	0.07 (n/a)	0.05 (n/a)	0.06 (n/a)	0.03 (n/a)	0.01 (n/a)	188.40 (n/a)	112.72 (n/a)	93.80 (n/a)	79.70 (n/a)	43.58 (n/a)

dequant_8_cols_1_channels_2048_tile_256

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`130b6ea` — 2025-12-05 21:15:49	0.04 (-13.65%)	0.04 (-13.67%)	0.04 (-18.56%)	0.03 (-13.52%)	0.00 (-23.88%)	170.40 (+15.68%)	140.12 (+15.33%)	135.80 (+22.78%)	119.60 (+15.78%)	17.94 (+1.87%)
`0a6c11c` — 2025-12-04 00:39:10	0.05 (n/a)	0.04 (n/a)	0.05 (n/a)	0.04 (n/a)	0.01 (n/a)	147.30 (n/a)	121.50 (n/a)	110.60 (n/a)	103.30 (n/a)	17.61 (n/a)

dequant_8_cols_1_channels_2048_tile_256_0

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`e199ef2` — 2026-02-06 17:34:56	0.05 (+17.42%)	0.03 (-14.15%)	0.03 (-26.36%)	0.03 (-14.71%)	0.01 (+130.70%)	195.70 (+17.26%)	169.86 (+22.50%)	184.50 (+35.76%)	100.10 (-14.88%)	39.31 (+119.04%)
`1531781` — 2026-01-23 18:49:24	0.04 (n/a)	0.04 (n/a)	0.04 (n/a)	0.03 (n/a)	0.00 (n/a)	166.90 (n/a)	138.66 (n/a)	135.90 (n/a)	117.60 (n/a)	17.94 (n/a)

dequant_8_cols_2_channels_2048_tile_128

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`130b6ea` — 2025-12-05 21:15:49	0.05 (+17.21%)	0.03 (-1.70%)	0.03 (-3.04%)	0.03 (-16.20%)	0.01 (+95.59%)	198.20 (+19.33%)	160.00 (+4.92%)	166.20 (+3.17%)	106.50 (-14.66%)	29.86 (+87.19%)
`0a6c11c` — 2025-12-04 00:39:10	0.04 (n/a)	0.03 (n/a)	0.03 (n/a)	0.03 (n/a)	0.00 (n/a)	166.10 (n/a)	152.50 (n/a)	161.10 (n/a)	124.80 (n/a)	15.95 (n/a)

dequant_8_cols_2_channels_2048_tile_128_0

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`e199ef2` — 2026-02-06 17:34:56	0.03 (-38.24%)	0.03 (-27.81%)	0.03 (-31.24%)	0.03 (+14.10%)	0.00 (-82.36%)	201.50 (-12.35%)	188.50 (+29.46%)	190.50 (+45.42%)	171.90 (+61.86%)	11.67 (-76.19%)
`1531781` — 2026-01-23 18:49:24	0.05 (n/a)	0.04 (n/a)	0.04 (n/a)	0.02 (n/a)	0.01 (n/a)	229.90 (n/a)	145.60 (n/a)	131.00 (n/a)	106.20 (n/a)	49.03 (n/a)

eltwise_add_1_cols_2_channels_2048_tile_2048

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`e199ef2` — 2026-02-06 17:34:56	0.10 (-38.58%)	0.09 (-26.67%)	0.09 (-9.12%)	0.07 (-16.95%)	0.01 (-64.28%)	187.60 (+20.41%)	141.20 (+28.62%)	132.60 (+10.04%)	120.20 (+62.87%)	26.56 (-22.74%)
`1531781` — 2026-01-23 18:49:24	0.17 (n/a)	0.12 (n/a)	0.10 (n/a)	0.08 (n/a)	0.04 (n/a)	155.80 (n/a)	109.78 (n/a)	120.50 (n/a)	73.80 (n/a)	34.38 (n/a)

eltwise_add_2_cols_2_channels_2048_tile_1024

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`e199ef2` — 2026-02-06 17:34:56	0.10 (-32.42%)	0.08 (-40.61%)	0.07 (-51.08%)	0.06 (-31.10%)	0.02 (-27.78%)	212.90 (+45.13%)	170.76 (+68.74%)	187.40 (+104.36%)	120.40 (+47.91%)	37.92 (+45.88%)
`1531781` — 2026-01-23 18:49:24	0.15 (n/a)	0.13 (n/a)	0.13 (n/a)	0.08 (n/a)	0.03 (n/a)	146.70 (n/a)	101.20 (n/a)	91.70 (n/a)	81.40 (n/a)	26.00 (n/a)

eltwise_add_4_cols_2_channels_2048_tile_512

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`e199ef2` — 2026-02-06 17:34:56	0.11 (-33.26%)	0.08 (-45.03%)	0.08 (-50.53%)	0.06 (-43.01%)	0.02 (-10.45%)	203.60 (+75.52%)	161.76 (+87.09%)	163.60 (+102.22%)	109.00 (+49.93%)	40.16 (+133.18%)
`1531781` — 2026-01-23 18:49:24	0.17 (n/a)	0.15 (n/a)	0.15 (n/a)	0.11 (n/a)	0.02 (n/a)	116.00 (n/a)	86.46 (n/a)	80.90 (n/a)	72.70 (n/a)	17.22 (n/a)

eltwise_add_8_cols_2_channels_2048_tile_256

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`e199ef2` — 2026-02-06 17:34:56	0.09 (-46.29%)	0.07 (-46.74%)	0.07 (-49.22%)	0.06 (-46.85%)	0.01 (-46.30%)	206.00 (+88.13%)	169.00 (+87.78%)	169.30 (+96.86%)	140.20 (+86.19%)	25.10 (+86.81%)
`1531781` — 2026-01-23 18:49:24	0.16 (n/a)	0.14 (n/a)	0.14 (n/a)	0.11 (n/a)	0.02 (n/a)	109.50 (n/a)	90.00 (n/a)	86.00 (n/a)	75.30 (n/a)	13.44 (n/a)

eltwise_mul_1_cols_2_channels_2048_tile_2048

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`e199ef2` — 2026-02-06 17:34:56	0.11 (-17.73%)	0.09 (-31.78%)	0.08 (-43.06%)	0.07 (-31.58%)	0.02 (+37.31%)	172.60 (+46.15%)	146.70 (+49.88%)	162.60 (+75.59%)	109.90 (+21.57%)	27.98 (+142.25%)
`1531781` — 2026-01-23 18:49:24	0.14 (n/a)	0.13 (n/a)	0.13 (n/a)	0.10 (n/a)	0.01 (n/a)	118.10 (n/a)	97.88 (n/a)	92.60 (n/a)	90.40 (n/a)	11.55 (n/a)

eltwise_mul_2_cols_2_channels_2048_tile_1024

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`e199ef2` — 2026-02-06 17:34:56	0.14 (+1.61%)	0.10 (-22.39%)	0.09 (-27.90%)	0.06 (-35.75%)	0.03 (+71.36%)	202.40 (+55.69%)	139.72 (+36.85%)	129.50 (+38.80%)	89.60 (-1.65%)	43.83 (+165.19%)
`1531781` — 2026-01-23 18:49:24	0.13 (n/a)	0.12 (n/a)	0.13 (n/a)	0.09 (n/a)	0.02 (n/a)	130.00 (n/a)	102.10 (n/a)	93.30 (n/a)	91.10 (n/a)	16.53 (n/a)

eltwise_mul_4_cols_2_channels_2048_tile_512

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`e199ef2` — 2026-02-06 17:34:56	0.12 (-26.95%)	0.08 (-42.37%)	0.07 (-50.41%)	0.05 (-49.23%)	0.03 (+13.45%)	224.60 (+97.02%)	156.88 (+82.04%)	164.30 (+101.60%)	101.60 (+36.93%)	46.26 (+191.85%)
`1531781` — 2026-01-23 18:49:24	0.17 (n/a)	0.15 (n/a)	0.15 (n/a)	0.11 (n/a)	0.02 (n/a)	114.00 (n/a)	86.18 (n/a)	81.50 (n/a)	74.20 (n/a)	15.85 (n/a)

eltwise_mul_8_cols_2_channels_2048_tile_256

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`e199ef2` — 2026-02-06 17:34:56	0.07 (-45.27%)	0.06 (-48.35%)	0.06 (-46.12%)	0.04 (-55.40%)	0.01 (-28.89%)	347.10 (+124.22%)	224.16 (+99.64%)	194.10 (+85.56%)	174.50 (+82.72%)	70.57 (+192.38%)
`1531781` — 2026-01-23 18:49:24	0.13 (n/a)	0.11 (n/a)	0.12 (n/a)	0.08 (n/a)	0.02 (n/a)	154.80 (n/a)	112.28 (n/a)	104.60 (n/a)	95.50 (n/a)	24.14 (n/a)

gelu_1_cols_1_channels_2048_tile_2048

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`e199ef2` — 2026-02-06 17:34:56	0.07 (-37.03%)	0.05 (-52.15%)	0.04 (-66.20%)	0.04 (-54.02%)	0.02 (-13.45%)	221.00 (+117.52%)	182.90 (+118.26%)	214.30 (+195.99%)	114.70 (+58.86%)	49.07 (+211.52%)
`1531781` — 2026-01-23 18:49:24	0.11 (n/a)	0.10 (n/a)	0.11 (n/a)	0.08 (n/a)	0.02 (n/a)	101.60 (n/a)	83.80 (n/a)	72.40 (n/a)	72.20 (n/a)	15.75 (n/a)

gelu_1_cols_2_channels_2048_tile_1024

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`e199ef2` — 2026-02-06 17:34:56	0.06 (-51.97%)	0.05 (-55.02%)	0.05 (-61.79%)	0.04 (-50.67%)	0.01 (-54.42%)	222.90 (+102.64%)	177.56 (+121.23%)	180.70 (+161.51%)	138.30 (+108.28%)	35.31 (+90.47%)
`1531781` — 2026-01-23 18:49:24	0.12 (n/a)	0.11 (n/a)	0.12 (n/a)	0.07 (n/a)	0.02 (n/a)	110.00 (n/a)	80.26 (n/a)	69.10 (n/a)	66.40 (n/a)	18.54 (n/a)

gelu_2_cols_1_channels_2048_tile_1024

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`e199ef2` — 2026-02-06 17:34:56	0.06 (-49.16%)	0.05 (-53.27%)	0.05 (-59.07%)	0.04 (-37.87%)	0.01 (-66.01%)	221.70 (+60.89%)	173.62 (+104.21%)	166.70 (+144.43%)	133.30 (+96.61%)	32.78 (+8.40%)
`1531781` — 2026-01-23 18:49:24	0.12 (n/a)	0.10 (n/a)	0.12 (n/a)	0.06 (n/a)	0.03 (n/a)	137.80 (n/a)	85.02 (n/a)	68.20 (n/a)	67.80 (n/a)	30.24 (n/a)

gelu_2_cols_2_channels_2048_tile_512

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`e199ef2` — 2026-02-06 17:34:56	0.06 (-48.95%)	0.05 (-56.47%)	0.05 (-59.99%)	0.04 (-48.42%)	0.01 (-54.03%)	209.10 (+93.97%)	175.14 (+128.05%)	173.80 (+150.07%)	131.40 (+95.83%)	29.17 (+67.66%)
`1531781` — 2026-01-23 18:49:24	0.12 (n/a)	0.11 (n/a)	0.12 (n/a)	0.08 (n/a)	0.02 (n/a)	107.80 (n/a)	76.80 (n/a)	69.50 (n/a)	67.10 (n/a)	17.40 (n/a)

gelu_4_cols_1_channels_2048_tile_512

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`e199ef2` — 2026-02-06 17:34:56	0.07 (-45.93%)	0.06 (-48.05%)	0.05 (-55.56%)	0.04 (-23.02%)	0.01 (-62.98%)	198.80 (+29.85%)	153.24 (+79.10%)	155.90 (+124.96%)	116.60 (+85.08%)	31.55 (-16.83%)
`1531781` — 2026-01-23 18:49:24	0.13 (n/a)	0.11 (n/a)	0.12 (n/a)	0.05 (n/a)	0.03 (n/a)	153.10 (n/a)	85.56 (n/a)	69.30 (n/a)	63.00 (n/a)	37.94 (n/a)

gelu_4_cols_2_channels_2048_tile_256

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`e199ef2` — 2026-02-06 17:34:56	0.07 (-42.63%)	0.05 (-56.14%)	0.04 (-63.38%)	0.03 (-57.18%)	0.01 (-28.27%)	258.00 (+133.48%)	187.40 (+134.84%)	192.40 (+172.91%)	122.10 (+74.43%)	49.89 (+185.36%)
`1531781` — 2026-01-23 18:49:24	0.12 (n/a)	0.11 (n/a)	0.12 (n/a)	0.07 (n/a)	0.02 (n/a)	110.50 (n/a)	79.80 (n/a)	70.50 (n/a)	70.00 (n/a)	17.48 (n/a)

gelu_8_cols_1_channels_2048_tile_256

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`e199ef2` — 2026-02-06 17:34:56	0.05 (-53.97%)	0.05 (-51.97%)	0.05 (-56.82%)	0.04 (+4.23%)	0.01 (-85.12%)	199.80 (-4.03%)	172.66 (+74.86%)	166.40 (+131.43%)	152.60 (+117.38%)	18.80 (-69.28%)
`1531781` — 2026-01-23 18:49:24	0.12 (n/a)	0.10 (n/a)	0.11 (n/a)	0.04 (n/a)	0.03 (n/a)	208.20 (n/a)	98.74 (n/a)	71.90 (n/a)	70.20 (n/a)	61.20 (n/a)

gelu_8_cols_2_channels_2048_tile_128

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`e199ef2` — 2026-02-06 17:34:56	0.05 (-51.96%)	0.04 (-57.95%)	0.04 (-62.76%)	0.03 (-44.97%)	0.01 (-65.93%)	242.50 (+81.78%)	217.74 (+133.03%)	221.00 (+168.53%)	170.50 (+108.18%)	28.04 (+24.81%)
`1531781` — 2026-01-23 18:49:24	0.10 (n/a)	0.09 (n/a)	0.10 (n/a)	0.06 (n/a)	0.02 (n/a)	133.40 (n/a)	93.44 (n/a)	82.30 (n/a)	81.90 (n/a)	22.47 (n/a)

gemm_1792x896x1152_64x32x48_8cols_ccolmaj

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)	Throughput (max)	Throughput (mean)	Throughput (median)	Throughput (min)	Throughput (stddev)
`e199ef2` — 2026-02-06 17:34:56	5.53 (+7.41%)	5.01 (+2.25%)	5.35 (+7.31%)	4.38 (+2.64%)	0.57 (+60.19%)	2145.10 (-2.57%)	1899.16 (-1.59%)	1756.70 (-6.81%)	1700.60 (-6.90%)	224.78 (+45.92%)	2175.28 (+7.41%)	1969.14 (+2.25%)	2105.91 (+7.31%)	1724.61 (+2.64%)	224.22 (+60.19%)
`1531781` — 2026-01-23 18:49:24	5.15 (n/a)	4.90 (n/a)	4.99 (n/a)	4.27 (n/a)	0.36 (n/a)	2201.70 (n/a)	1929.94 (n/a)	1885.00 (n/a)	1826.70 (n/a)	154.05 (n/a)	2025.19 (n/a)	1925.76 (n/a)	1962.51 (n/a)	1680.20 (n/a)	139.97 (n/a)

gemm_192x384x64_48x96x16_4cols

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)	Throughput (max)	Throughput (mean)	Throughput (median)	Throughput (min)	Throughput (stddev)
`e199ef2` — 2026-02-06 17:34:56	1.58 (-33.01%)	1.02 (-44.84%)	0.94 (-47.04%)	0.52 (-60.53%)	0.39 (-1.36%)	425.10 (+153.34%)	247.04 (+98.84%)	234.80 (+88.75%)	140.00 (+49.41%)	107.61 (+282.77%)	67.43 (-33.01%)	43.54 (-44.84%)	40.19 (-47.04%)	22.20 (-60.53%)	16.51 (-1.36%)
`1531781` — 2026-01-23 18:49:24	2.36 (n/a)	1.85 (n/a)	1.78 (n/a)	1.32 (n/a)	0.39 (n/a)	167.80 (n/a)	124.24 (n/a)	124.40 (n/a)	93.70 (n/a)	28.11 (n/a)	100.67 (n/a)	78.94 (n/a)	75.88 (n/a)	56.25 (n/a)	16.73 (n/a)

gemm_192x384x64_48x96x16_4cols_bcolmaj_ccolmaj

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)	Throughput (max)	Throughput (mean)	Throughput (median)	Throughput (min)	Throughput (stddev)
`e199ef2` — 2026-02-06 17:34:56	1.39 (-37.23%)	0.96 (-53.49%)	0.96 (-56.50%)	0.49 (-69.46%)	0.32 (+21.46%)	452.50 (+227.42%)	259.52 (+137.66%)	231.00 (+129.85%)	159.00 (+59.32%)	112.21 (+578.47%)	59.34 (-37.23%)	40.82 (-53.49%)	40.85 (-56.50%)	20.85 (-69.46%)	13.68 (+21.46%)
`1531781` — 2026-01-23 18:49:24	2.22 (n/a)	2.06 (n/a)	2.20 (n/a)	1.60 (n/a)	0.26 (n/a)	138.20 (n/a)	109.20 (n/a)	100.50 (n/a)	99.80 (n/a)	16.54 (n/a)	94.54 (n/a)	87.77 (n/a)	93.91 (n/a)	68.30 (n/a)	11.26 (n/a)

gemm_2048x2048x2048_64x64x32_8_cols_0_bcolmaj_0_ccolmaj_0

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	GFLOP/s (max)	GFLOP/s (mean)	GFLOP/s (median)	GFLOP/s (min)	GFLOP/s (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`3a053d3` — 2025-11-26 01:12:16	3.72 (-0.11%)	3.55 (-2.75%)	3.56 (-2.75%)	3.36 (-5.33%)	0.12 (+76.10%)	2725.23 (-0.11%)	2602.52 (-2.75%)	2612.11 (-2.75%)	2459.89 (-5.33%)	86.27 (+76.10%)	6984.00 (+5.63%)	6608.60 (+2.91%)	6577.00 (+2.83%)	6304.00 (+0.11%)	221.87 (+87.15%)
`9e98049` — 2025-11-25 20:50:55	3.72 (n/a)	3.65 (n/a)	3.66 (n/a)	3.54 (n/a)	0.07 (n/a)	2728.26 (n/a)	2676.06 (n/a)	2686.03 (n/a)	2598.29 (n/a)	48.99 (n/a)	6612.00 (n/a)	6422.00 (n/a)	6396.00 (n/a)	6297.00 (n/a)	118.55 (n/a)

gemm_2048x2048x2048_64x64x32_8_cols_0_bcolmaj_1_ccolmaj_0

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	GFLOP/s (max)	GFLOP/s (mean)	GFLOP/s (median)	GFLOP/s (min)	GFLOP/s (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`3a053d3` — 2025-11-26 01:12:16	3.40 (-3.14%)	3.38 (-0.82%)	3.38 (-0.82%)	3.35 (+4.93%)	0.02 (-84.51%)	2493.81 (-3.14%)	2475.27 (-0.82%)	2479.06 (-0.82%)	2453.92 (+4.93%)	13.33 (-84.51%)	7001.00 (-4.70%)	6940.80 (+0.71%)	6930.00 (+0.83%)	6889.00 (+3.24%)	37.43 (-84.76%)
`9e98049` — 2025-11-25 20:50:55	3.51 (n/a)	3.40 (n/a)	3.41 (n/a)	3.19 (n/a)	0.12 (n/a)	2574.53 (n/a)	2495.72 (n/a)	2499.62 (n/a)	2338.67 (n/a)	86.02 (n/a)	7346.00 (n/a)	6892.20 (n/a)	6873.00 (n/a)	6673.00 (n/a)	245.57 (n/a)

gemm_2048x2048x2048_64x64x32_8_cols_1_bcolmaj_0_ccolmaj_0

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	GFLOP/s (max)	GFLOP/s (mean)	GFLOP/s (median)	GFLOP/s (min)	GFLOP/s (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`3a053d3` — 2025-11-26 01:12:16	3.02 (+0.52%)	2.87 (-2.05%)	2.86 (-2.24%)	2.70 (-6.16%)	0.11 (+131.66%)	2216.19 (+0.52%)	2104.44 (-2.05%)	2094.08 (-2.24%)	1980.84 (-6.16%)	80.25 (+131.67%)	8673.00 (+6.56%)	8175.60 (+2.21%)	8204.00 (+2.29%)	7752.00 (-0.51%)	314.03 (+145.36%)
`9e98049` — 2025-11-25 20:50:55	3.01 (n/a)	2.93 (n/a)	2.92 (n/a)	2.88 (n/a)	0.05 (n/a)	2204.81 (n/a)	2148.41 (n/a)	2142.13 (n/a)	2110.81 (n/a)	34.64 (n/a)	8139.00 (n/a)	7998.60 (n/a)	8020.00 (n/a)	7792.00 (n/a)	127.99 (n/a)

gemm_2048x2048x2048_64x64x64_1cols

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)	Throughput (max)	Throughput (mean)	Throughput (median)	Throughput (min)	Throughput (stddev)
`e199ef2` — 2026-02-06 17:34:56	0.54 (+4.11%)	0.54 (+3.94%)	0.54 (+4.08%)	0.54 (+3.72%)	0.00 (+344.71%)	46677.10 (-3.59%)	46538.78 (-3.79%)	46465.20 (-3.92%)	46436.50 (-3.95%)	121.03 (+311.71%)	369.96 (+4.11%)	369.15 (+3.94%)	369.74 (+4.08%)	368.06 (+3.72%)	0.96 (+344.70%)
`1531781` — 2026-01-23 18:49:24	0.52 (n/a)	0.52 (n/a)	0.52 (n/a)	0.52 (n/a)	0.00 (n/a)	48415.10 (n/a)	48370.52 (n/a)	48362.20 (n/a)	48344.40 (n/a)	29.40 (n/a)	355.36 (n/a)	355.17 (n/a)	355.23 (n/a)	354.85 (n/a)	0.22 (n/a)

gemm_2048x2048x2048_64x64x64_2_cols_0_bcolmaj_0_ccolmaj_0

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`130b6ea` — 2025-12-05 21:15:49	1.04 (+0.94%)	1.03 (+0.23%)	1.03 (+0.09%)	1.02 (+0.07%)	0.00 (+197.51%)	24576.00 (-0.07%)	24485.82 (-0.23%)	24530.40 (-0.09%)	24266.40 (-0.94%)	111.26 (+194.35%)
`0a6c11c` — 2025-12-04 00:39:10	1.03 (n/a)	1.03 (n/a)	1.03 (n/a)	1.02 (n/a)	0.00 (n/a)	24594.30 (n/a)	24542.14 (n/a)	24551.90 (n/a)	24495.50 (n/a)	37.80 (n/a)

gemm_2048x2048x2048_64x64x64_2_cols_0_bcolmaj_0_ccolmaj_0_0

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)	Throughput (max)	Throughput (mean)	Throughput (median)	Throughput (min)	Throughput (stddev)
`a4b6ffe` — 2026-01-05 19:24:05	1.04 (+0.28%)	1.03 (-0.15%)	1.03 (+0.05%)	1.02 (-0.52%)	0.01 (+42.72%)	24719.80 (+0.52%)	24512.42 (+0.16%)	24529.20 (-0.05%)	24092.60 (-0.28%)	252.92 (+43.15%)	713.08 (+0.28%)	700.92 (-0.15%)	700.38 (+0.05%)	694.98 (-0.52%)	7.30 (+42.71%)
`50ad37c` — 2025-12-29 20:35:53	1.04 (n/a)	1.03 (n/a)	1.03 (n/a)	1.02 (n/a)	0.01 (n/a)	24592.30 (n/a)	24473.70 (n/a)	24540.90 (n/a)	24160.80 (n/a)	176.68 (n/a)	711.06 (n/a)	702.00 (n/a)	700.05 (n/a)	698.59 (n/a)	5.12 (n/a)

gemm_2048x2048x2048_64x64x64_2_cols_0_bcolmaj_1_ccolmaj_0

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`130b6ea` — 2025-12-05 21:15:49	0.99 (+1.69%)	0.97 (+0.29%)	0.97 (-0.04%)	0.97 (-0.15%)	0.01 (+434.92%)	26025.30 (+0.15%)	25839.34 (-0.29%)	25916.70 (+0.04%)	25446.00 (-1.67%)	202.05 (+425.44%)
`0a6c11c` — 2025-12-04 00:39:10	0.97 (n/a)	0.97 (n/a)	0.97 (n/a)	0.97 (n/a)	0.00 (n/a)	25986.90 (n/a)	25913.52 (n/a)	25905.70 (n/a)	25877.00 (n/a)	38.45 (n/a)

gemm_2048x2048x2048_64x64x64_2_cols_0_bcolmaj_1_ccolmaj_0_0

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)	Throughput (max)	Throughput (mean)	Throughput (median)	Throughput (min)	Throughput (stddev)
`a4b6ffe` — 2026-01-05 19:24:05	0.97 (-0.01%)	0.97 (+0.03%)	0.97 (+0.04%)	0.97 (+0.00%)	0.00 (-6.43%)	25927.90 (-0.00%)	25903.20 (-0.03%)	25898.60 (-0.04%)	25892.80 (+0.01%)	14.05 (-6.56%)	663.50 (-0.01%)	663.23 (+0.03%)	663.35 (+0.04%)	662.60 (+0.00%)	0.36 (-6.43%)
`50ad37c` — 2025-12-29 20:35:53	0.97 (n/a)	0.97 (n/a)	0.97 (n/a)	0.97 (n/a)	0.00 (n/a)	25928.10 (n/a)	25910.14 (n/a)	25908.10 (n/a)	25891.20 (n/a)	15.04 (n/a)	663.54 (n/a)	663.06 (n/a)	663.11 (n/a)	662.60 (n/a)	0.38 (n/a)

gemm_2048x2048x2048_64x64x64_2_cols_1_bcolmaj_0_ccolmaj_0

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`130b6ea` — 2025-12-05 21:15:49	0.92 (-1.02%)	0.92 (-0.66%)	0.92 (-0.10%)	0.91 (-0.78%)	0.01 (-4.54%)	27751.40 (+0.79%)	27480.74 (+0.66%)	27354.80 (+0.10%)	27320.50 (+1.03%)	182.69 (-2.84%)
`0a6c11c` — 2025-12-04 00:39:10	0.93 (n/a)	0.92 (n/a)	0.92 (n/a)	0.91 (n/a)	0.01 (n/a)	27534.80 (n/a)	27299.32 (n/a)	27326.70 (n/a)	27042.30 (n/a)	188.04 (n/a)

gemm_2048x2048x2048_64x64x64_2_cols_1_bcolmaj_0_ccolmaj_0_0

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)	Throughput (max)	Throughput (mean)	Throughput (median)	Throughput (min)	Throughput (stddev)
`a4b6ffe` — 2026-01-05 19:24:05	0.92 (+0.11%)	0.91 (-0.12%)	0.91 (-0.21%)	0.91 (-0.24%)	0.00 (+30.34%)	27723.50 (+0.24%)	27548.94 (+0.12%)	27522.90 (+0.21%)	27403.60 (-0.11%)	124.26 (+30.58%)	626.92 (+0.11%)	623.62 (-0.12%)	624.20 (-0.21%)	619.69 (-0.24%)	2.81 (+30.34%)
`50ad37c` — 2025-12-29 20:35:53	0.92 (n/a)	0.91 (n/a)	0.92 (n/a)	0.91 (n/a)	0.00 (n/a)	27657.10 (n/a)	27515.58 (n/a)	27464.70 (n/a)	27434.60 (n/a)	95.16 (n/a)	626.21 (n/a)	624.37 (n/a)	625.52 (n/a)	621.17 (n/a)	2.16 (n/a)

gemm_2048x2048x2048_64x64x64_2cols_bcolmaj

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)	Throughput (max)	Throughput (mean)	Throughput (median)	Throughput (min)	Throughput (stddev)
`e199ef2` — 2026-02-06 17:34:56	0.94 (+1.68%)	0.94 (+2.71%)	0.94 (+2.91%)	0.93 (+2.83%)	0.01 (-34.04%)	27060.50 (-2.75%)	26806.12 (-2.64%)	26742.00 (-2.83%)	26658.30 (-1.65%)	171.91 (-36.88%)	644.45 (+1.68%)	640.91 (+2.71%)	642.43 (+2.91%)	634.87 (+2.83%)	4.10 (-34.04%)
`1531781` — 2026-01-23 18:49:24	0.93 (n/a)	0.91 (n/a)	0.91 (n/a)	0.90 (n/a)	0.01 (n/a)	27826.70 (n/a)	27533.84 (n/a)	27519.90 (n/a)	27105.20 (n/a)	272.35 (n/a)	633.82 (n/a)	624.00 (n/a)	624.27 (n/a)	617.39 (n/a)	6.21 (n/a)

gemm_2048x2048x2048_64x64x64_8_cols_0_bcolmaj_0_ccolmaj_0

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	GFLOP/s (max)	GFLOP/s (mean)	GFLOP/s (median)	GFLOP/s (min)	GFLOP/s (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`31ab24e` — 2025-11-22 02:01:56	3.65 (-1.06%)	3.59 (+0.83%)	3.63 (-0.26%)	3.45 (+2.66%)	0.07 (-44.71%)	2678.08 (-1.06%)	2632.13 (+0.83%)	2658.60 (-0.26%)	2528.31 (+2.66%)	53.86 (-44.71%)	6795.00 (-2.59%)	6529.80 (-0.92%)	6462.00 (+0.26%)	6415.00 (+1.07%)	137.14 (-45.27%)
`d9a64c5` — 2025-11-21 20:49:35	3.69 (n/a)	3.56 (n/a)	3.64 (n/a)	3.36 (n/a)	0.13 (n/a)	2706.77 (n/a)	2610.50 (n/a)	2665.61 (n/a)	2462.71 (n/a)	97.41 (n/a)	6976.00 (n/a)	6590.40 (n/a)	6445.00 (n/a)	6347.00 (n/a)	250.58 (n/a)

gemm_2048x2048x2048_64x64x64_8_cols_0_bcolmaj_0_ccolmaj_0_0

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)	Throughput (max)	Throughput (mean)	Throughput (median)	Throughput (min)	Throughput (stddev)
`d48746f` — 2026-01-12 22:17:10	3.98 (+0.20%)	3.84 (-2.95%)	3.76 (-5.27%)	3.72 (-4.97%)	0.13 (+473.97%)	6760.00 (+5.23%)	6559.50 (+3.13%)	6697.10 (+5.56%)	6321.00 (-0.20%)	218.68 (+499.40%)	2717.91 (+0.20%)	2621.44 (-2.95%)	2565.27 (-5.27%)	2541.40 (-4.97%)	88.37 (+473.97%)
`331dcca` — 2026-01-08 18:34:10	3.97 (n/a)	3.96 (n/a)	3.97 (n/a)	3.92 (n/a)	0.02 (n/a)	6423.80 (n/a)	6360.32 (n/a)	6344.30 (n/a)	6333.90 (n/a)	36.48 (n/a)	2712.39 (n/a)	2701.17 (n/a)	2707.91 (n/a)	2674.40 (n/a)	15.40 (n/a)

gemm_2048x2048x2048_64x64x64_8_cols_0_bcolmaj_1_ccolmaj_0_0

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)	Throughput (max)	Throughput (mean)	Throughput (median)	Throughput (min)	Throughput (stddev)
`d48746f` — 2026-01-12 22:17:10	3.75 (+0.28%)	3.70 (+1.25%)	3.71 (-0.21%)	3.65 (+4.77%)	0.04 (-66.90%)	6901.80 (-4.56%)	6795.46 (-1.30%)	6790.20 (+0.21%)	6715.90 (-0.28%)	67.05 (-68.40%)	2558.08 (+0.28%)	2528.33 (+1.25%)	2530.10 (-0.21%)	2489.18 (+4.77%)	24.81 (-66.90%)
`331dcca` — 2026-01-08 18:34:10	3.74 (n/a)	3.66 (n/a)	3.71 (n/a)	3.48 (n/a)	0.11 (n/a)	7231.20 (n/a)	6885.18 (n/a)	6775.70 (n/a)	6734.90 (n/a)	212.21 (n/a)	2550.87 (n/a)	2497.04 (n/a)	2535.50 (n/a)	2375.78 (n/a)	74.96 (n/a)

gemm_2048x2048x2048_64x64x64_8_cols_1_bcolmaj_0_ccolmaj_0_0

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)	Throughput (max)	Throughput (mean)	Throughput (median)	Throughput (min)	Throughput (stddev)
`d48746f` — 2026-01-12 22:17:10	3.62 (-0.77%)	3.54 (+0.14%)	3.60 (-0.04%)	3.44 (+2.17%)	0.09 (-25.77%)	7307.40 (-2.13%)	7103.46 (-0.18%)	6982.50 (+0.04%)	6948.20 (+0.77%)	182.95 (-26.55%)	2472.55 (-0.77%)	2419.80 (+0.14%)	2460.41 (-0.04%)	2351.04 (+2.17%)	61.76 (-25.77%)
`331dcca` — 2026-01-08 18:34:10	3.65 (n/a)	3.54 (n/a)	3.61 (n/a)	3.37 (n/a)	0.12 (n/a)	7466.20 (n/a)	7116.40 (n/a)	6979.60 (n/a)	6895.10 (n/a)	249.09 (n/a)	2491.61 (n/a)	2416.45 (n/a)	2461.45 (n/a)	2301.00 (n/a)	83.20 (n/a)

gemm_2048x2048x2048_64x64x64_8cols_bcolmaj_ccolmaj

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)	Throughput (max)	Throughput (mean)	Throughput (median)	Throughput (min)	Throughput (stddev)
`e199ef2` — 2026-02-06 17:34:56	3.51 (+1.39%)	3.46 (+1.30%)	3.49 (+1.11%)	3.38 (+2.65%)	0.05 (-27.03%)	7453.40 (-2.59%)	7272.60 (-1.30%)	7220.70 (-1.10%)	7164.20 (-1.37%)	113.93 (-29.90%)	2398.00 (+1.39%)	2362.72 (+1.30%)	2379.24 (+1.11%)	2304.96 (+2.65%)	36.60 (-27.03%)
`1531781` — 2026-01-23 18:49:24	3.46 (n/a)	3.42 (n/a)	3.45 (n/a)	3.29 (n/a)	0.07 (n/a)	7651.20 (n/a)	7368.34 (n/a)	7300.80 (n/a)	7263.70 (n/a)	162.53 (n/a)	2365.17 (n/a)	2332.46 (n/a)	2353.14 (n/a)	2245.39 (n/a)	50.15 (n/a)

gemm_384x1536x1792_32x48x64_4cols_bcolmaj

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)	Throughput (max)	Throughput (mean)	Throughput (median)	Throughput (min)	Throughput (stddev)
`e199ef2` — 2026-02-06 17:34:56	4.42 (+1.55%)	3.98 (-6.17%)	4.13 (-4.51%)	3.16 (-17.85%)	0.52 (+137.98%)	2550.10 (+21.73%)	2057.74 (+7.98%)	1950.00 (+4.73%)	1823.70 (-1.53%)	300.85 (+184.13%)	1159.17 (+1.55%)	1043.25 (-6.17%)	1084.05 (-4.51%)	828.95 (-17.85%)	136.87 (+137.98%)
`1531781` — 2026-01-23 18:49:24	4.35 (n/a)	4.24 (n/a)	4.33 (n/a)	3.85 (n/a)	0.22 (n/a)	2094.90 (n/a)	1905.64 (n/a)	1862.00 (n/a)	1852.00 (n/a)	105.88 (n/a)	1141.45 (n/a)	1111.86 (n/a)	1135.29 (n/a)	1009.08 (n/a)	57.51 (n/a)

gemm_896x1792x640_32x64x80_8cols_ccolmaj

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)	Throughput (max)	Throughput (mean)	Throughput (median)	Throughput (min)	Throughput (stddev)
`e199ef2` — 2026-02-06 17:34:56	6.67 (-1.95%)	5.40 (-16.02%)	5.18 (-22.62%)	4.86 (-8.95%)	0.73 (+18.08%)	1369.70 (+9.83%)	1246.94 (+19.59%)	1283.20 (+29.24%)	997.80 (+1.99%)	146.90 (+28.20%)	2059.78 (-1.95%)	1669.40 (-16.02%)	1601.67 (-22.62%)	1500.46 (-8.95%)	225.34 (+18.09%)
`1531781` — 2026-01-23 18:49:24	6.80 (n/a)	6.43 (n/a)	6.70 (n/a)	5.33 (n/a)	0.62 (n/a)	1247.10 (n/a)	1042.70 (n/a)	992.90 (n/a)	978.30 (n/a)	114.59 (n/a)	2100.72 (n/a)	1987.76 (n/a)	2069.85 (n/a)	1647.93 (n/a)	190.83 (n/a)

layer_norm_1_cols_1_channels_2048_tile_2048

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`e199ef2` — 2026-02-06 17:34:56	0.07 (-38.93%)	0.06 (-36.29%)	0.05 (-32.96%)	0.05 (-33.57%)	0.01 (-54.51%)	180.20 (+50.54%)	149.54 (+53.47%)	150.00 (+49.25%)	117.10 (+63.78%)	25.09 (+10.68%)
`1531781` — 2026-01-23 18:49:24	0.11 (n/a)	0.09 (n/a)	0.08 (n/a)	0.07 (n/a)	0.02 (n/a)	119.70 (n/a)	97.44 (n/a)	100.50 (n/a)	71.50 (n/a)	22.67 (n/a)

layer_norm_1_cols_2_channels_2048_tile_1024

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`e199ef2` — 2026-02-06 17:34:56	0.08 (-29.71%)	0.06 (-33.92%)	0.05 (-36.86%)	0.04 (-33.93%)	0.01 (-29.47%)	209.10 (+51.41%)	150.06 (+51.91%)	153.80 (+58.39%)	108.70 (+42.28%)	38.40 (+54.34%)
`1531781` — 2026-01-23 18:49:24	0.11 (n/a)	0.09 (n/a)	0.08 (n/a)	0.06 (n/a)	0.02 (n/a)	138.10 (n/a)	98.78 (n/a)	97.10 (n/a)	76.40 (n/a)	24.88 (n/a)

layer_norm_2_cols_1_channels_2048_tile_1024

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`e199ef2` — 2026-02-06 17:34:56	0.06 (-46.02%)	0.06 (-37.20%)	0.06 (-33.36%)	0.04 (-29.27%)	0.01 (-57.13%)	198.10 (+41.40%)	152.52 (+55.03%)	141.60 (+50.00%)	126.10 (+85.17%)	29.04 (+11.23%)
`1531781` — 2026-01-23 18:49:24	0.12 (n/a)	0.09 (n/a)	0.09 (n/a)	0.06 (n/a)	0.02 (n/a)	140.10 (n/a)	98.38 (n/a)	94.40 (n/a)	68.10 (n/a)	26.11 (n/a)

layer_norm_2_cols_2_channels_2048_tile_512

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`e199ef2` — 2026-02-06 17:34:56	0.06 (-53.14%)	0.05 (-48.55%)	0.05 (-46.34%)	0.04 (-39.09%)	0.01 (-65.59%)	214.10 (+64.19%)	172.52 (+87.73%)	174.30 (+86.42%)	132.30 (+113.39%)	30.69 (+19.92%)
`1531781` — 2026-01-23 18:49:24	0.13 (n/a)	0.09 (n/a)	0.09 (n/a)	0.06 (n/a)	0.03 (n/a)	130.40 (n/a)	91.90 (n/a)	93.50 (n/a)	62.00 (n/a)	25.59 (n/a)

layer_norm_4_cols_1_channels_2048_tile_512

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`e199ef2` — 2026-02-06 17:34:56	0.06 (-44.49%)	0.06 (-42.00%)	0.05 (-49.19%)	0.05 (-12.66%)	0.00 (-80.02%)	161.40 (+14.47%)	149.52 (+63.34%)	153.30 (+96.79%)	131.60 (+80.27%)	11.31 (-60.20%)
`1531781` — 2026-01-23 18:49:24	0.11 (n/a)	0.09 (n/a)	0.11 (n/a)	0.06 (n/a)	0.02 (n/a)	141.00 (n/a)	91.54 (n/a)	77.90 (n/a)	73.00 (n/a)	28.43 (n/a)

layer_norm_4_cols_2_channels_2048_tile_256

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`e199ef2` — 2026-02-06 17:34:56	0.06 (-48.62%)	0.05 (-45.43%)	0.04 (-51.93%)	0.04 (-27.27%)	0.01 (-72.67%)	204.10 (+37.53%)	178.44 (+73.01%)	182.60 (+107.97%)	143.50 (+94.71%)	23.24 (-28.04%)
`1531781` — 2026-01-23 18:49:24	0.11 (n/a)	0.09 (n/a)	0.09 (n/a)	0.06 (n/a)	0.02 (n/a)	148.40 (n/a)	103.14 (n/a)	87.80 (n/a)	73.70 (n/a)	32.30 (n/a)

layer_norm_8_cols_1_channels_2048_tile_256

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`e199ef2` — 2026-02-06 17:34:56	0.05 (-61.86%)	0.04 (-55.58%)	0.04 (-58.59%)	0.04 (-21.84%)	0.00 (-92.76%)	196.90 (+27.94%)	186.48 (+106.37%)	186.90 (+141.47%)	176.40 (+162.11%)	8.20 (-77.13%)
`1531781` — 2026-01-23 18:49:24	0.12 (n/a)	0.10 (n/a)	0.11 (n/a)	0.05 (n/a)	0.03 (n/a)	153.90 (n/a)	90.36 (n/a)	77.40 (n/a)	67.30 (n/a)	35.86 (n/a)

layer_norm_8_cols_2_channels_2048_tile_128

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`e199ef2` — 2026-02-06 17:34:56	0.05 (-43.60%)	0.04 (-50.02%)	0.04 (-49.87%)	0.03 (-54.27%)	0.01 (-27.15%)	306.80 (+118.67%)	218.34 (+105.01%)	194.80 (+99.59%)	163.80 (+77.27%)	56.47 (+184.66%)
`1531781` — 2026-01-23 18:49:24	0.09 (n/a)	0.08 (n/a)	0.08 (n/a)	0.06 (n/a)	0.01 (n/a)	140.30 (n/a)	106.50 (n/a)	97.60 (n/a)	92.40 (n/a)	19.84 (n/a)

matrix_vector_mul_128x128_32_1col

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)
`130b6ea` — 2025-12-05 21:15:49	0.37 (+17.13%)	0.30 (+24.87%)	0.30 (+38.03%)	0.23 (+12.47%)	0.06 (+35.23%)
`0a6c11c` — 2025-12-04 00:39:10	0.32 (n/a)	0.24 (n/a)	0.22 (n/a)	0.20 (n/a)	0.04 (n/a)

matrix_vector_mul_128x128_32_1col0

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Throughput (max)	Throughput (mean)	Throughput (median)	Throughput (min)	Throughput (stddev)
`331dcca` — 2026-01-08 18:34:10	0.26 (-1.81%)	0.24 (+4.06%)	0.24 (+0.52%)	0.21 (+16.97%)	0.02 (-48.16%)	0.26 (-1.81%)	0.24 (+4.06%)	0.24 (+0.52%)	0.20 (+16.97%)	0.02 (-48.16%)
`a4b6ffe` — 2026-01-05 19:24:05	0.26 (n/a)	0.23 (n/a)	0.24 (n/a)	0.18 (n/a)	0.04 (n/a)	0.26 (n/a)	0.23 (n/a)	0.24 (n/a)	0.17 (n/a)	0.04 (n/a)

matrix_vector_mul_128x128_32tsi_128tso_1col0

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Throughput (max)	Throughput (mean)	Throughput (median)	Throughput (min)	Throughput (stddev)
`e199ef2` — 2026-02-06 17:34:56	0.26 (-41.49%)	0.22 (-36.65%)	0.22 (-35.58%)	0.15 (-32.33%)	0.04 (-45.00%)	0.26 (-41.49%)	0.22 (-36.65%)	0.22 (-35.58%)	0.15 (-32.33%)	0.04 (-45.00%)
`1531781` — 2026-01-23 18:49:24	0.45 (n/a)	0.35 (n/a)	0.35 (n/a)	0.22 (n/a)	0.08 (n/a)	0.44 (n/a)	0.34 (n/a)	0.34 (n/a)	0.22 (n/a)	0.08 (n/a)

matrix_vector_mul_2048x8192_1_1col

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)
`130b6ea` — 2025-12-05 21:15:49	13.57 (+3.47%)	13.15 (+1.09%)	13.10 (+0.16%)	12.52 (-1.58%)	0.37 (+153.19%)
`0a6c11c` — 2025-12-04 00:39:10	13.11 (n/a)	13.01 (n/a)	13.07 (n/a)	12.72 (n/a)	0.15 (n/a)

matrix_vector_mul_2048x8192_1_1col0

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Throughput (max)	Throughput (mean)	Throughput (median)	Throughput (min)	Throughput (stddev)
`331dcca` — 2026-01-08 18:34:10	13.43 (-0.07%)	13.30 (+4.62%)	13.27 (+1.29%)	13.16 (+16.70%)	0.12 (-86.69%)	13.42 (-0.07%)	13.29 (+4.62%)	13.26 (+1.29%)	13.16 (+16.70%)	0.12 (-86.69%)
`a4b6ffe` — 2026-01-05 19:24:05	13.44 (n/a)	12.71 (n/a)	13.10 (n/a)	11.28 (n/a)	0.87 (n/a)	13.43 (n/a)	12.70 (n/a)	13.09 (n/a)	11.27 (n/a)	0.87 (n/a)

matrix_vector_mul_2048x8192_1_2col

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)
`130b6ea` — 2025-12-05 21:15:49	25.06 (+2.47%)	23.71 (+1.81%)	24.13 (+0.50%)	20.90 (+5.41%)	1.48 (-15.60%)
`0a6c11c` — 2025-12-04 00:39:10	24.46 (n/a)	23.29 (n/a)	24.01 (n/a)	19.83 (n/a)	1.75 (n/a)

matrix_vector_mul_2048x8192_1_2col0

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Throughput (max)	Throughput (mean)	Throughput (median)	Throughput (min)	Throughput (stddev)
`331dcca` — 2026-01-08 18:34:10	24.46 (-0.60%)	21.07 (-8.03%)	19.75 (-17.83%)	18.08 (+1.18%)	3.03 (+7.07%)	24.44 (-0.60%)	21.06 (-8.03%)	19.74 (-17.83%)	18.07 (+1.18%)	3.03 (+7.07%)
`a4b6ffe` — 2026-01-05 19:24:05	24.60 (n/a)	22.91 (n/a)	24.04 (n/a)	17.87 (n/a)	2.83 (n/a)	24.59 (n/a)	22.90 (n/a)	24.02 (n/a)	17.86 (n/a)	2.83 (n/a)

matrix_vector_mul_2048x8192_1_4col

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)
`130b6ea` — 2025-12-05 21:15:49	42.05 (+1.54%)	40.50 (+12.60%)	40.49 (+2.36%)	38.98 (+97.53%)	0.97 (-88.09%)
`0a6c11c` — 2025-12-04 00:39:10	41.41 (n/a)	35.97 (n/a)	39.55 (n/a)	19.73 (n/a)	8.15 (n/a)

matrix_vector_mul_2048x8192_1_4col0

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Throughput (max)	Throughput (mean)	Throughput (median)	Throughput (min)	Throughput (stddev)
`331dcca` — 2026-01-08 18:34:10	42.58 (+1.51%)	41.37 (+14.29%)	41.25 (+4.89%)	40.33 (+91.42%)	0.92 (-89.18%)	42.56 (+1.51%)	41.34 (+14.29%)	41.23 (+4.89%)	40.30 (+91.42%)	0.92 (-89.18%)
`a4b6ffe` — 2026-01-05 19:24:05	41.95 (n/a)	36.20 (n/a)	39.33 (n/a)	21.07 (n/a)	8.54 (n/a)	41.92 (n/a)	36.17 (n/a)	39.31 (n/a)	21.05 (n/a)	8.53 (n/a)

matrix_vector_mul_2048x8192_1_8col

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)
`130b6ea` — 2025-12-05 21:15:49	46.62 (+3.11%)	44.82 (+0.17%)	44.78 (+0.17%)	42.60 (-3.80%)	1.60 (+367.72%)
`0a6c11c` — 2025-12-04 00:39:10	45.21 (n/a)	44.75 (n/a)	44.70 (n/a)	44.29 (n/a)	0.34 (n/a)

matrix_vector_mul_2048x8192_1_8col0

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Throughput (max)	Throughput (mean)	Throughput (median)	Throughput (min)	Throughput (stddev)
`331dcca` — 2026-01-08 18:34:10	46.04 (+4.14%)	45.06 (+3.47%)	44.55 (+2.76%)	44.30 (+2.84%)	0.87 (+66.58%)	46.01 (+4.14%)	45.03 (+3.47%)	44.52 (+2.76%)	44.27 (+2.84%)	0.87 (+66.58%)
`a4b6ffe` — 2026-01-05 19:24:05	44.21 (n/a)	43.55 (n/a)	43.35 (n/a)	43.07 (n/a)	0.52 (n/a)	44.18 (n/a)	43.52 (n/a)	43.32 (n/a)	43.05 (n/a)	0.52 (n/a)

matrix_vector_mul_2048x8192_1tsi_1024tso_2col0

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Throughput (max)	Throughput (mean)	Throughput (median)	Throughput (min)	Throughput (stddev)
`e199ef2` — 2026-02-06 17:34:56	24.38 (-1.71%)	23.78 (-3.29%)	23.90 (-2.61%)	22.76 (-6.71%)	0.62 (+252.59%)	24.37 (-1.71%)	23.77 (-3.29%)	23.89 (-2.61%)	22.75 (-6.71%)	0.62 (+252.58%)
`1531781` — 2026-01-23 18:49:24	24.80 (n/a)	24.59 (n/a)	24.55 (n/a)	24.40 (n/a)	0.18 (n/a)	24.79 (n/a)	24.58 (n/a)	24.53 (n/a)	24.38 (n/a)	0.18 (n/a)

matrix_vector_mul_2048x8192_1tsi_2048tso_1col0

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Throughput (max)	Throughput (mean)	Throughput (median)	Throughput (min)	Throughput (stddev)
`e199ef2` — 2026-02-06 17:34:56	13.46 (-0.61%)	12.60 (-5.52%)	12.76 (-5.10%)	10.95 (-16.40%)	0.97 (+370.33%)	13.45 (-0.61%)	12.60 (-5.52%)	12.75 (-5.10%)	10.94 (-16.40%)	0.97 (+370.33%)
`1531781` — 2026-01-23 18:49:24	13.54 (n/a)	13.34 (n/a)	13.45 (n/a)	13.10 (n/a)	0.21 (n/a)	13.53 (n/a)	13.33 (n/a)	13.44 (n/a)	13.09 (n/a)	0.21 (n/a)

matrix_vector_mul_2048x8192_1tsi_256tso_8col0

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Throughput (max)	Throughput (mean)	Throughput (median)	Throughput (min)	Throughput (stddev)
`e199ef2` — 2026-02-06 17:34:56	43.34 (-7.41%)	42.19 (-8.10%)	41.77 (-9.92%)	41.41 (-7.60%)	0.87 (-2.84%)	43.32 (-7.41%)	42.17 (-8.10%)	41.75 (-9.92%)	41.38 (-7.60%)	0.87 (-2.84%)
`1531781` — 2026-01-23 18:49:24	46.81 (n/a)	45.91 (n/a)	46.38 (n/a)	44.81 (n/a)	0.89 (n/a)	46.79 (n/a)	45.88 (n/a)	46.35 (n/a)	44.79 (n/a)	0.89 (n/a)

matrix_vector_mul_2048x8192_1tsi_512tso_4col0

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Throughput (max)	Throughput (mean)	Throughput (median)	Throughput (min)	Throughput (stddev)
`e199ef2` — 2026-02-06 17:34:56	41.95 (+1.11%)	39.55 (-3.08%)	39.11 (-4.91%)	38.20 (-4.37%)	1.42 (+118.78%)	41.92 (+1.11%)	39.53 (-3.08%)	39.08 (-4.91%)	38.18 (-4.37%)	1.42 (+118.78%)
`1531781` — 2026-01-23 18:49:24	41.49 (n/a)	40.81 (n/a)	41.13 (n/a)	39.95 (n/a)	0.65 (n/a)	41.46 (n/a)	40.79 (n/a)	41.10 (n/a)	39.93 (n/a)	0.65 (n/a)

matrix_vector_mul_8192x2048_4_1col

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)
`130b6ea` — 2025-12-05 21:15:49	13.29 (+1.63%)	13.19 (+6.08%)	13.19 (+1.17%)	13.09 (+26.86%)	0.08 (-92.94%)
`0a6c11c` — 2025-12-04 00:39:10	13.08 (n/a)	12.44 (n/a)	13.03 (n/a)	10.32 (n/a)	1.07 (n/a)

matrix_vector_mul_8192x2048_4_1col0

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Throughput (max)	Throughput (mean)	Throughput (median)	Throughput (min)	Throughput (stddev)
`331dcca` — 2026-01-08 18:34:10	13.68 (+3.42%)	13.30 (+8.54%)	13.38 (+7.25%)	12.72 (+14.96%)	0.35 (-66.09%)	13.67 (+3.42%)	13.29 (+8.54%)	13.38 (+7.25%)	12.71 (+14.96%)	0.35 (-66.09%)
`a4b6ffe` — 2026-01-05 19:24:05	13.23 (n/a)	12.25 (n/a)	12.48 (n/a)	11.07 (n/a)	1.04 (n/a)	13.22 (n/a)	12.24 (n/a)	12.47 (n/a)	11.06 (n/a)	1.04 (n/a)

matrix_vector_mul_8192x2048_4_2col

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)
`130b6ea` — 2025-12-05 21:15:49	25.58 (+4.24%)	24.23 (+6.59%)	24.03 (+0.29%)	23.58 (+37.29%)	0.70 (-74.97%)
`0a6c11c` — 2025-12-04 00:39:10	24.54 (n/a)	22.73 (n/a)	23.96 (n/a)	17.18 (n/a)	2.79 (n/a)

matrix_vector_mul_8192x2048_4_2col0

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Throughput (max)	Throughput (mean)	Throughput (median)	Throughput (min)	Throughput (stddev)
`331dcca` — 2026-01-08 18:34:10	25.08 (+1.32%)	24.68 (+13.42%)	24.61 (+3.26%)	24.47 (+48.32%)	0.24 (-93.56%)	25.07 (+1.32%)	24.67 (+13.42%)	24.59 (+3.26%)	24.46 (+48.32%)	0.24 (-93.56%)
`a4b6ffe` — 2026-01-05 19:24:05	24.76 (n/a)	21.76 (n/a)	23.83 (n/a)	16.50 (n/a)	3.74 (n/a)	24.74 (n/a)	21.75 (n/a)	23.82 (n/a)	16.49 (n/a)	3.73 (n/a)

matrix_vector_mul_8192x2048_4_4col

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)
`130b6ea` — 2025-12-05 21:15:49	40.94 (-2.19%)	39.74 (-1.10%)	39.76 (-0.60%)	38.17 (-2.14%)	1.04 (+4.39%)
`0a6c11c` — 2025-12-04 00:39:10	41.86 (n/a)	40.18 (n/a)	40.00 (n/a)	39.00 (n/a)	1.00 (n/a)

matrix_vector_mul_8192x2048_4_4col0

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Throughput (max)	Throughput (mean)	Throughput (median)	Throughput (min)	Throughput (stddev)
`331dcca` — 2026-01-08 18:34:10	40.74 (-1.09%)	37.16 (-7.15%)	40.25 (+1.47%)	24.51 (-37.44%)	7.07 (+736.13%)	40.72 (-1.09%)	37.13 (-7.15%)	40.23 (+1.47%)	24.50 (-37.44%)	7.07 (+736.13%)
`a4b6ffe` — 2026-01-05 19:24:05	41.19 (n/a)	40.02 (n/a)	39.67 (n/a)	39.18 (n/a)	0.85 (n/a)	41.17 (n/a)	39.99 (n/a)	39.64 (n/a)	39.16 (n/a)	0.85 (n/a)

matrix_vector_mul_8192x2048_4_8col

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)
`130b6ea` — 2025-12-05 21:15:49	46.71 (-1.04%)	44.91 (-2.34%)	44.71 (-2.98%)	43.80 (-1.44%)	0.96 (+6.93%)
`0a6c11c` — 2025-12-04 00:39:10	47.20 (n/a)	45.98 (n/a)	46.08 (n/a)	44.44 (n/a)	0.90 (n/a)

matrix_vector_mul_8192x2048_4_8col0

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Throughput (max)	Throughput (mean)	Throughput (median)	Throughput (min)	Throughput (stddev)
`331dcca` — 2026-01-08 18:34:10	46.28 (+5.68%)	44.89 (+14.59%)	44.74 (+4.26%)	44.20 (+62.80%)	0.84 (-87.96%)	46.25 (+5.68%)	44.86 (+14.59%)	44.71 (+4.26%)	44.17 (+62.80%)	0.84 (-87.96%)
`a4b6ffe` — 2026-01-05 19:24:05	43.79 (n/a)	39.17 (n/a)	42.91 (n/a)	27.15 (n/a)	6.98 (n/a)	43.77 (n/a)	39.15 (n/a)	42.88 (n/a)	27.13 (n/a)	6.98 (n/a)

matrix_vector_mul_8192x2048_4tsi_1024tso_1col0

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Throughput (max)	Throughput (mean)	Throughput (median)	Throughput (min)	Throughput (stddev)
`e199ef2` — 2026-02-06 17:34:56	13.66 (+1.48%)	12.91 (-3.11%)	13.09 (-1.74%)	12.18 (-7.72%)	0.58 (+414.89%)	13.65 (+1.48%)	12.90 (-3.11%)	13.08 (-1.74%)	12.17 (-7.72%)	0.58 (+414.89%)
`1531781` — 2026-01-23 18:49:24	13.46 (n/a)	13.33 (n/a)	13.32 (n/a)	13.20 (n/a)	0.11 (n/a)	13.45 (n/a)	13.32 (n/a)	13.31 (n/a)	13.19 (n/a)	0.11 (n/a)

matrix_vector_mul_8192x2048_4tsi_1024tso_2col0

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Throughput (max)	Throughput (mean)	Throughput (median)	Throughput (min)	Throughput (stddev)
`e199ef2` — 2026-02-06 17:34:56	24.54 (-2.87%)	23.84 (-3.69%)	23.79 (-3.34%)	23.01 (-6.24%)	0.57 (+91.92%)	24.53 (-2.87%)	23.82 (-3.69%)	23.77 (-3.34%)	23.00 (-6.24%)	0.57 (+91.92%)
`1531781` — 2026-01-23 18:49:24	25.27 (n/a)	24.75 (n/a)	24.61 (n/a)	24.55 (n/a)	0.30 (n/a)	25.25 (n/a)	24.74 (n/a)	24.59 (n/a)	24.53 (n/a)	0.30 (n/a)

matrix_vector_mul_8192x2048_4tsi_1024tso_4col0

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Throughput (max)	Throughput (mean)	Throughput (median)	Throughput (min)	Throughput (stddev)
`e199ef2` — 2026-02-06 17:34:56	40.21 (-1.59%)	37.97 (-5.08%)	39.00 (-3.55%)	32.17 (-17.23%)	3.31 (+273.80%)	40.19 (-1.59%)	37.95 (-5.08%)	38.97 (-3.55%)	32.15 (-17.23%)	3.31 (+273.80%)
`1531781` — 2026-01-23 18:49:24	40.86 (n/a)	40.01 (n/a)	40.43 (n/a)	38.86 (n/a)	0.89 (n/a)	40.84 (n/a)	39.98 (n/a)	40.41 (n/a)	38.84 (n/a)	0.89 (n/a)

matrix_vector_mul_8192x2048_4tsi_1024tso_8col0

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Throughput (max)	Throughput (mean)	Throughput (median)	Throughput (min)	Throughput (stddev)
`e199ef2` — 2026-02-06 17:34:56	45.08 (-2.18%)	42.30 (-6.21%)	41.61 (-8.52%)	41.35 (-5.80%)	1.58 (+62.90%)	45.05 (-2.18%)	42.28 (-6.21%)	41.59 (-8.52%)	41.32 (-5.80%)	1.57 (+62.90%)
`1531781` — 2026-01-23 18:49:24	46.08 (n/a)	45.11 (n/a)	45.49 (n/a)	43.89 (n/a)	0.97 (n/a)	46.05 (n/a)	45.08 (n/a)	45.46 (n/a)	43.87 (n/a)	0.97 (n/a)

mem_copy_16_cores_2_chans_2048_tile_128_False

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`130b6ea` — 2025-12-05 21:15:49	0.07 (+22.03%)	0.06 (+27.86%)	0.06 (+35.49%)	0.05 (+15.00%)	0.01 (+49.69%)	173.30 (-13.00%)	137.94 (-21.32%)	133.60 (-26.19%)	123.40 (-18.06%)	18.40 (+8.53%)
`0a6c11c` — 2025-12-04 00:39:10	0.05 (n/a)	0.05 (n/a)	0.05 (n/a)	0.04 (n/a)	0.00 (n/a)	199.20 (n/a)	175.32 (n/a)	181.00 (n/a)	150.60 (n/a)	16.96 (n/a)

mem_copy_16_cores_2_chans_2048_tile_128_False0

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`e199ef2` — 2026-02-06 17:34:56	0.05 (-52.48%)	0.04 (-54.29%)	0.04 (-55.70%)	0.02 (-49.89%)	0.01 (-57.40%)	389.80 (+99.59%)	238.94 (+114.07%)	195.50 (+125.75%)	178.10 (+110.52%)	87.41 (+83.42%)
`1531781` — 2026-01-23 18:49:24	0.10 (n/a)	0.08 (n/a)	0.09 (n/a)	0.04 (n/a)	0.02 (n/a)	195.30 (n/a)	111.62 (n/a)	86.60 (n/a)	84.60 (n/a)	47.65 (n/a)

mem_copy_1_cols_1_channels_2048_tile_2048

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`003b49d` — 2025-11-14 19:30:54	0.05 (-16.99%)	0.05 (-12.47%)	0.05 (-7.32%)	0.04 (-11.17%)	0.00 (-29.61%)	197.00 (+12.57%)	171.00 (+13.70%)	164.00 (+7.89%)	153.00 (+20.47%)	17.49 (-4.32%)
`0999a9a` — 2025-11-13 19:24:08	0.06 (n/a)	0.05 (n/a)	0.05 (n/a)	0.04 (n/a)	0.01 (n/a)	175.00 (n/a)	150.40 (n/a)	152.00 (n/a)	127.00 (n/a)	18.28 (n/a)

mem_copy_1_cols_2_channels_2048_tile_1024

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`003b49d` — 2025-11-14 19:30:54	0.06 (+5.56%)	0.05 (-1.59%)	0.05 (-1.39%)	0.03 (-14.07%)	0.01 (+25.03%)	263.00 (+16.37%)	166.00 (+4.53%)	144.00 (+1.41%)	126.00 (-5.26%)	49.38 (+43.17%)
`0999a9a` — 2025-11-13 19:24:08	0.06 (n/a)	0.05 (n/a)	0.05 (n/a)	0.03 (n/a)	0.01 (n/a)	226.00 (n/a)	158.80 (n/a)	142.00 (n/a)	133.00 (n/a)	34.49 (n/a)

mem_copy_1_cores_1_chans_2048_tile_2048_False

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`130b6ea` — 2025-12-05 21:15:49	0.08 (+10.93%)	0.07 (+7.79%)	0.07 (+3.55%)	0.05 (-1.73%)	0.01 (+36.73%)	150.20 (+1.76%)	121.68 (-6.36%)	125.60 (-3.38%)	97.90 (-9.85%)	18.56 (+21.98%)
`0a6c11c` — 2025-12-04 00:39:10	0.08 (n/a)	0.06 (n/a)	0.06 (n/a)	0.06 (n/a)	0.01 (n/a)	147.60 (n/a)	129.94 (n/a)	130.00 (n/a)	108.60 (n/a)	15.21 (n/a)

mem_copy_1_cores_1_chans_2048_tile_2048_False0

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`e199ef2` — 2026-02-06 17:34:56	0.06 (-38.96%)	0.05 (-38.35%)	0.06 (-42.97%)	0.04 (-38.23%)	0.01 (-55.18%)	191.20 (+61.90%)	154.62 (+60.19%)	148.10 (+75.47%)	134.90 (+63.71%)	22.15 (+21.10%)
`1531781` — 2026-01-23 18:49:24	0.10 (n/a)	0.09 (n/a)	0.10 (n/a)	0.07 (n/a)	0.02 (n/a)	118.10 (n/a)	96.52 (n/a)	84.40 (n/a)	82.40 (n/a)	18.29 (n/a)

mem_copy_2_cols_1_channels_2048_tile_1024

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`003b49d` — 2025-11-14 19:30:54	0.06 (-15.32%)	0.05 (-17.09%)	0.05 (-23.81%)	0.04 (-15.54%)	0.01 (-18.18%)	193.00 (+18.40%)	160.80 (+20.36%)	168.00 (+31.25%)	124.00 (+18.10%)	22.92 (+11.39%)
`0999a9a` — 2025-11-13 19:24:08	0.07 (n/a)	0.06 (n/a)	0.06 (n/a)	0.05 (n/a)	0.01 (n/a)	163.00 (n/a)	133.60 (n/a)	128.00 (n/a)	105.00 (n/a)	20.58 (n/a)

mem_copy_2_cols_2_channels_2048_tile_512

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`003b49d` — 2025-11-14 19:30:54	0.06 (-23.02%)	0.05 (-6.98%)	0.05 (-2.13%)	0.04 (-15.31%)	0.01 (-33.29%)	196.00 (+18.07%)	147.00 (+6.52%)	141.00 (+2.17%)	126.00 (+29.90%)	25.27 (+8.13%)
`0999a9a` — 2025-11-13 19:24:08	0.08 (n/a)	0.06 (n/a)	0.06 (n/a)	0.05 (n/a)	0.01 (n/a)	166.00 (n/a)	138.00 (n/a)	138.00 (n/a)	97.00 (n/a)	23.37 (n/a)

mem_copy_2_cores_1_chans_2048_tile_1024_False

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`130b6ea` — 2025-12-05 21:15:49	0.10 (+31.37%)	0.08 (+29.12%)	0.08 (+29.06%)	0.07 (+41.80%)	0.01 (+0.23%)	109.90 (-29.46%)	98.30 (-23.38%)	100.80 (-22.52%)	79.40 (-23.87%)	10.54 (-46.17%)
`0a6c11c` — 2025-12-04 00:39:10	0.08 (n/a)	0.07 (n/a)	0.06 (n/a)	0.05 (n/a)	0.01 (n/a)	155.80 (n/a)	128.30 (n/a)	130.10 (n/a)	104.30 (n/a)	19.58 (n/a)

mem_copy_2_cores_1_chans_2048_tile_1024_False0

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`e199ef2` — 2026-02-06 17:34:56	0.05 (-53.23%)	0.05 (-47.65%)	0.05 (-47.49%)	0.04 (-39.57%)	0.00 (-79.69%)	193.80 (+65.50%)	168.64 (+82.91%)	165.50 (+90.45%)	150.40 (+113.64%)	16.77 (-28.06%)
`1531781` — 2026-01-23 18:49:24	0.12 (n/a)	0.09 (n/a)	0.09 (n/a)	0.07 (n/a)	0.02 (n/a)	117.10 (n/a)	92.20 (n/a)	86.90 (n/a)	70.40 (n/a)	23.32 (n/a)

mem_copy_2_cores_2_chans_2048_tile_1024_False

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`130b6ea` — 2025-12-05 21:15:49	0.10 (+25.02%)	0.07 (+10.01%)	0.07 (+7.53%)	0.05 (-1.99%)	0.01 (+77.08%)	150.10 (+2.04%)	114.60 (-7.22%)	111.70 (-6.99%)	83.90 (-20.02%)	21.23 (+43.54%)
`0a6c11c` — 2025-12-04 00:39:10	0.08 (n/a)	0.07 (n/a)	0.07 (n/a)	0.06 (n/a)	0.01 (n/a)	147.10 (n/a)	123.52 (n/a)	120.10 (n/a)	104.90 (n/a)	14.79 (n/a)

mem_copy_2_cores_2_chans_2048_tile_1024_False0

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`e199ef2` — 2026-02-06 17:34:56	0.06 (-39.60%)	0.05 (-39.55%)	0.06 (-34.45%)	0.04 (-39.20%)	0.01 (-32.95%)	230.60 (+64.48%)	169.56 (+66.79%)	140.40 (+52.44%)	137.30 (+65.62%)	43.58 (+79.85%)
`1531781` — 2026-01-23 18:49:24	0.10 (n/a)	0.08 (n/a)	0.09 (n/a)	0.06 (n/a)	0.02 (n/a)	140.20 (n/a)	101.66 (n/a)	92.10 (n/a)	82.90 (n/a)	24.23 (n/a)

mem_copy_4_cols_1_channels_2048_tile_512

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`003b49d` — 2025-11-14 19:30:54	0.07 (+11.97%)	0.05 (+3.09%)	0.05 (+2.10%)	0.03 (-11.35%)	0.01 (+41.65%)	229.00 (+12.81%)	156.00 (-0.51%)	143.00 (-2.05%)	117.00 (-10.69%)	38.21 (+48.71%)
`0999a9a` — 2025-11-13 19:24:08	0.06 (n/a)	0.05 (n/a)	0.05 (n/a)	0.04 (n/a)	0.01 (n/a)	203.00 (n/a)	156.80 (n/a)	146.00 (n/a)	131.00 (n/a)	25.69 (n/a)

mem_copy_4_cols_2_channels_2048_tile_256

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`003b49d` — 2025-11-14 19:30:54	0.06 (+9.02%)	0.05 (+5.59%)	0.05 (+0.00%)	0.04 (+3.63%)	0.01 (+25.48%)	193.00 (-3.50%)	147.40 (-4.66%)	147.00 (+0.00%)	122.00 (-8.27%)	25.24 (+7.48%)
`0999a9a` — 2025-11-13 19:24:08	0.06 (n/a)	0.05 (n/a)	0.05 (n/a)	0.04 (n/a)	0.01 (n/a)	200.00 (n/a)	154.60 (n/a)	147.00 (n/a)	133.00 (n/a)	23.48 (n/a)

mem_copy_4_cores_1_chans_2048_tile_512_False

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`130b6ea` — 2025-12-05 21:15:49	0.09 (-3.98%)	0.07 (+4.00%)	0.07 (+7.04%)	0.06 (+3.67%)	0.01 (-15.76%)	129.70 (-3.57%)	113.30 (-4.50%)	116.90 (-6.55%)	90.40 (+4.15%)	15.04 (-13.80%)
`0a6c11c` — 2025-12-04 00:39:10	0.09 (n/a)	0.07 (n/a)	0.07 (n/a)	0.06 (n/a)	0.01 (n/a)	134.50 (n/a)	118.64 (n/a)	125.10 (n/a)	86.80 (n/a)	17.45 (n/a)

mem_copy_4_cores_1_chans_2048_tile_512_False0

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`e199ef2` — 2026-02-06 17:34:56	0.05 (-57.30%)	0.05 (-54.11%)	0.05 (-55.46%)	0.04 (-37.48%)	0.01 (-74.40%)	205.30 (+60.02%)	175.72 (+107.81%)	160.70 (+124.44%)	153.80 (+134.09%)	25.82 (-2.26%)
`1531781` — 2026-01-23 18:49:24	0.12 (n/a)	0.10 (n/a)	0.11 (n/a)	0.06 (n/a)	0.03 (n/a)	128.30 (n/a)	84.56 (n/a)	71.60 (n/a)	65.70 (n/a)	26.42 (n/a)

mem_copy_4_cores_2_chans_2048_tile_512_False

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`130b6ea` — 2025-12-05 21:15:49	0.08 (+9.57%)	0.07 (+16.85%)	0.07 (+19.58%)	0.06 (+24.52%)	0.01 (+1.05%)	134.50 (-19.70%)	116.64 (-14.92%)	116.80 (-16.39%)	98.00 (-8.75%)	15.71 (-25.13%)
`0a6c11c` — 2025-12-04 00:39:10	0.08 (n/a)	0.06 (n/a)	0.06 (n/a)	0.05 (n/a)	0.01 (n/a)	167.50 (n/a)	137.10 (n/a)	139.70 (n/a)	107.40 (n/a)	20.99 (n/a)

mem_copy_4_cores_2_chans_2048_tile_512_False0

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`e199ef2` — 2026-02-06 17:34:56	0.06 (-43.91%)	0.05 (-48.82%)	0.04 (-55.46%)	0.04 (-35.80%)	0.01 (-53.53%)	207.40 (+55.71%)	179.02 (+92.54%)	187.20 (+124.46%)	146.40 (+78.32%)	28.55 (+26.93%)
`1531781` — 2026-01-23 18:49:24	0.10 (n/a)	0.09 (n/a)	0.10 (n/a)	0.06 (n/a)	0.02 (n/a)	133.20 (n/a)	92.98 (n/a)	83.40 (n/a)	82.10 (n/a)	22.49 (n/a)

mem_copy_8_cols_1_channels_2048_tile_256

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`003b49d` — 2025-11-14 19:30:54	0.05 (-13.48%)	0.05 (-17.79%)	0.05 (-15.33%)	0.03 (-25.09%)	0.01 (+6.67%)	267.00 (+33.50%)	176.80 (+24.51%)	150.00 (+18.11%)	141.00 (+15.57%)	47.41 (+61.41%)
`0999a9a` — 2025-11-13 19:24:08	0.06 (n/a)	0.06 (n/a)	0.06 (n/a)	0.04 (n/a)	0.01 (n/a)	200.00 (n/a)	142.00 (n/a)	127.00 (n/a)	122.00 (n/a)	29.37 (n/a)

mem_copy_8_cols_2_channels_2048_tile_128

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`003b49d` — 2025-11-14 19:30:54	0.05 (+13.38%)	0.05 (+23.85%)	0.05 (+37.91%)	0.05 (+30.72%)	0.00 (-56.85%)	166.00 (-23.50%)	153.40 (-20.60%)	153.00 (-27.49%)	142.00 (-11.80%)	7.63 (-70.60%)
`0999a9a` — 2025-11-13 19:24:08	0.05 (n/a)	0.04 (n/a)	0.04 (n/a)	0.04 (n/a)	0.01 (n/a)	217.00 (n/a)	193.20 (n/a)	211.00 (n/a)	161.00 (n/a)	25.96 (n/a)

mem_copy_8_cores_1_chans_2048_tile_256_False

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`130b6ea` — 2025-12-05 21:15:49	0.09 (+16.31%)	0.08 (+15.34%)	0.08 (+14.60%)	0.07 (+19.30%)	0.01 (-5.41%)	117.30 (-16.15%)	107.80 (-13.61%)	108.50 (-12.78%)	94.10 (-13.99%)	8.29 (-32.01%)
`0a6c11c` — 2025-12-04 00:39:10	0.07 (n/a)	0.07 (n/a)	0.07 (n/a)	0.06 (n/a)	0.01 (n/a)	139.90 (n/a)	124.78 (n/a)	124.40 (n/a)	109.40 (n/a)	12.20 (n/a)

mem_copy_8_cores_1_chans_2048_tile_256_False0

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`e199ef2` — 2026-02-06 17:34:56	0.06 (-41.48%)	0.04 (-54.05%)	0.04 (-59.22%)	0.02 (-66.46%)	0.01 (+0.87%)	391.50 (+198.17%)	238.42 (+138.95%)	217.40 (+145.10%)	148.50 (+70.89%)	97.26 (+414.52%)
`1531781` — 2026-01-23 18:49:24	0.09 (n/a)	0.08 (n/a)	0.09 (n/a)	0.06 (n/a)	0.01 (n/a)	131.30 (n/a)	99.78 (n/a)	88.70 (n/a)	86.90 (n/a)	18.90 (n/a)

mem_copy_8_cores_2_chans_2048_tile_256_False

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`130b6ea` — 2025-12-05 21:15:49	0.09 (+12.61%)	0.07 (-2.24%)	0.07 (-8.85%)	0.06 (+5.13%)	0.01 (+35.50%)	136.50 (-4.88%)	120.86 (+2.96%)	125.10 (+9.64%)	91.60 (-11.15%)	15.72 (+10.46%)
`0a6c11c` — 2025-12-04 00:39:10	0.08 (n/a)	0.07 (n/a)	0.07 (n/a)	0.06 (n/a)	0.01 (n/a)	143.50 (n/a)	117.38 (n/a)	114.10 (n/a)	103.10 (n/a)	14.23 (n/a)

mem_copy_8_cores_2_chans_2048_tile_256_False0

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`e199ef2` — 2026-02-06 17:34:56	0.05 (-49.15%)	0.04 (-57.03%)	0.04 (-55.46%)	0.03 (-54.45%)	0.01 (-36.53%)	317.10 (+119.60%)	234.80 (+140.33%)	197.90 (+124.63%)	157.40 (+96.75%)	73.16 (+177.70%)
`1531781` — 2026-01-23 18:49:24	0.10 (n/a)	0.09 (n/a)	0.09 (n/a)	0.06 (n/a)	0.02 (n/a)	144.40 (n/a)	97.70 (n/a)	88.10 (n/a)	80.00 (n/a)	26.35 (n/a)

mha

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`130b6ea` — 2025-12-05 21:15:49	0.18 (+0.16%)	0.18 (+0.19%)	0.18 (+0.12%)	0.18 (+0.32%)	0.00 (-52.24%)	47253.90 (-0.32%)	47213.70 (-0.19%)	47210.90 (-0.12%)	47167.70 (-0.16%)	28.71 (-52.47%)
`0a6c11c` — 2025-12-04 00:39:10	0.18 (n/a)	0.18 (n/a)	0.18 (n/a)	0.18 (n/a)	0.00 (n/a)	47405.00 (n/a)	47301.72 (n/a)	47265.90 (n/a)	47241.80 (n/a)	60.40 (n/a)

mha0

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`e199ef2` — 2026-02-06 17:34:56	0.18 (-0.81%)	0.18 (-0.82%)	0.18 (-0.78%)	0.18 (-0.78%)	0.00 (+10.80%)	47741.40 (+0.79%)	47625.06 (+0.83%)	47605.30 (+0.79%)	47457.20 (+0.82%)	118.82 (+12.68%)
`1531781` — 2026-01-23 18:49:24	0.18 (n/a)	0.18 (n/a)	0.18 (n/a)	0.18 (n/a)	0.00 (n/a)	47367.50 (n/a)	47232.68 (n/a)	47232.00 (n/a)	47073.00 (n/a)	105.45 (n/a)

relu_1_cols_1_channels_2048_tile_2048

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`e199ef2` — 2026-02-06 17:34:56	0.07 (-36.45%)	0.05 (-41.06%)	0.05 (-47.87%)	0.04 (-28.25%)	0.01 (-54.56%)	213.10 (+39.37%)	164.34 (+61.72%)	157.60 (+91.73%)	114.90 (+57.40%)	35.91 (+0.45%)
`1531781` — 2026-01-23 18:49:24	0.11 (n/a)	0.09 (n/a)	0.10 (n/a)	0.05 (n/a)	0.03 (n/a)	152.90 (n/a)	101.62 (n/a)	82.20 (n/a)	73.00 (n/a)	35.75 (n/a)

relu_2_cols_1_channels_2048_tile_1024

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`e199ef2` — 2026-02-06 17:34:56	0.06 (-39.42%)	0.05 (-43.39%)	0.05 (-43.82%)	0.03 (-37.60%)	0.01 (-44.12%)	241.30 (+60.23%)	177.30 (+75.09%)	162.60 (+78.09%)	137.10 (+65.18%)	40.73 (+45.37%)
`1531781` — 2026-01-23 18:49:24	0.10 (n/a)	0.08 (n/a)	0.09 (n/a)	0.05 (n/a)	0.02 (n/a)	150.60 (n/a)	101.26 (n/a)	91.30 (n/a)	83.00 (n/a)	28.02 (n/a)

relu_4_cols_1_channels_2048_tile_512

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`e199ef2` — 2026-02-06 17:34:56	0.05 (-45.81%)	0.05 (-43.02%)	0.05 (-46.87%)	0.05 (-21.63%)	0.00 (-82.37%)	173.30 (+27.61%)	163.50 (+69.96%)	162.40 (+88.18%)	153.30 (+84.48%)	9.24 (-58.84%)
`1531781` — 2026-01-23 18:49:24	0.10 (n/a)	0.09 (n/a)	0.09 (n/a)	0.06 (n/a)	0.02 (n/a)	135.80 (n/a)	96.20 (n/a)	86.30 (n/a)	83.10 (n/a)	22.46 (n/a)

relu_8_cols_1_channels_2048_tile_256

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`e199ef2` — 2026-02-06 17:34:56	0.04 (-59.76%)	0.04 (-56.17%)	0.04 (-60.87%)	0.04 (-23.81%)	0.00 (-94.70%)	202.70 (+31.28%)	197.58 (+111.72%)	200.30 (+155.48%)	188.40 (+148.55%)	5.79 (-83.10%)
`1531781` — 2026-01-23 18:49:24	0.11 (n/a)	0.09 (n/a)	0.10 (n/a)	0.05 (n/a)	0.02 (n/a)	154.40 (n/a)	93.32 (n/a)	78.40 (n/a)	75.80 (n/a)	34.24 (n/a)

rms_norm_1_cols_1_channels_2048_tile_2048

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`e199ef2` — 2026-02-06 17:34:56	0.07 (-19.44%)	0.05 (-37.71%)	0.05 (-45.73%)	0.03 (-52.32%)	0.02 (+44.28%)	275.90 (+109.81%)	174.62 (+73.41%)	170.20 (+84.20%)	112.80 (+24.09%)	64.04 (+269.22%)
`1531781` — 2026-01-23 18:49:24	0.09 (n/a)	0.08 (n/a)	0.09 (n/a)	0.06 (n/a)	0.01 (n/a)	131.50 (n/a)	100.70 (n/a)	92.40 (n/a)	90.90 (n/a)	17.35 (n/a)

rms_norm_1_cols_2_channels_2048_tile_1024

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`e199ef2` — 2026-02-06 17:34:56	0.05 (-53.33%)	0.05 (-55.81%)	0.04 (-63.93%)	0.04 (-30.16%)	0.01 (-71.61%)	207.90 (+43.18%)	184.54 (+113.24%)	199.80 (+177.12%)	151.20 (+114.47%)	28.21 (-13.99%)
`1531781` — 2026-01-23 18:49:24	0.12 (n/a)	0.10 (n/a)	0.11 (n/a)	0.06 (n/a)	0.03 (n/a)	145.20 (n/a)	86.54 (n/a)	72.10 (n/a)	70.50 (n/a)	32.80 (n/a)

rms_norm_2_cols_1_channels_2048_tile_1024

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`e199ef2` — 2026-02-06 17:34:56	0.05 (-43.53%)	0.05 (-45.34%)	0.05 (-47.93%)	0.04 (-39.19%)	0.01 (-57.71%)	213.10 (+64.43%)	177.88 (+81.10%)	174.50 (+91.97%)	159.00 (+77.06%)	21.59 (+23.04%)
`1531781` — 2026-01-23 18:49:24	0.09 (n/a)	0.09 (n/a)	0.09 (n/a)	0.06 (n/a)	0.01 (n/a)	129.60 (n/a)	98.22 (n/a)	90.90 (n/a)	89.80 (n/a)	17.55 (n/a)

rms_norm_2_cols_2_channels_2048_tile_512

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`e199ef2` — 2026-02-06 17:34:56	0.06 (-57.03%)	0.04 (-63.10%)	0.04 (-63.23%)	0.02 (-72.10%)	0.01 (-26.15%)	335.00 (+258.29%)	213.72 (+187.57%)	198.80 (+171.96%)	145.20 (+132.69%)	72.54 (+525.20%)
`1531781` — 2026-01-23 18:49:24	0.13 (n/a)	0.11 (n/a)	0.11 (n/a)	0.09 (n/a)	0.02 (n/a)	93.50 (n/a)	74.32 (n/a)	73.10 (n/a)	62.40 (n/a)	11.60 (n/a)

rms_norm_4_cols_1_channels_2048_tile_512

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`e199ef2` — 2026-02-06 17:34:56	0.06 (-37.47%)	0.04 (-49.13%)	0.04 (-54.30%)	0.03 (-44.55%)	0.01 (-29.03%)	235.80 (+80.41%)	193.18 (+98.17%)	197.40 (+118.85%)	137.60 (+60.00%)	35.78 (+91.73%)
`1531781` — 2026-01-23 18:49:24	0.10 (n/a)	0.09 (n/a)	0.09 (n/a)	0.06 (n/a)	0.01 (n/a)	130.70 (n/a)	97.48 (n/a)	90.20 (n/a)	86.00 (n/a)	18.66 (n/a)

rms_norm_4_cols_2_channels_2048_tile_256

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`e199ef2` — 2026-02-06 17:34:56	0.05 (-57.63%)	0.04 (-61.87%)	0.04 (-59.72%)	0.02 (-67.97%)	0.01 (-45.26%)	392.00 (+212.35%)	231.42 (+175.83%)	182.40 (+148.16%)	172.10 (+136.08%)	93.34 (+301.27%)
`1531781` — 2026-01-23 18:49:24	0.11 (n/a)	0.10 (n/a)	0.11 (n/a)	0.07 (n/a)	0.02 (n/a)	125.50 (n/a)	83.90 (n/a)	73.50 (n/a)	72.90 (n/a)	23.26 (n/a)

rms_norm_8_cols_1_channels_2048_tile_256

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`e199ef2` — 2026-02-06 17:34:56	0.05 (-47.54%)	0.04 (-48.46%)	0.04 (-54.55%)	0.04 (-29.18%)	0.00 (-76.66%)	217.90 (+41.22%)	200.30 (+87.79%)	208.60 (+120.04%)	179.30 (+90.74%)	16.54 (-37.92%)
`1531781` — 2026-01-23 18:49:24	0.09 (n/a)	0.08 (n/a)	0.09 (n/a)	0.05 (n/a)	0.01 (n/a)	154.30 (n/a)	106.66 (n/a)	94.80 (n/a)	94.00 (n/a)	26.64 (n/a)

rms_norm_8_cols_2_channels_2048_tile_128

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`e199ef2` — 2026-02-06 17:34:56	0.05 (-43.58%)	0.04 (-50.82%)	0.04 (-59.44%)	0.03 (-29.32%)	0.01 (-59.77%)	237.20 (+41.44%)	214.88 (+96.49%)	225.80 (+146.51%)	157.60 (+77.28%)	32.93 (-1.75%)
`1531781` — 2026-01-23 18:49:24	0.09 (n/a)	0.08 (n/a)	0.09 (n/a)	0.05 (n/a)	0.02 (n/a)	167.70 (n/a)	109.36 (n/a)	91.60 (n/a)	88.90 (n/a)	33.52 (n/a)

rope_1_cols_2_channels_4096_tile_4096_0

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`1b2d230` — 2026-01-13 22:38:03	0.19 (-21.66%)	0.17 (-8.53%)	0.18 (+2.09%)	0.14 (-6.71%)	0.02 (-38.07%)	174.10 (+7.20%)	150.06 (+8.11%)	139.60 (-2.04%)	129.80 (+27.63%)	20.36 (-10.65%)
`d48746f` — 2026-01-12 22:17:10	0.24 (n/a)	0.18 (n/a)	0.17 (n/a)	0.15 (n/a)	0.04 (n/a)	162.40 (n/a)	138.80 (n/a)	142.50 (n/a)	101.70 (n/a)	22.79 (n/a)

rope_1c_32rows_512cols_32arows_0m

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`e199ef2` — 2026-02-06 17:34:56	0.74 (-32.54%)	0.55 (-44.06%)	0.46 (-56.74%)	0.44 (-40.68%)	0.15 (+0.10%)	225.50 (+68.54%)	187.58 (+84.74%)	215.80 (+131.30%)	132.10 (+48.26%)	46.35 (+149.23%)
`1531781` — 2026-01-23 18:49:24	1.10 (n/a)	0.99 (n/a)	1.05 (n/a)	0.73 (n/a)	0.15 (n/a)	133.80 (n/a)	101.54 (n/a)	93.30 (n/a)	89.10 (n/a)	18.60 (n/a)

rope_1c_32rows_512cols_8arows_0m

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`e199ef2` — 2026-02-06 17:34:56	0.61 (+10.25%)	0.50 (-5.13%)	0.49 (-7.43%)	0.43 (-10.58%)	0.07 (+127.99%)	171.90 (+11.84%)	150.18 (+6.63%)	151.40 (+8.07%)	120.20 (-9.28%)	18.94 (+125.53%)
`1531781` — 2026-01-23 18:49:24	0.56 (n/a)	0.52 (n/a)	0.53 (n/a)	0.48 (n/a)	0.03 (n/a)	153.70 (n/a)	140.84 (n/a)	140.10 (n/a)	132.50 (n/a)	8.40 (n/a)

rope_2_cols_2_channels_4096_tile_2048_0

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`1b2d230` — 2026-01-13 22:38:03	0.21 (-0.56%)	0.17 (-10.62%)	0.16 (-13.58%)	0.13 (-13.33%)	0.03 (+17.05%)	183.70 (+15.39%)	151.76 (+12.77%)	155.20 (+15.73%)	117.30 (+0.60%)	24.24 (+35.73%)
`d48746f` — 2026-01-12 22:17:10	0.21 (n/a)	0.19 (n/a)	0.18 (n/a)	0.15 (n/a)	0.02 (n/a)	159.20 (n/a)	134.58 (n/a)	134.10 (n/a)	116.60 (n/a)	17.86 (n/a)

rope_2c_32rows_512cols_32arows_0m

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`e199ef2` — 2026-02-06 17:34:56	0.78 (-37.24%)	0.58 (-39.49%)	0.54 (-40.46%)	0.48 (-22.64%)	0.12 (-52.46%)	202.90 (+29.24%)	173.08 (+60.11%)	181.90 (+67.96%)	125.40 (+59.34%)	30.10 (-3.19%)
`1531781` — 2026-01-23 18:49:24	1.25 (n/a)	0.97 (n/a)	0.91 (n/a)	0.63 (n/a)	0.25 (n/a)	157.00 (n/a)	108.10 (n/a)	108.30 (n/a)	78.70 (n/a)	31.09 (n/a)

rope_2c_32rows_512cols_8arows_0m

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`e199ef2` — 2026-02-06 17:34:56	0.59 (-6.71%)	0.41 (-27.04%)	0.38 (-33.40%)	0.31 (-36.01%)	0.11 (+106.38%)	235.30 (+56.24%)	186.20 (+42.59%)	195.70 (+50.08%)	125.20 (+7.19%)	41.86 (+233.14%)
`1531781` — 2026-01-23 18:49:24	0.63 (n/a)	0.57 (n/a)	0.57 (n/a)	0.49 (n/a)	0.05 (n/a)	150.60 (n/a)	130.58 (n/a)	130.40 (n/a)	116.80 (n/a)	12.57 (n/a)

rope_4_cols_2_channels_4096_tile_1024_0

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`1b2d230` — 2026-01-13 22:38:03	0.22 (+18.92%)	0.17 (+6.24%)	0.17 (+5.57%)	0.14 (-0.50%)	0.03 (+90.77%)	171.10 (+0.47%)	147.40 (-4.39%)	148.10 (-5.31%)	110.10 (-15.89%)	23.59 (+60.67%)
`d48746f` — 2026-01-12 22:17:10	0.19 (n/a)	0.16 (n/a)	0.16 (n/a)	0.14 (n/a)	0.02 (n/a)	170.30 (n/a)	154.16 (n/a)	156.40 (n/a)	130.90 (n/a)	14.69 (n/a)

rope_8_cols_2_channels_4096_tile_512_0

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`1b2d230` — 2026-01-13 22:38:03	0.17 (-1.30%)	0.15 (+6.63%)	0.15 (-8.78%)	0.14 (+39.68%)	0.01 (-67.35%)	174.90 (-28.41%)	163.42 (-10.36%)	167.40 (+9.63%)	146.50 (+1.31%)	11.19 (-76.17%)
`d48746f` — 2026-01-12 22:17:10	0.17 (n/a)	0.14 (n/a)	0.16 (n/a)	0.10 (n/a)	0.03 (n/a)	244.30 (n/a)	182.30 (n/a)	152.70 (n/a)	144.60 (n/a)	46.98 (n/a)

rope_8c_32rows_512cols_32arows_0m

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`e199ef2` — 2026-02-06 17:34:56	0.71 (-46.62%)	0.55 (-38.20%)	0.51 (-40.04%)	0.47 (-21.44%)	0.10 (-64.74%)	208.10 (+27.28%)	181.56 (+54.68%)	192.60 (+66.75%)	137.80 (+87.23%)	27.14 (-15.39%)
`1531781` — 2026-01-23 18:49:24	1.34 (n/a)	0.89 (n/a)	0.85 (n/a)	0.60 (n/a)	0.27 (n/a)	163.50 (n/a)	117.38 (n/a)	115.50 (n/a)	73.60 (n/a)	32.08 (n/a)

rope_8c_32rows_512cols_8arows_0m

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`e199ef2` — 2026-02-06 17:34:56	0.45 (-16.91%)	0.36 (-29.72%)	0.40 (-24.95%)	0.22 (-53.50%)	0.09 (+230.11%)	328.80 (+115.04%)	216.00 (+50.71%)	186.20 (+33.29%)	163.10 (+20.37%)	66.12 (+785.74%)
`1531781` — 2026-01-23 18:49:24	0.54 (n/a)	0.52 (n/a)	0.53 (n/a)	0.48 (n/a)	0.03 (n/a)	152.90 (n/a)	143.32 (n/a)	139.70 (n/a)	135.50 (n/a)	7.46 (n/a)

sigmoid_1_cols_1_channels_2048_tile_2048

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`e199ef2` — 2026-02-06 17:34:56	0.09 (-25.37%)	0.06 (-40.58%)	0.05 (-54.72%)	0.04 (-25.70%)	0.02 (-36.96%)	206.20 (+34.60%)	148.74 (+63.59%)	151.40 (+120.70%)	91.60 (+34.11%)	41.94 (+13.98%)
`1531781` — 2026-01-23 18:49:24	0.12 (n/a)	0.10 (n/a)	0.12 (n/a)	0.05 (n/a)	0.03 (n/a)	153.20 (n/a)	90.92 (n/a)	68.60 (n/a)	68.30 (n/a)	36.80 (n/a)

sigmoid_2_cols_1_channels_2048_tile_1024

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`e199ef2` — 2026-02-06 17:34:56	0.07 (-44.51%)	0.06 (-42.22%)	0.06 (-46.65%)	0.04 (-31.86%)	0.01 (-58.41%)	206.70 (+46.80%)	148.86 (+65.69%)	129.20 (+87.52%)	122.80 (+80.32%)	36.16 (+12.84%)
`1531781` — 2026-01-23 18:49:24	0.12 (n/a)	0.10 (n/a)	0.12 (n/a)	0.06 (n/a)	0.03 (n/a)	140.80 (n/a)	89.84 (n/a)	68.90 (n/a)	68.10 (n/a)	32.04 (n/a)

sigmoid_4_cols_1_channels_2048_tile_512

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`e199ef2` — 2026-02-06 17:34:56	0.08 (-14.28%)	0.06 (-31.63%)	0.05 (-46.93%)	0.05 (-5.75%)	0.02 (-25.97%)	179.80 (+6.14%)	152.16 (+42.77%)	164.70 (+88.44%)	98.00 (+16.67%)	31.72 (-12.77%)
`1531781` — 2026-01-23 18:49:24	0.10 (n/a)	0.08 (n/a)	0.09 (n/a)	0.05 (n/a)	0.02 (n/a)	169.40 (n/a)	106.58 (n/a)	87.40 (n/a)	84.00 (n/a)	36.37 (n/a)

sigmoid_8_cols_1_channels_2048_tile_256

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`e199ef2` — 2026-02-06 17:34:56	0.07 (-46.29%)	0.05 (-47.43%)	0.05 (-56.60%)	0.04 (-23.19%)	0.01 (-63.55%)	204.80 (+30.20%)	158.88 (+77.04%)	158.30 (+130.42%)	122.80 (+86.34%)	33.53 (-13.68%)
`1531781` — 2026-01-23 18:49:24	0.12 (n/a)	0.10 (n/a)	0.12 (n/a)	0.05 (n/a)	0.03 (n/a)	157.30 (n/a)	89.74 (n/a)	68.70 (n/a)	65.90 (n/a)	38.85 (n/a)

silu_1_cols_1_channels_2048_tile_2048

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`e199ef2` — 2026-02-06 17:34:56	0.07 (-28.99%)	0.05 (-36.40%)	0.07 (-34.42%)	0.03 (-56.09%)	0.02 (-10.88%)	326.40 (+127.77%)	181.20 (+72.77%)	125.40 (+52.55%)	113.50 (+40.82%)	91.21 (+182.94%)
`1531781` — 2026-01-23 18:49:24	0.10 (n/a)	0.08 (n/a)	0.10 (n/a)	0.06 (n/a)	0.02 (n/a)	143.30 (n/a)	104.88 (n/a)	82.20 (n/a)	80.60 (n/a)	32.24 (n/a)

silu_2_cols_1_channels_2048_tile_1024

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`e199ef2` — 2026-02-06 17:34:56	0.06 (-35.82%)	0.05 (-34.76%)	0.06 (-42.86%)	0.04 (-28.25%)	0.01 (-61.19%)	193.60 (+39.38%)	152.36 (+47.78%)	145.70 (+75.12%)	127.10 (+55.76%)	24.95 (-11.86%)
`1531781` — 2026-01-23 18:49:24	0.10 (n/a)	0.08 (n/a)	0.10 (n/a)	0.06 (n/a)	0.02 (n/a)	138.90 (n/a)	103.10 (n/a)	83.20 (n/a)	81.60 (n/a)	28.30 (n/a)

silu_4_cols_1_channels_2048_tile_512

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`e199ef2` — 2026-02-06 17:34:56	0.07 (-45.70%)	0.05 (-51.64%)	0.05 (-59.91%)	0.04 (-25.18%)	0.01 (-67.80%)	183.60 (+33.62%)	163.08 (+95.63%)	174.50 (+149.29%)	125.40 (+84.14%)	23.19 (-23.39%)
`1531781` — 2026-01-23 18:49:24	0.12 (n/a)	0.11 (n/a)	0.12 (n/a)	0.06 (n/a)	0.03 (n/a)	137.40 (n/a)	83.36 (n/a)	70.00 (n/a)	68.10 (n/a)	30.27 (n/a)

silu_8_cols_1_channels_2048_tile_256

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`e199ef2` — 2026-02-06 17:34:56	0.07 (-39.76%)	0.05 (-47.57%)	0.05 (-49.68%)	0.04 (-49.06%)	0.01 (-18.58%)	215.50 (+96.27%)	174.10 (+95.18%)	174.80 (+98.86%)	118.00 (+65.96%)	37.92 (+161.44%)
`1531781` — 2026-01-23 18:49:24	0.12 (n/a)	0.09 (n/a)	0.09 (n/a)	0.07 (n/a)	0.02 (n/a)	109.80 (n/a)	89.20 (n/a)	87.90 (n/a)	71.10 (n/a)	14.50 (n/a)

softmax_1_cols_2_channels_4096_tile_2048

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`1531781` — 2026-01-23 18:49:24	0.21 (+20.19%)	0.16 (+13.84%)	0.16 (+8.19%)	0.12 (+26.61%)	0.04 (+17.18%)	135.50 (-20.99%)	104.68 (-12.65%)	102.70 (-7.56%)	76.50 (-16.85%)	22.55 (-26.21%)
`0f0da0b` — 2026-01-14 16:31:00	0.18 (n/a)	0.14 (n/a)	0.15 (n/a)	0.10 (n/a)	0.03 (n/a)	171.50 (n/a)	119.84 (n/a)	111.10 (n/a)	92.00 (n/a)	30.56 (n/a)

softmax_2_cols_2_channels_32768_tile_1024

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`e199ef2` — 2026-02-06 17:34:56	1.01 (n/a)	0.90 (n/a)	0.95 (n/a)	0.71 (n/a)	0.12 (n/a)	185.30 (n/a)	147.78 (n/a)	138.00 (n/a)	129.50 (n/a)	22.45 (n/a)

softmax_2_cols_2_channels_32768_tile_512

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`e199ef2` — 2026-02-06 17:34:56	0.83 (n/a)	0.75 (n/a)	0.75 (n/a)	0.63 (n/a)	0.08 (n/a)	209.10 (n/a)	177.08 (n/a)	175.20 (n/a)	158.50 (n/a)	20.00 (n/a)

softmax_2_cols_2_channels_4096_tile_1024

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`1531781` — 2026-01-23 18:49:24	0.22 (+40.26%)	0.18 (+43.33%)	0.20 (+64.08%)	0.13 (+24.34%)	0.04 (+68.98%)	127.80 (-19.62%)	94.74 (-28.95%)	83.40 (-39.08%)	73.90 (-28.67%)	24.46 (-5.31%)
`0f0da0b` — 2026-01-14 16:31:00	0.16 (n/a)	0.13 (n/a)	0.12 (n/a)	0.10 (n/a)	0.03 (n/a)	159.00 (n/a)	133.34 (n/a)	136.90 (n/a)	103.60 (n/a)	25.83 (n/a)

softmax_2_cols_2_channels_4096_tile_512

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`1531781` — 2026-01-23 18:49:24	0.24 (+85.21%)	0.17 (+69.97%)	0.15 (+54.16%)	0.11 (+37.21%)	0.05 (+203.82%)	148.10 (-27.12%)	102.78 (-37.88%)	106.70 (-35.10%)	69.60 (-45.96%)	31.81 (+15.98%)
`0f0da0b` — 2026-01-14 16:31:00	0.13 (n/a)	0.10 (n/a)	0.10 (n/a)	0.08 (n/a)	0.02 (n/a)	203.20 (n/a)	165.46 (n/a)	164.40 (n/a)	128.80 (n/a)	27.43 (n/a)

softmax_4_cols_4_channels_32768_tile_2048

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`e199ef2` — 2026-02-06 17:34:56	0.71 (n/a)	0.58 (n/a)	0.55 (n/a)	0.48 (n/a)	0.09 (n/a)	271.80 (n/a)	230.78 (n/a)	237.90 (n/a)	184.10 (n/a)	32.22 (n/a)

swiglu

No metrics available.

swiglu_decode_1x2048x2048

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`130b6ea` — 2025-12-05 21:15:49	0.00 (-15.46%)	0.00 (-27.20%)	0.00 (-28.17%)	0.00 (-33.90%)	0.00 (+1834.83%)	5187.20 (+51.28%)	4720.84 (+38.38%)	4750.90 (+39.22%)	4015.10 (+18.29%)	390.23 (+3298.45%)
`0a6c11c` — 2025-12-04 00:39:10	0.00 (n/a)	0.00 (n/a)	0.00 (n/a)	0.00 (n/a)	0.00 (n/a)	3428.80 (n/a)	3411.62 (n/a)	3412.60 (n/a)	3394.20 (n/a)	11.48 (n/a)

swiglu_decode_1x2048x2048_0

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`e199ef2` — 2026-02-06 17:34:56	n/a (n/a)	n/a (n/a)	n/a (n/a)	n/a (n/a)	n/a (n/a)	n/a (n/a)	n/a (n/a)	n/a (n/a)	n/a (n/a)	n/a (n/a)
`1531781` — 2026-01-23 18:49:24	0.00 (n/a)	0.00 (n/a)	0.00 (n/a)	0.00 (n/a)	0.00 (n/a)	5111.40 (n/a)	4324.88 (n/a)	4736.10 (n/a)	3410.70 (n/a)	843.97 (n/a)

swiglu_prefill_256x2048x2048_0

No metrics available.

tanh_1_cols_1_channels_2048_tile_2048

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`e199ef2` — 2026-02-06 17:34:56	0.06 (-34.50%)	0.06 (-28.83%)	0.06 (-20.29%)	0.05 (-13.38%)	0.00 (-70.60%)	162.60 (+15.48%)	148.62 (+36.17%)	145.40 (+25.45%)	130.70 (+52.69%)	12.89 (-45.17%)
`1531781` — 2026-01-23 18:49:24	0.10 (n/a)	0.08 (n/a)	0.07 (n/a)	0.06 (n/a)	0.02 (n/a)	140.80 (n/a)	109.14 (n/a)	115.90 (n/a)	85.60 (n/a)	23.52 (n/a)

tanh_2_cols_1_channels_2048_tile_1024

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`e199ef2` — 2026-02-06 17:34:56	0.08 (-20.62%)	0.06 (-38.87%)	0.05 (-42.02%)	0.04 (-50.00%)	0.02 (+86.18%)	206.70 (+100.10%)	153.96 (+74.91%)	151.40 (+72.44%)	98.00 (+25.96%)	45.54 (+372.58%)
`1531781` — 2026-01-23 18:49:24	0.11 (n/a)	0.09 (n/a)	0.09 (n/a)	0.08 (n/a)	0.01 (n/a)	103.30 (n/a)	88.02 (n/a)	87.80 (n/a)	77.80 (n/a)	9.64 (n/a)

tanh_4_cols_1_channels_2048_tile_512

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`e199ef2` — 2026-02-06 17:34:56	0.10 (-0.85%)	0.06 (-32.26%)	0.05 (-43.08%)	0.04 (-34.82%)	0.02 (+63.17%)	186.20 (+53.38%)	144.96 (+55.20%)	151.60 (+75.67%)	85.40 (+0.95%)	36.69 (+132.16%)
`1531781` — 2026-01-23 18:49:24	0.10 (n/a)	0.09 (n/a)	0.09 (n/a)	0.07 (n/a)	0.01 (n/a)	121.40 (n/a)	93.40 (n/a)	86.30 (n/a)	84.60 (n/a)	15.80 (n/a)

tanh_8_cols_1_channels_2048_tile_256

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`e199ef2` — 2026-02-06 17:34:56	0.06 (-47.17%)	0.05 (-47.42%)	0.04 (-56.50%)	0.04 (-17.93%)	0.01 (-72.14%)	199.60 (+21.86%)	179.84 (+74.77%)	184.80 (+129.85%)	135.70 (+89.26%)	25.67 (-35.81%)
`1531781` — 2026-01-23 18:49:24	0.11 (n/a)	0.09 (n/a)	0.10 (n/a)	0.05 (n/a)	0.03 (n/a)	163.80 (n/a)	102.90 (n/a)	80.40 (n/a)	71.70 (n/a)	39.99 (n/a)

transpose_2048_M_64_N_1_cols_1_channels_64_m_64_n_8_s

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`130b6ea` — 2025-12-05 21:15:49	3.93 (-7.02%)	3.68 (+2.48%)	3.66 (+1.82%)	3.45 (+15.62%)	0.15 (-63.64%)	151.80 (-13.50%)	142.70 (-3.59%)	143.40 (-1.78%)	133.50 (+7.49%)	5.90 (-66.22%)
`0a6c11c` — 2025-12-04 00:39:10	4.22 (n/a)	3.59 (n/a)	3.59 (n/a)	2.99 (n/a)	0.42 (n/a)	175.50 (n/a)	148.02 (n/a)	146.00 (n/a)	124.20 (n/a)	17.46 (n/a)

transpose_2048_M_64_N_1_cols_1_channels_64_m_64_n_8_s0

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`e199ef2` — 2026-02-06 17:34:56	3.11 (-37.04%)	2.81 (-35.38%)	2.76 (-36.91%)	2.57 (-31.32%)	0.20 (-59.26%)	204.30 (+45.62%)	187.54 (+53.75%)	190.00 (+58.47%)	168.30 (+58.77%)	13.23 (-6.51%)
`1531781` — 2026-01-23 18:49:24	4.95 (n/a)	4.34 (n/a)	4.37 (n/a)	3.74 (n/a)	0.50 (n/a)	140.30 (n/a)	121.98 (n/a)	119.90 (n/a)	106.00 (n/a)	14.15 (n/a)

transpose_2048_M_64_N_1_cols_2_channels_64_m_64_n_8_s

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`130b6ea` — 2025-12-05 21:15:49	5.11 (+15.57%)	3.85 (-3.32%)	3.40 (-18.64%)	3.29 (+7.09%)	0.71 (+49.14%)	159.60 (-6.61%)	140.18 (+4.80%)	154.20 (+22.97%)	102.60 (-13.49%)	22.70 (+19.77%)
`0a6c11c` — 2025-12-04 00:39:10	4.42 (n/a)	3.99 (n/a)	4.18 (n/a)	3.07 (n/a)	0.48 (n/a)	170.90 (n/a)	133.76 (n/a)	125.40 (n/a)	118.60 (n/a)	18.95 (n/a)

transpose_2048_M_64_N_1_cols_2_channels_64_m_64_n_8_s0

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`e199ef2` — 2026-02-06 17:34:56	4.35 (-28.45%)	2.95 (-40.85%)	2.75 (-46.04%)	2.12 (-48.11%)	0.83 (+11.38%)	246.80 (+92.66%)	187.26 (+75.24%)	191.00 (+85.44%)	120.60 (+39.75%)	44.77 (+185.71%)
`1531781` — 2026-01-23 18:49:24	6.07 (n/a)	4.99 (n/a)	5.09 (n/a)	4.09 (n/a)	0.74 (n/a)	128.10 (n/a)	106.86 (n/a)	103.00 (n/a)	86.30 (n/a)	15.67 (n/a)

weighted_rms_norm_1_cols_2_channels_2048_weights_2048

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`e199ef2` — 2026-02-06 17:34:56	n/a (n/a)	n/a (n/a)	n/a (n/a)	n/a (n/a)	n/a (n/a)	n/a (n/a)	n/a (n/a)	n/a (n/a)	n/a (n/a)	n/a (n/a)
`1531781` — 2026-01-23 18:49:24	0.11 (n/a)	0.09 (n/a)	0.09 (n/a)	0.05 (n/a)	0.02 (n/a)	155.00 (n/a)	99.92 (n/a)	89.60 (n/a)	76.50 (n/a)	31.36 (n/a)

weighted_rms_norm_2_cols_2_channels_2048_weights_1024

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`e199ef2` — 2026-02-06 17:34:56	n/a (n/a)	n/a (n/a)	n/a (n/a)	n/a (n/a)	n/a (n/a)	n/a (n/a)	n/a (n/a)	n/a (n/a)	n/a (n/a)	n/a (n/a)
`1531781` — 2026-01-23 18:49:24	0.10 (n/a)	0.09 (n/a)	0.09 (n/a)	0.05 (n/a)	0.02 (n/a)	166.70 (n/a)	102.72 (n/a)	88.50 (n/a)	83.10 (n/a)	35.87 (n/a)

weighted_rms_norm_4_cols_2_channels_2048_weights_512

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`e199ef2` — 2026-02-06 17:34:56	n/a (n/a)	n/a (n/a)	n/a (n/a)	n/a (n/a)	n/a (n/a)	n/a (n/a)	n/a (n/a)	n/a (n/a)	n/a (n/a)	n/a (n/a)
`1531781` — 2026-01-23 18:49:24	0.09 (n/a)	0.09 (n/a)	0.09 (n/a)	0.07 (n/a)	0.01 (n/a)	119.00 (n/a)	96.48 (n/a)	89.90 (n/a)	89.30 (n/a)	12.81 (n/a)

weighted_rms_norm_8_cols_2_channels_2048_weights_256

Commit/Date	Bandwidth (max)	Bandwidth (mean)	Bandwidth (median)	Bandwidth (min)	Bandwidth (stddev)	Latency (max)	Latency (mean)	Latency (median)	Latency (min)	Latency (stddev)
`e199ef2` — 2026-02-06 17:34:56	n/a (n/a)	n/a (n/a)	n/a (n/a)	n/a (n/a)	n/a (n/a)	n/a (n/a)	n/a (n/a)	n/a (n/a)	n/a (n/a)	n/a (n/a)
`1531781` — 2026-01-23 18:49:24	0.11 (n/a)	0.10 (n/a)	0.10 (n/a)	0.06 (n/a)	0.02 (n/a)	135.80 (n/a)	89.92 (n/a)	80.00 (n/a)	74.30 (n/a)	25.76 (n/a)

github-actions · 2026-02-06T17:47:47Z

📊 Test Results for Test Example Applications

e199ef2 (2026_02_06_17_46_55)

IRONCLAD

Tested on 2026_02_06_17_46_55 at commit e199ef2.

Test	Checks	TTFT (mean)	TPS (mean)
llama_3.2_1b_prompt_13_tokens_1	❌ 0/1	n/a	n/a
llama_3.2_1b_prompt_13_tokens_40	❌ 0/1	n/a	n/a
llama_3.2_1b_prompt_2048_tokens_1	❌ 0/1	n/a	n/a
llama_3.2_1b_prompt_2048_tokens_40	❌ 0/1	n/a	n/a

📈 Trends (vs main branch) for Test Example Applications

e199ef2 (2026_02_06_17_46_55)

IRONCLAD Trends

llama_3.2_1b

Commit/Date	Num Tokens (max)	Num Tokens (mean)	Num Tokens (median)	Num Tokens (min)	Num Tokens (stddev)	TPS (max)	TPS (mean)	TPS (median)	TPS (min)	TPS (stddev)	TTFT (max)	TTFT (mean)	TTFT (median)	TTFT (min)	TTFT (stddev)	Total (max)	Total (mean)	Total (median)	Total (min)	Total (stddev)
`130b6ea` — 2025-12-05 21:33:12	40.00 (+0.00%)	40.00 (+0.00%)	40.00 (+0.00%)	40.00 (+0.00%)	0.00 (n/a)	4.71 (-0.42%)	4.64 (-0.09%)	4.64 (+0.65%)	4.55 (-0.22%)	0.05 (-17.66%)	4.41 (-0.34%)	4.39 (-0.19%)	4.38 (-0.33%)	4.37 (-0.15%)	0.01 (-25.90%)	12.96 (-0.00%)	12.80 (+0.07%)	12.80 (-0.23%)	12.67 (+0.44%)	0.09 (-21.12%)
`0a6c11c` — 2025-12-03 23:35:15	40.00 (n/a)	40.00 (n/a)	40.00 (n/a)	40.00 (n/a)	0.00 (n/a)	4.73 (n/a)	4.64 (n/a)	4.61 (n/a)	4.56 (n/a)	0.06 (n/a)	4.42 (n/a)	4.40 (n/a)	4.40 (n/a)	4.37 (n/a)	0.02 (n/a)	12.96 (n/a)	12.79 (n/a)	12.83 (n/a)	12.62 (n/a)	0.12 (n/a)

llama_3.2_1b_prompt_13_tokens_1

Commit/Date	Num_Tokens (max)	Num_Tokens (mean)	Num_Tokens (median)	Num_Tokens (min)	Num_Tokens (stddev)	TPS (max)	TPS (mean)	TPS (median)	TPS (min)	TPS (stddev)	TTFT (max)	TTFT (mean)	TTFT (median)	TTFT (min)	TTFT (stddev)
`e199ef2` — 2026-02-06 17:46:50	n/a (n/a)	n/a (n/a)	n/a (n/a)	n/a (n/a)	n/a (n/a)	n/a (n/a)	n/a (n/a)	n/a (n/a)	n/a (n/a)	n/a (n/a)	n/a (n/a)	n/a (n/a)	n/a (n/a)	n/a (n/a)	n/a (n/a)
`1531781` — 2026-01-23 18:40:34	1.00 (n/a)	1.00 (n/a)	1.00 (n/a)	1.00 (n/a)	0.00 (n/a)	0.00 (n/a)	0.00 (n/a)	0.00 (n/a)	0.00 (n/a)	0.00 (n/a)	0.61 (n/a)	0.61 (n/a)	0.61 (n/a)	0.61 (n/a)	0.00 (n/a)

llama_3.2_1b_prompt_13_tokens_40

Commit/Date	Num_Tokens (max)	Num_Tokens (mean)	Num_Tokens (median)	Num_Tokens (min)	Num_Tokens (stddev)	TPS (max)	TPS (mean)	TPS (median)	TPS (min)	TPS (stddev)	TTFT (max)	TTFT (mean)	TTFT (median)	TTFT (min)	TTFT (stddev)
`e199ef2` — 2026-02-06 17:46:50	n/a (n/a)	n/a (n/a)	n/a (n/a)	n/a (n/a)	n/a (n/a)	n/a (n/a)	n/a (n/a)	n/a (n/a)	n/a (n/a)	n/a (n/a)	n/a (n/a)	n/a (n/a)	n/a (n/a)	n/a (n/a)	n/a (n/a)
`1531781` — 2026-01-23 18:40:34	40.00 (n/a)	40.00 (n/a)	40.00 (n/a)	40.00 (n/a)	0.00 (n/a)	4.34 (n/a)	4.34 (n/a)	4.34 (n/a)	4.34 (n/a)	0.00 (n/a)	0.61 (n/a)	0.61 (n/a)	0.61 (n/a)	0.61 (n/a)	0.00 (n/a)

llama_3.2_1b_prompt_2048_tokens_1

Commit/Date	Num_Tokens (max)	Num_Tokens (mean)	Num_Tokens (median)	Num_Tokens (min)	Num_Tokens (stddev)	TPS (max)	TPS (mean)	TPS (median)	TPS (min)	TPS (stddev)	TTFT (max)	TTFT (mean)	TTFT (median)	TTFT (min)	TTFT (stddev)
`e199ef2` — 2026-02-06 17:46:50	n/a (n/a)	n/a (n/a)	n/a (n/a)	n/a (n/a)	n/a (n/a)	n/a (n/a)	n/a (n/a)	n/a (n/a)	n/a (n/a)	n/a (n/a)	n/a (n/a)	n/a (n/a)	n/a (n/a)	n/a (n/a)	n/a (n/a)
`1531781` — 2026-01-23 18:40:34	1.00 (n/a)	1.00 (n/a)	1.00 (n/a)	1.00 (n/a)	0.00 (n/a)	0.00 (n/a)	0.00 (n/a)	0.00 (n/a)	0.00 (n/a)	0.00 (n/a)	2.70 (n/a)	2.70 (n/a)	2.70 (n/a)	2.70 (n/a)	0.00 (n/a)

llama_3.2_1b_prompt_2048_tokens_40

Commit/Date	Num_Tokens (max)	Num_Tokens (mean)	Num_Tokens (median)	Num_Tokens (min)	Num_Tokens (stddev)	TPS (max)	TPS (mean)	TPS (median)	TPS (min)	TPS (stddev)	TTFT (max)	TTFT (mean)	TTFT (median)	TTFT (min)	TTFT (stddev)
`e199ef2` — 2026-02-06 17:46:50	n/a (n/a)	n/a (n/a)	n/a (n/a)	n/a (n/a)	n/a (n/a)	n/a (n/a)	n/a (n/a)	n/a (n/a)	n/a (n/a)	n/a (n/a)	n/a (n/a)	n/a (n/a)	n/a (n/a)	n/a (n/a)	n/a (n/a)
`1531781` — 2026-01-23 18:40:34	40.00 (n/a)	40.00 (n/a)	40.00 (n/a)	40.00 (n/a)	0.00 (n/a)	4.22 (n/a)	4.22 (n/a)	4.22 (n/a)	4.22 (n/a)	0.00 (n/a)	2.70 (n/a)	2.70 (n/a)	2.70 (n/a)	2.70 (n/a)	0.00 (n/a)

andrej added 30 commits January 14, 2026 14:46

rework profiling

cf6485d

vibe-coded flame graph visualization

af46d81

plot updates

1f184eb

simplified implementation (no KV cache yet)

eedc527

add KV cache

5372fca

start with simplified llama for NPU

f0d3289

offload last layer GEMM/GEMV

fd6f759

add profile path analyzer

ec0d372

change profiling to allow annotation using contexts; remove nn.Module

280917a

refactoring started; RMSNorm offloaded, last layer GEMM started

7b56ef8

last layer GEMM offloaded

1daaa39

fixes:

83380c2

cleanup

3aa1e02

RMSNorm offloaded verywhere, cleanup

002d487

less buffer copying

db9c4b8

offload first residual

b0942e8

simplify

02e5307

offload second residual

e815c50

SwiGLU offloaded

ebe8f32

offload RoPE

7515199

offload attention query projection linear layer

7ef6b03

offload attention key projection linear layer -- slight decrease in o…

d43eeea

…utput quality

add batching to GEMV, fix issue when K<vector_size, offload attention…

de2995c

… score calculation in llama

add strided_copy operator

74300bd

add patchable callable

b1eab7c

simplify llama_npu.py, make GEMV operator input shapes simpler

119bb7e

fix strided copy; offload KV cache concat + transpose to NPU

7fee60d

offload repeat_interleave

b72432b

offload normalization/scaling + softmax (with -inf masking on CPU for…

8b4b2b3

… now)

make softmax run-time parametrizable

5575d4f

andrej and others added 19 commits January 23, 2026 16:59

transpose fused onto the end; 2.6 TPS

834f33b

fuse attention context gemv - 2.7 TPS

f345e8d

fuse attn output onto end - 2.7 TPS

b46838f

fuse GQA + post attention - 2.5 TPS

99ef9fa

fuse rms norm onto beginning of transformer block -- full transformer…

8eaa9bc

… block fused -- 2.5 TPS

[WRONG RESULTS] 16x-fused transformer block

165a93b

remove unnecessary syncs, remove unused ops -- 4.4 TPS

86d7de8

[decode end-to-end fused] offload last rms norm and last linear layer…

77bac5a

… -- 4.4 TPS

cleanup

6211124

remove old llama implementation

2da438d

naive attempt at porting operations

38e1b4c

most of the operators running in local tests

ba15525

the great reformatting

d483d4f

a few more formatting fixes

9f29a92

some refactoring

404211e

some work with swiglu

12b1dc6

add licenses

1aa050e

Merge branch 'devel' into simplifying-refactor: Reconcile operator ab…

b5898f0

…straction with new iron/ structure

Move remaining operators to iron/operators/

a8e07fc

try to minimize changes that are not central to the refactor

68a2e45

hunhoffe added 2 commits February 6, 2026 10:22

remove extra test file

b9cbcba

fixup imports

b8815ab

fix another import path

93e04a5

Conversation

hunhoffe commented Feb 5, 2026

Added

Changed

Removed

PR Merge Checklist

Uh oh!

github-actions bot commented Feb 6, 2026

IRONCLAD

IRONCLAD Trends

axpy_1_cols_2_channels_2048_tile_2048_3.0

axpy_1_cols_2_channels_2048_tile_2048_3.0_0

axpy_2_cols_2_channels_2048_tile_1024_3.0

axpy_2_cols_2_channels_2048_tile_1024_3.0_0

axpy_4_cols_2_channels_2048_tile_512_3.0

axpy_4_cols_2_channels_2048_tile_512_3.0_0

axpy_8_cols_2_channels_2048_tile_256_3.0

axpy_8_cols_2_channels_2048_tile_256_3.0_0

dequant_1_cols_1_channels_2048_tile_2048

dequant_1_cols_1_channels_2048_tile_2048_0

dequant_1_cols_2_channels_2048_tile_1024

dequant_1_cols_2_channels_2048_tile_1024_0

dequant_2_cols_1_channels_2048_tile_1024

dequant_2_cols_1_channels_2048_tile_1024_0

dequant_2_cols_2_channels_2048_tile_512

dequant_2_cols_2_channels_2048_tile_512_0

dequant_4_cols_1_channels_2048_tile_512

dequant_4_cols_1_channels_2048_tile_512_0

dequant_4_cols_2_channels_2048_tile_256

dequant_4_cols_2_channels_2048_tile_256_0

dequant_8_cols_1_channels_2048_tile_256

dequant_8_cols_1_channels_2048_tile_256_0

dequant_8_cols_2_channels_2048_tile_128

dequant_8_cols_2_channels_2048_tile_128_0

eltwise_add_1_cols_2_channels_2048_tile_2048

eltwise_add_2_cols_2_channels_2048_tile_1024

eltwise_add_4_cols_2_channels_2048_tile_512

eltwise_add_8_cols_2_channels_2048_tile_256

eltwise_mul_1_cols_2_channels_2048_tile_2048

eltwise_mul_2_cols_2_channels_2048_tile_1024

eltwise_mul_4_cols_2_channels_2048_tile_512

eltwise_mul_8_cols_2_channels_2048_tile_256

gelu_1_cols_1_channels_2048_tile_2048

gelu_1_cols_2_channels_2048_tile_1024

gelu_2_cols_1_channels_2048_tile_1024

gelu_2_cols_2_channels_2048_tile_512

gelu_4_cols_1_channels_2048_tile_512

gelu_4_cols_2_channels_2048_tile_256

gelu_8_cols_1_channels_2048_tile_256

gelu_8_cols_2_channels_2048_tile_128

gemm_1792x896x1152_64x32x48_8cols_ccolmaj

gemm_192x384x64_48x96x16_4cols

gemm_192x384x64_48x96x16_4cols_bcolmaj_ccolmaj

gemm_2048x2048x2048_64x64x32_8_cols_0_bcolmaj_0_ccolmaj_0

gemm_2048x2048x2048_64x64x32_8_cols_0_bcolmaj_1_ccolmaj_0

gemm_2048x2048x2048_64x64x32_8_cols_1_bcolmaj_0_ccolmaj_0

gemm_2048x2048x2048_64x64x64_1cols

gemm_2048x2048x2048_64x64x64_2_cols_0_bcolmaj_0_ccolmaj_0

gemm_2048x2048x2048_64x64x64_2_cols_0_bcolmaj_0_ccolmaj_0_0

gemm_2048x2048x2048_64x64x64_2_cols_0_bcolmaj_1_ccolmaj_0

gemm_2048x2048x2048_64x64x64_2_cols_0_bcolmaj_1_ccolmaj_0_0

gemm_2048x2048x2048_64x64x64_2_cols_1_bcolmaj_0_ccolmaj_0

gemm_2048x2048x2048_64x64x64_2_cols_1_bcolmaj_0_ccolmaj_0_0

gemm_2048x2048x2048_64x64x64_2cols_bcolmaj

gemm_2048x2048x2048_64x64x64_8_cols_0_bcolmaj_0_ccolmaj_0

gemm_2048x2048x2048_64x64x64_8_cols_0_bcolmaj_0_ccolmaj_0_0

gemm_2048x2048x2048_64x64x64_8_cols_0_bcolmaj_1_ccolmaj_0_0

gemm_2048x2048x2048_64x64x64_8_cols_1_bcolmaj_0_ccolmaj_0_0

gemm_2048x2048x2048_64x64x64_8cols_bcolmaj_ccolmaj

gemm_384x1536x1792_32x48x64_4cols_bcolmaj

gemm_896x1792x640_32x64x80_8cols_ccolmaj

layer_norm_1_cols_1_channels_2048_tile_2048

layer_norm_1_cols_2_channels_2048_tile_1024

layer_norm_2_cols_1_channels_2048_tile_1024

layer_norm_2_cols_2_channels_2048_tile_512

layer_norm_4_cols_1_channels_2048_tile_512

layer_norm_4_cols_2_channels_2048_tile_256

layer_norm_8_cols_1_channels_2048_tile_256

layer_norm_8_cols_2_channels_2048_tile_128

matrix_vector_mul_128x128_32_1col