Skip to content

[Unity][BYOC] Fuse attention pattern with strided_slice #14649

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Merged
masahi merged 2 commits into from
Apr 18, 2023
Merged

[Unity][BYOC] Fuse attention pattern with strided_slice #14649

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
2 commits
Select commit Hold shift + click to select a range
aed469b
[Unity][BYOC] Fuse attention pattern with `strided_slice`
cyx-6 Apr 17, 2023
da5a27d
remove useless code
cyx-6 Apr 17, 2023
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
12 changes: 10 additions & 2 deletions python/tvm/relax/backend/contrib/cutlass.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -247,11 +247,19 @@ def attention_patterns():
),
(
"cutlass.stacked_attention",
*make_stacked_attention_pattern(),
*make_stacked_attention_pattern(start_op="split"),
),
(
"cutlass.stacked_attention",
*make_stacked_attention_pattern(with_bias=True),
*make_stacked_attention_pattern(start_op="split", with_bias=True),
),
(
"cutlass.stacked_attention",
*make_stacked_attention_pattern(start_op="strided_slice"),
),
(
"cutlass.stacked_attention",
*make_stacked_attention_pattern(start_op="strided_slice", with_bias=True),
),
]

Expand Down
23 changes: 18 additions & 5 deletions python/tvm/relax/backend/patterns.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -197,12 +197,15 @@ def make_attention_pattern(with_bias: bool = False):
return out, annotations


def make_stacked_attention_pattern(with_bias: bool = False):
def make_stacked_attention_pattern(start_op: str, with_bias: bool = False):
"""
Create pattern for fused multi head attention with stacked input.

Parameters
----------
start_op: str
The starting op for pattern, i.e. `R.split` or `R.strided_slice`.

with_bias: bool
Whether or not to include bias addition

Expand All @@ -217,13 +220,23 @@ def make_stacked_attention_pattern(with_bias: bool = False):
check function and codegen.
"""
stacked_qkv = wildcard()
qkv_tuple = is_op("relax.split")(stacked_qkv)
if start_op == "split":
qkv_tuple = is_op("relax.split")(stacked_qkv)
query_raw = is_tuple_get_item(qkv_tuple, 0)
key_raw = is_tuple_get_item(qkv_tuple, 1)
value_raw = is_tuple_get_item(qkv_tuple, 2)
elif start_op == "strided_slice":
query_raw = is_op("relax.strided_slice")(stacked_qkv)
key_raw = is_op("relax.strided_slice")(stacked_qkv)
value_raw = is_op("relax.strided_slice")(stacked_qkv)
Copy link
Member

@vinx13 vinx13 Apr 17, 2023

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

we will also need to check the begin/end in strides_slice if value has a different sequence length

else:
raise NotImplementedError()
query_reshape_list = wildcard()
key_reshape_list = wildcard()
value_reshape_list = wildcard()
query = is_op("relax.reshape")(is_tuple_get_item(qkv_tuple, 0), query_reshape_list)
key = is_op("relax.reshape")(is_tuple_get_item(qkv_tuple, 1), key_reshape_list)
value = is_op("relax.reshape")(is_tuple_get_item(qkv_tuple, 2), value_reshape_list)
query = is_op("relax.reshape")(query_raw, query_reshape_list)
key = is_op("relax.reshape")(key_raw, key_reshape_list)
value = is_op("relax.reshape")(value_raw, value_reshape_list)
annotations = {
"stacked_qkv": stacked_qkv,
"query_reshape_list": query_reshape_list,
Expand Down
44 changes: 36 additions & 8 deletions tests/python/relax/test_codegen_cutlass.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -660,7 +660,7 @@ def get_numpy_stacked_attention_ref(b, s, n, h, h_v, bias_shape, bias_reshape, q
return qkv, bias, ref.transpose(0, 2, 1, 3) # b, s, n, h_v


def get_relax_stacked_attention_module(qkv, b, s, n, h, h_v, bias=None, qk_scale=None):
def get_relax_stacked_attention_module(qkv, b, s, n, h, h_v, op, bias=None, qk_scale=None):
dtype = str(qkv.dtype)

from tvm.script.ir_builder import IRBuilder
Expand All @@ -676,10 +676,22 @@ def get_relax_stacked_attention_module(qkv, b, s, n, h, h_v, bias=None, qk_scale
if bias is not None:
bias = R.arg("bias", R.Tensor(bias.shape, dtype))
with R.dataflow() as frame:
qkv_tuple = R.split(qkv, [n * h, n * h * 2], axis=2)
q = R.reshape(qkv_tuple[0], [b, s, n, h])
k = R.reshape(qkv_tuple[1], [b, s, n, h])
v = R.reshape(qkv_tuple[2], [b, s, n, h_v])
if op == "split":
qkv_tuple = R.split(qkv, [n * h, n * h * 2], axis=2)
q = R.reshape(qkv_tuple[0], [b, s, n, h])
k = R.reshape(qkv_tuple[1], [b, s, n, h])
v = R.reshape(qkv_tuple[2], [b, s, n, h_v])
elif op == "strided_slice":
q = R.reshape(R.strided_slice(qkv, [2], [0], [n * h], [1]), [b, s, n, h])
k = R.reshape(
R.strided_slice(qkv, [2], [n * h], [n * h * 2], [1]), [b, s, n, h]
)
v = R.reshape(
R.strided_slice(qkv, [2], [n * h * 2], [n * h * 2 + n * h_v], [1]),
[b, s, n, h_v],
)
else:
raise NotImplementedError()
result = R.emit(R.nn.attention(q, k, v, bias, qk_scale))
R.output(result)

Expand All @@ -700,15 +712,31 @@ def stacked_attention_size(request):
return request.param


def test_stacked_attention_offload(stacked_attention_size):
def test_stacked_attention_split_offload(stacked_attention_size):
b, s, n, (h, h_v), bias_shape, bias_reshape, scale = stacked_attention_size
qkv, bias, ref = get_numpy_stacked_attention_ref(
b, s, n, h, h_v, bias_shape, bias_reshape, scale, "float32"
)
if scale == "none":
mod = get_relax_stacked_attention_module(qkv, b, s, n, h, h_v, "split", bias)
else:
mod = get_relax_stacked_attention_module(qkv, b, s, n, h, h_v, "split", bias, scale)
if bias is None:
out = get_result_with_relax_cutlass_offload(mod, qkv)
else:
out = get_result_with_relax_cutlass_offload(mod, qkv, bias)
tvm.testing.assert_allclose(out, ref, rtol=1e-2, atol=1e-2)


def test_stacked_attention_strided_slice_offload(stacked_attention_size):
b, s, n, (h, h_v), bias_shape, bias_reshape, scale = stacked_attention_size
qkv, bias, ref = get_numpy_stacked_attention_ref(
b, s, n, h, h_v, bias_shape, bias_reshape, scale, "float32"
)
if scale == "none":
mod = get_relax_stacked_attention_module(qkv, b, s, n, h, h_v, bias)
mod = get_relax_stacked_attention_module(qkv, b, s, n, h, h_v, "strided_slice", bias)
else:
mod = get_relax_stacked_attention_module(qkv, b, s, n, h, h_v, bias, scale)
mod = get_relax_stacked_attention_module(qkv, b, s, n, h, h_v, "strided_slice", bias, scale)
if bias is None:
out = get_result_with_relax_cutlass_offload(mod, qkv)
else:
Expand Down