Running quicktest.py fails with [XRT] ERROR: CreateFile failed with error 3: no such process

[sashakeel]

Running quick test after installing the Ryzen AI Software 1.6 results into XRT error. Any ideas?

C:\Program Files\RyzenAI\1.6.0\quicktest>python quicktest.py
Setting environment for PHX/HPT
WARNING: Logging before InitGoogleLogging() is written to STDERR
I20251016 16:51:29.392673 14268 register_ssmlp.cpp:124] Registering Custom Operator: com.amd:SSMLP
I20251016 16:51:29.392673 14268 register_matmulnbits.cpp:110] Registering Custom Operator: com.amd:MatMulNBits
I20251016 16:51:29.439898 14268 vitisai_compile_model.cpp:1266] Vitis AI EP Load ONNX Model Success
I20251016 16:51:29.440778 14268 vitisai_compile_model.cpp:1267] Graph Input Node Name/Shape (1)
I20251016 16:51:29.440778 14268 vitisai_compile_model.cpp:1271]          input : [-1x3x32x32]
I20251016 16:51:29.440778 14268 vitisai_compile_model.cpp:1277] Graph Output Node Name/Shape (1)
I20251016 16:51:29.440778 14268 vitisai_compile_model.cpp:1281]          output : [-1x10]
Using TXN FORMAT 0.1
[Vitis AI EP] No. of Operators :   NPU   398 VITIS_EP_CPU     2
[Vitis AI EP] No. of Subgraphs :   NPU     1 Actually running on NPU      1
XRT build version: 2.14.0
Build hash: 1cbc60cf4360cebfbbf169f8ea6228bb98aaed4c
Build date: 2023-10-06 10:29:32
Git branch: HEAD
PID: 11824
UID: 0
[Thu Oct 16 14:51:40 2025 GMT]
HOST:
EXE:
[XRT] ERROR: CreateFile failed with error 3: no such process

[dwithchenna]

@sashakeel This could be an error related to permissions, can you confirm if you are able to run quicktest.py in administrator mode.

If you can share more details about the systems and software version, it will help us reproduce the issue. Thanks!

[sashakeel]

@dwithchenna running in administrator mode still gives the same error.

System Information:

Operating System: Microsoft Windows 11 IoT Enterprise LTSC
Processor: AMD Ryzen 5 8640U w/ Radeon 760M Graphics, 3501 Mhz, 6 Core(s), 12 Logical Processor(s)
Software: Ryzen AI Software 1.6

[sashakeel]

@dwithchenna any update?

[amdjiahangpan]

I also encountered this problem in 760PRO, but not in HX370/HX395. I tried to give administrator privileges, but still got an error. So I had to roll back the 760PRO environment to 1.5.0.

[xXxTheUserxXx]

AMD Ryzen™ 5 PRO 8640U has the same problem. One major advantage of environment 1.6.0 is that it uses Python 3.12. Please ensure it is running for your Pro Series Technologies.

[sashakeel]

Running quick test after installing the Ryzen AI Software 1.6 results into XRT error. Any ideas?

C:\Program Files\RyzenAI\1.6.0\quicktest>python quicktest.py
Setting environment for PHX/HPT
WARNING: Logging before InitGoogleLogging() is written to STDERR
I20251016 16:51:29.392673 14268 register_ssmlp.cpp:124] Registering Custom Operator: com.amd:SSMLP
I20251016 16:51:29.392673 14268 register_matmulnbits.cpp:110] Registering Custom Operator: com.amd:MatMulNBits
I20251016 16:51:29.439898 14268 vitisai_compile_model.cpp:1266] Vitis AI EP Load ONNX Model Success
I20251016 16:51:29.440778 14268 vitisai_compile_model.cpp:1267] Graph Input Node Name/Shape (1)
I20251016 16:51:29.440778 14268 vitisai_compile_model.cpp:1271]          input : [-1x3x32x32]
I20251016 16:51:29.440778 14268 vitisai_compile_model.cpp:1277] Graph Output Node Name/Shape (1)
I20251016 16:51:29.440778 14268 vitisai_compile_model.cpp:1281]          output : [-1x10]
Using TXN FORMAT 0.1
[Vitis AI EP] No. of Operators :   NPU   398 VITIS_EP_CPU     2
[Vitis AI EP] No. of Subgraphs :   NPU     1 Actually running on NPU      1
XRT build version: 2.14.0
Build hash: 1cbc60cf4360cebfbbf169f8ea6228bb98aaed4c
Build date: 2023-10-06 10:29:32
Git branch: HEAD
PID: 11824
UID: 0
[Thu Oct 16 14:51:40 2025 GMT]
HOST:
EXE:
[XRT] ERROR: CreateFile failed with error 3: no such process

When attempting to run inference with my custom INT8 ONNX model, I encounter error that is identical to the one produced by the official quicktest.py script.

I20251112 09:48:18.988478  9836 vitisai_compile_model.cpp:1266] Vitis AI EP Load ONNX Model Success
I20251112 09:48:18.988478  9836 vitisai_compile_model.cpp:1267] Graph Input Node Name/Shape (1)
I20251112 09:48:18.988478  9836 vitisai_compile_model.cpp:1271]          images : [1x3x640x640]
I20251112 09:48:18.988478  9836 vitisai_compile_model.cpp:1277] Graph Output Node Name/Shape (1)
I20251112 09:48:18.988478  9836 vitisai_compile_model.cpp:1281]          output0 : [1x5x8400]
Using TXN FORMAT 0.1
[Vitis AI EP] No. of Operators :   CPU   172    NPU   929 VITIS_EP_CPU   158
[Vitis AI EP] No. of Subgraphs :   NPU    75 Actually running on NPU     75
2025-11-12 09:49:05.0439408 [W:onnxruntime:, session_state.cc:1307 onnxruntime::VerifyEachNodeIsAssignedToAnEp] Some nodes were not assigned to the preferred execution providers which may or may not have an negative impact on performance. e.g. ORT explicitly assigns shape related ops to CPU to improve perf.
2025-11-12 09:49:05.0499446 [W:onnxruntime:, session_state.cc:1309 onnxruntime::VerifyEachNodeIsAssignedToAnEp] Rerunning with verbose output on a non-minimal build will show node assignments.
XRT build version: 2.14.0
Build hash: 1cbc60cf4360cebfbbf169f8ea6228bb98aaed4c
Build date: 2023-10-06 10:29:32
Git branch: HEAD
PID: 9112
UID: 0
[Wed Nov 12 08:49:05 2025 GMT]
HOST:
EXE:
[XRT] ERROR: CreateFile failed with error 3: no such process
2

Any update on this would be highly appreciated.

[sashakeel]

I originally reported this issue while using Ryzen AI 1.6, but it appears to still be present. Last week, Ryzen AI 1.7 was released, so I upgraded hoping the problem would be fixed — however, the issue still persists.

Below is the updated error output from Ryzen AI 1.7:

(env-base-ryzen-ai-1-7) C:\Program Files\RyzenAI\1.7.0\quicktest>python quicktest.py
Setting environment for PHX/HPT
←[0;93m2026-01-29 16:50:23.3877562 [W:onnxruntime:, RedundantOpReductionPass.cpp:846 RedundantOpReductionPass.cpp] xir::Op{name = (/avgpool/GlobalAveragePool_output_0_Mul_vaip_163), type = qlinear-pool}'s input and output is unchanged, so it will be removed.←[m
←[0;93m2026-01-29 16:50:23.6880734 [W:onnxruntime:, PartitionPass.cpp:12507 PartitionPass.cpp] xir::Op{name = (output)_replaced_input270, type = dequantize-linear} is partitioned to CPU as :  doesn't supported by target [AMD_AIE2P_4x8_CMC_Overlay].←[m
INFO: [aiecompiler 77-749] Reading logical device aie2p_8x4_device
Using TXN FORMAT 0.1
←[0;93m2026-01-29 16:50:25.5948335 [W:onnxruntime:, pass_main.cpp:250 pass_main.cpp] skip mmap in sg: 1. runner libgraph-engine.so not support mmap ←[m
XRT build version: 2.14.0
Build hash: 1cbc60cf4360cebfbbf169f8ea6228bb98aaed4c
Build date: 2023-10-06 10:29:32
Git branch: HEAD
PID: 13444
UID: 0
[Thu Jan 29 15:50:26 2026 GMT]
HOST:
EXE:
[XRT] ERROR: CreateFile failed with error 3: no such process

System Information:

Operating System: Microsoft Windows 11 IoT Enterprise LTSC
Processor: AMD Ryzen 5 8640U w/ Radeon 760M Graphics, 3501 Mhz, 6 Core(s), 12 Logical Processor(s)
Software: Ryzen AI Software 1.7

Is there already a known workaround or recommended way to overcome this issue?

[ThomasXilinx]

Which version of the NPU driver is installed on your system?

You need to have version 32.0.203.280, as explained in the installation instructions :
https://ryzenai.docs.amd.com/en/latest/inst.html#install-npu-drivers

The XRT build version and build date at the end of your log suggests you have an older version.