RyzenAI 1.4.0 quick test reports no NPU on 7840U

[zeerd]

1.3.1 has NPU reported, but 1.40 has not.

Driver was undated to 32.0.203.257 3/12/2025.

(ryzen-ai-1.4.0) PS C:\Program Files\RyzenAI\1.4.0\quicktest> python.exe .\quicktest.py
Setting environment for PHX/HPT
XLNX_VART_FIRMWARE= C:\Program Files\RyzenAI\1.4.0\voe-4.0-win_amd64\xclbins\phoenix\1x4.xclbin
NUM_OF_DPU_RUNNERS= 1
XLNX_TARGET_NAME= AMD_AIE2_Nx4_Overlay
XLNX_ENABLE_CACHE= 0
WARNING: Logging before InitGoogleLogging() is written to STDERR
I20250330 16:01:12.021073  6200 vitisai_compile_model.cpp:1144] Vitis AI EP Load ONNX Model Success
I20250330 16:01:12.021073  6200 vitisai_compile_model.cpp:1145] Graph Input Node Name/Shape (1)
I20250330 16:01:12.021073  6200 vitisai_compile_model.cpp:1149]          input : [-1x3x32x32]
I20250330 16:01:12.022074  6200 vitisai_compile_model.cpp:1155] Graph Output Node Name/Shape (1)
I20250330 16:01:12.022074  6200 vitisai_compile_model.cpp:1159]          output : [-1x10]
F20250330 16:01:31.203575  6200 target_factory.cpp:155] Cannot find or create target with fingerprint=0x08000205000e16c6
[Vitis AI EP] No. of Operators :   CPU   400
Test Passed
(ryzen-ai-1.4.0) PS C:\Program Files\RyzenAI\1.4.0\quicktest> conda activate ryzen-ai-1.3.1
(ryzen-ai-1.3.1) PS C:\Program Files\RyzenAI\1.4.0\quicktest> python.exe .\quicktest.py
Setting environment for PHX/HPT
XLNX_VART_FIRMWARE= C:\Program Files\RyzenAI\1.4.0\voe-4.0-win_amd64\xclbins\phoenix\1x4.xclbin
NUM_OF_DPU_RUNNERS= 1
XLNX_TARGET_NAME= AMD_AIE2_Nx4_Overlay
XLNX_ENABLE_CACHE= 0
WARNING: Logging before InitGoogleLogging() is written to STDERR
I20250330 16:02:05.704517 18980 vitisai_compile_model.cpp:1046] Vitis AI EP Load ONNX Model Success
I20250330 16:02:05.704517 18980 vitisai_compile_model.cpp:1047] Graph Input Node Name/Shape (1)
I20250330 16:02:05.704517 18980 vitisai_compile_model.cpp:1051]          input : [-1x3x32x32]
I20250330 16:02:05.704517 18980 vitisai_compile_model.cpp:1057] Graph Output Node Name/Shape (1)
I20250330 16:02:05.704517 18980 vitisai_compile_model.cpp:1061]          output : [-1x10]
[Vitis AI EP] No. of Operators :   CPU     2    NPU   398
[Vitis AI EP] No. of Subgraphs :   NPU     1 Actually running on NPU     1
2025-03-30 16:02:52.6865601 [W:onnxruntime:, session_state.cc:1166 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.
2025-03-30 16:02:52.7080564 [W:onnxruntime:, session_state.cc:1168 onnxruntime::VerifyEachNodeIsAssignedToAnEp] Rerunning with verbose output on a non-minimal build will show node assignments.
Test Passed
(ryzen-ai-1.3.1) PS C:\Program Files\RyzenAI\1.4.0\quicktest>

[uday610]

There should be a cache directory from your previous run, modelcachekey_quicktest . Can you remove this and run in 1.4 ?

[zeerd]

Thanks, but where should I find the modelcachekey_quicktest?
I am not good at the commands of power-shell for Windows to find something...

[UltraRabbit]

@zeerd The cache path would be C:\temp\%USERNAME%\vaip\.cache if you didn't modify the default temporary folder location of windows. You may also edit the quicktest.py file and change the modelcachekey_quicktest into another value.