search

NVIDIA / cuda-python

CUDA Python Low-level Bindings
https://nvidia.github.io/cuda-python/
Other
811 stars 63 forks source link

Sample code execution failure: globalToShmemAsyncCopy.py #6

Closed jordanhubbard closed 2 years ago

jordanhubbard commented 2 years ago

jkh@megamind-> env CUDA_HOME=~/Src/cuda-python python globalToShmemAsyncCopy.py [globalToShmemAsyncCopy] - Starting... Traceback (most recent call last): File "/home/jkh/Src/cuda-python/examples/0_Simple/globalToShmemAsyncCopy.py", line 1054, in main() File "/home/jkh/Src/cuda-python/examples/0_Simple/globalToShmemAsyncCopy.py", line 1022, in main major = checkCudaErrors(cudart.cudaDeviceGetAttribute(cudart.cudaDeviceAttr.cudaDevAttrComputeCapabilityMajor, devID)) File "/home/jkh/Src/cuda-python/examples/common/helper_cuda.py", line 24, in checkCudaErrors raise RuntimeError("CUDA error code={}({})".format(result[0].value, _cudaGetErrorEnum(result[0])))

Note that I can get some of the other samples to pass, so this does not seem like a generic error.

vzhurba01 commented 2 years ago

v11.5.0 fixes initialization bug in this test.