search

inducer / pycuda

CUDA integration for Python, plus shiny features
http://mathema.tician.de/software/pycuda
Other
1.85k stars 288 forks source link

pycuda._driver.LogicError: cuMemHostAlloc failed: OS call failed or operation not supported on this OS #449

Open kukugpt opened 4 months ago

kukugpt commented 4 months ago

During loop processing tensor model inference, in the middle of the loop will appear “pycuda._driver.LogicError: cuMemHostAlloc failed: OS call failed or operation not supported on this OSXXX failure of TensorRT X.Y when running XXX on GPU XXX” when host_mem = cuda.pagelocked_empty(size, dtype)

Environment TensorRT Version: 8.6.1

NVIDIA GPU: T4

NVIDIA Driver Version: 515.65.01

CUDA Version:11.7

CUDNN Version:8.2.1

Operating System: CentOS

inducer commented 4 months ago

I don't know how to think about this issue without a reproducer. Also, given that the error mentions TensorRT (which PyCUDA doesn't interact with), there's a good chance it's unrelated.