DeepDetect builds with `TENSORRT_OSS` do require CUDA 10.2
beniz opened this issue · comments
Emmanuel Benazera commented
Otherwise the build error is as follows:
deepdetect/build/tensorrt-oss/src/tensorrt-oss/samples/common/sampleDevice.h:189:56: error: ‘cudaStreamCaptureModeGlobal’ was not declared in this scope
cudaCheck(cudaStreamBeginCapture(stream.get(), cudaStreamCaptureModeGlobal));
^~~~~~~~~~~~~~~~~~~~~~~~~~~
deepdetect/build/tensorrt-oss/src/tensorrt-oss/samples/common/sampleDevice.h:189:56: note: suggested alternative: ‘cudaStreamCaptureStatusNone’
cudaCheck(cudaStreamBeginCapture(stream.get(), cudaStreamCaptureModeGlobal));
^~~~~~~~~~~~~~~~~~~~~~~~~~~
cudaStreamCaptureStatusNone
FTR this also appears here: triton-inference-server/server#635
Using CUDA 10.2 fixes the issue. A check for CUDA 10.2 is probably needed in the CMakeFile
to catch the issue much earlier.
YaYaB commented
According to the issue on triton-infenrence-server it should work with cuda 10.1 no?
Emmanuel Benazera commented
dk, 10.1 is deprecated for us, and I don't believe we are using it anymore.