microsoft/DeepSpeedExamples

nvcc fatal : Unsupported gpu architecture 'compute_86' and nvcc fatal : Value 'c++17' is not defined for option 'std'

Xccanxin opened this issue · 1 comments

could you provide the cuda and torch version?
i've try cu116+torch1.13 / cu118+torch2.00 / cu12.2+torch2.02
they all fails due to the nvcc fatal error, it seems a version problem?

could you provide the cuda and torch version? i've try cu116+torch1.13 / cu118+torch2.00 / cu12.2+torch2.02 they all fails due to the nvcc fatal error, it seems a version problem?

seems it need a newer version of cuda, which can't be modify in anaconda envs. Install the cuda12.1 from https://developer.nvidia.com/cuda-toolkit-archive and it works