burkliu
burkliu
my conda env conda install gcc=11 gxx=11 python=3.8 git git-lfs make cmake cuda cudnn -c conda-forge -c nvidia conda install pytorch torchvision torchaudio cudatoolkit=11.6 -c pytorch -c conda-forge
error occurs when I call pip install kaldifeat
nvcc -arch sm_86 test.cu -o test && ./test 11.7
I tried conda install gcc=11 gxx=11 python=3.8 git git-lfs make cmake cuda=11.6 cudnn -c conda-forge -c nvidia use cuda=11.6 fail too
I create a new env using conda install gcc=11 gxx=11 python=3.8 git git-lfs make cmake cuda -c conda-forge -c nvidia/label/cuda-11.6.2 conda install pytorch torchvision torchaudio cudatoolkit=11.6 -c pytorch -c conda-forge...
nvcc -arch sm_86 test.cu -o test && ./test 11.6
> fail again after source activate-cuda-11.6.sh same error, missing mem and clock glibc
I find cuda.h in /cfs/user/burkliu/dev/anaconda3/envs/k2_p/include
g++ test.cc -o test && ./test test.cc:1:10: fatal error: cuda.h: No such file or directory 1 | #include | ^~~~~~~~ compilation terminated.
success as u said g++ -I $CUDA_HOME/include test.cc -o test && ./test 11.6