Hello all!
I’m trying to do ‘nvprof python myfile.py’ just as in
at 4:21’. But I see:
me@inspiron:/data$ nvprof python cuda_1.py
==3790== NVPROF is profiling process 3790, command: python cuda_1.py
==3790== Error: Internal profiling error 4168:999.
C[:5] = [2. 2. 2. 2. 2.]
C[-5:] = [2. 2. 2. 2. 2.]
VectorAdd took for 0.49910449981689453econds
======== Error: CUDA profiling error.
Can’t find much reference for this error within my SO.
Any thoughts on that?
Kind regards
It seems it’s not a Python-related error.
Also tried to reproduce the example from
at 6:30’ for a .cpp/.cu program, where I find:
me@inspiron:/data$ nvcc add.cu -o add_cuda
me@inspiron:/data$ nvprof ./add_cuda
==5887== NVPROF is profiling process 5887, command: ./add_cuda
==5887== Error: Internal profiling error 4168:999.
Max error: 0
======== Error: CUDA profiling error.
veraj
3
Are you using CUDA 9.1?
I’m afraid we haven’t official support Ubuntu18.04 for CUDA 9.1
Yes, I am indeed.
Well, thank you for your reply.