I want to generate the gpu program profile and then use NVVP to visualize it. However, when I try to use the following command based on GPU Tesla V100, I get the following error. Does Tesla V 100 support profile? My cuda version is 10.0.
nvprof --analysis-metrics -f -o sgemm.prof ./sgemm
==9078== NVPROF is profiling process 9078, command: ./sgemm -f
==9078== Some kernel(s) will be replayed on device 0 in order to collect all events/metrics.
==9078== Replaying kernel “sgemm_gpu_kernel(float const *, float const , float, int, int, int, float, float)” (1 of 2Replaying kernel “sgemm_gpu_kernel(float const *, float const , float, int, int, int, float, float)” (done)
==9078== Error: Internal profiling error 4183:7.
======== Error: CUDA profiling error.