I’m trying to profile a CUDA kernel running on a remote machine using Nsight Compute.
I installed CUDA 10 on both the local (where I run Nsight Compute) and the target machine (where I run the kernel on a Tesla V100 GPU).
So when I launch nv-nsight-cu and click on "run to the next kernel" I get the error: kernel profiling is not supported on this device, despite the command nv-nsight-cu-cli --list-chips outputs the architecture gv100 which is the one of the Tesla V100.
I tried to build the program with and without the --arch flag but nothing changes.
I tried also to use CUDA 11 but it's the same.
I’m able to profile the kernel using nvprof but I would like to use the UI of Nsight Compute.
What can I do?
question from:https://stackoverflow.com/questions/65600441/cuda-kernel-profiling-is-not-supported-on-tesla-v100
