I am following CUDA exercises and was able to create two version of reverseArray. One that use shared memory and one that doesn’t.
When I run CUDA Visual Profiler on the compiled exe (in release mode), I only see the GPU Time, CPU Time and Occupancy columns for these sessions. I did select the counters under Session Configuration and the profiler does run my exe 3 times. However, no counter is reported (gld coalesce, gld uncoalesce …)
Am I missing something?