Latest CUDA release includes instruction level profiling feature:
http://devblogs.nvidia.com/parallelforall/cuda-7-5-pinpoint-performance-problems-instruction-level-p...
Is this feature available, or planned, for CodeXL?
Thanks!
Hi,
That is a useful capability and we're planning to provide it in a future CodeXL release.
Doron
Thanks. Looking forward to using this when it's ready.
Cheers,
Aaron