I have already read the relevant explanation in PROFILER USER’S GUIDE([url]https://docs.nvidia.com/cuda/profiler-users-guide/index.html#metrics-reference[/url]). Anyone who has read the documentation knows that there are very few explanations about these items. Who can explain these items in detail, it is best to have related examples, or provide relevant documentation.
For example, the description of the document for inst_per_warp is the Average number of instructions executed by each warp. What is the meaning of the instructions here?