about Nv visual profiler v4.1

Accelerated Computing CUDA CUDA Programming and Performance
1

Could anyone clarify the mean of “Total Replay Overhead”, “Shared Moemory Replay Overhead”, “Global Memory Replay Overhead”, “Global Cache Replay Overhead” and “Local Cache Replay Overhead”, especially the different between “Global Memory Replay Overhead” and “Global Cache Replay Overhead”?

I am sorry that I can not find any information about that in NV Visual Profiler v4.1 help External Image .

Thank you!

2

Tooltip help is available for these in the “Details” view. Try hovering over the column headers. Is that level of information sufficient. Tooltip help is not available in the “Properties” view… is that where you were looking at these values?

3

Thank you, David!

I have got that tooltip.