negative Global Store Efficiency

Hi, everyone!

I’m using nvvp in CUDA Toolkit 4.2. When I am profiling a GPU kernel, I get negative “Global Store Efficiency” on the kernel displayed in the “Properties” tab, e.g. Global Store Efficiency -36.2%. I expect this number should be within the range of 0% and 100%. Does anyone have similar problems like this? Is it normal?

Thanks!

Like you, I would expect this number to be between 0% and 100%, so this looks like a bug to me. If this happens with CUDA 4.2, I would suggest filing a bug against the profiler, attaching a self-contained repro program that demonstrates the problem. Thank you for your help.

Last night I got CUDA 5 preview via the CUDA Registered Developer program. I updated and configured the new tools on my personal GPU (geforce gt 525M). This bug seems gone :D

Our server (Tesla M2090) is configured with CUDA 4.2, the bug is still there. I’m looking forward to the new CUDA 5!

The attached picture shows the differences.
bug.png