I am using
nsight to visualize trace timelines output by
However, it is extremely slow to run (just collapsing one line can take up to 20 seconds), making it very difficult to use.
Just for a bit of context I am profiling a mesh-tensorflow model.
I can provide the prof file if needed.
My local machine is running on ubuntu 16.04 with 8 cores (no GPU).
I would like to know if there are ways to make sure nsight is faster.