I create an NVTX range for each host iteration in my cuda application. But the timeline shown in the report file only have partial iterations. And the last nvtx range is very long, maybe it contains all the subsequent timeline data? (to be more precisely, total iteration in my host code is 3e5, but the timeline only have 1.2e4 nvtx iterations.)
Is this limit caused by Nsight System itself or caused by my OS resources, such as CPU memory limitation?