GPU Performance Profiling In Progress

Practical tools and workflows for profiling GPU utilization, memory usage, and performance bottlenecks
  1. Reading torch.profiler Correctly: Schedules, Labels, and CPU vs GPU Time
    torch.profiler integrates PyTorch operator attribution, optional source stacks, and CUDA activity in one trace. Its numbers mean what you think only if you control the capture window, the labels and the host syncs.
  2. GPU Profiling with Nsight Systems
    Nsight Systems puts host ranges, CUDA API calls, GPU kernels and memory operations on one timeline. It helps locate gaps in recorded GPU activity and inspect the host-side work around them.
  3. Introduction to GPU Profiling
    Learn how host work, CUDA submission, and GPU execution appear in a profile, and when to use PyTorch Profiler or Nsight Systems.
  4. PyTorch Profiler vs Nsight Systems
    Both tools record CUDA activity. The useful distinction is how they connect it to framework operations, tensor memory, host threads, and system behavior.