CONTRIBUTING.md

Building a Development Container

First, download any dependencies required in ci/install_deps.sh.

Build a test image

docker build -f amd64_10-1.dockerfile -t npt-dev .

Run an image

docker run -it --rm npt-dev

Delete all unused docker data:

docker prune system

Travis

CI is done through travis-ci.com. Travis builds the example code, as well as all docker images. Carl Pearson's docker hub account is used to push images up to cwpearson/nvidia-performance-tools on Docker Hub.

Resources

• Nvidia Docker Image Definitions

Roadmap

• Using Nsight Compute and Nsight Systems on EWS
• Instructions for remote profiling
• Nsight Systems: How to load missing source file
• Definitions for Various Performance Terms
  • Occupancy
  • Memory Hierarchy
  • Scheduling
    • Stall reasons
  • cudaStreams, cudaEvents
• CUDA Event and Stream timing examples
  • single-device
  • multi-device
• interacting with .qdrep files.
• interacting with .nsight-cuprof-report files.
• Best Practices
  • Fixing GPU frequency
  • initial CUDA runtime cost
  • Warmup Kernels
  • cuda-memcheck race condition and sync check?
• Is stream 0 the default stream?
• Nsight System with MPI
• Nsight System with multi-GPU
• Nsight Compute multi-GPU

Inspirations

• https://developer.download.nvidia.com/video/gputechconf/gtc/2020/presentations/s22141-what-the-profiler-is-telling-you-how-to-get-the-most-performance-out-of-your-hardware.pdf
  • https://developer.nvidia.com/gtc/2020/video/s22141
• https://developer.download.nvidia.com/video/gputechconf/gtc/2020/presentations/s21351-scaling-the-transformer-model-implementation-in-pytorch-across-multiple-nodes.pdf
  • https://developer.nvidia.com/gtc/2020/video/s21351