Skip to content
Zack Galbreath edited this page Oct 13, 2023 · 1 revision

Attendees

  • Aashish Chaudhary
  • Alec Scott
  • Jacob Nesbitt
  • John Parent
  • Massimiliano Culpo
  • Dan LaManna
  • Mike VanDenburgh
  • Ryan Krattiger
  • Scott Wittenburg
  • Todd Gamblin
  • Zack Galbreath

CI pipelines optimization

  • Try to reduce the number of ParaView builds we currently perform. This requires deciding which ones to keep and which ones to eliminate.
  • Look for ways to speed up jobs for Python packages. We're hopeful CloudFront caching will make a big impact here.
  • Starting using our timing data to look for other opportunities to speed things up.

Priorities

  • Start using CloudFront to cache our build caches and monitor the impact this change has on the speed of our CI pipelines (PR #39939)
  • Take a closer look at our timing data and use that to determine what opportunities exist to further speed up our pipelines.
  • Look for opportunities to reduce duplication between stacks (PR #38512, issue #32893)
  • Create a dashboard showing the number of jobs per stack over time
  • Add a column to cache.spack.io showing what stack a binary was built for
  • Store tarballs by content hash in our binary mirrors.
Clone this wiki locally