You can be a World Champion!
summary
A community-run implementation challenge proposes to turn fast theoretical Laplacian solvers into working software by running teams against public matrix generators (2D/3D meshes, image Laplacians, weighted expanders that force high condition numbers). Organizers time serial and parallel runs, report both wall-clock and core×time, require residuals near machine epsilon, run anonymized qualifiers, then hold a workshop and publish results so complete solvers and engineering tradeoffs (what’s fast, what fails) are exposed. laplacian implementation challenge
AI-written summary · about
chapters (11)
- 0:00Start
- 4:40Achieving Impact
- 8:20Overall Structure
- 11:20Many Events to Compete In (1)
- 17:29Many Events to Compete In (3)
- 19:16Organic Teams
- 19:58Teaming Up to Win
- 23:46Matchmaking
- 24:06Software-Archeology Virtual Teams
- 29:22Now for Serious Stuff
- 31:26Where (to Publish and to Meet)?
related talks
18mPortfolio Solving for Dafny
32mWhat the game industry taught me about performance
33mRunning Machine Learning in Production - a Journey to Success
12mHow do you deliver a feature on the biggest stage in the world?
38mSPONSOR WORKSHOP / Zhiyi Ma, Shagun Sodhani / Facebook
34mModern Machine Learning flow with Quilt and Polyaxon
33mProd, Do Stack: Production-Ready Solutions via Data Science Competitions
15mvLLM: Easy, Fast, and Cheap LLM Serving for Everyone