https://peertube.watch/videos/watch/ac465580-4293-4372-86fe-bac0cd3f15e3 We motivate using ML policies to replace fine-grained compiler optimization decisions, otherwise made by heuristics, and discuss: the current infrastructure available in the main repository of an open source compiler, Clang/LLVM;, lessons learned from the practical application of ML on two optimizations over the last few years: a register allocation policy, and a size-optimizing inliner policy, both trained with an open source solution, on both internal and open-source projects, such as Chrome and Fuchsia;, what we discovered as obstacles to the broader application of ML optimizations, and to broader community adoption., We'll conclude with current status and investigation directions. Speaker: Mircea Trofin, Track: AI and Machine Learning devroom, Room: UB2.252A (Lameere), When: Sunday 4 February 2024, 10:45–11:15 (Brussels time), Schedule: FOSDEM 2024 event page, Slides & links Slides, LLVM Discourse Tag, Large IR Corpus, LLVM MLGO Tooling: LLVM Dev Summit Talk, Gematria (latency prediction), Licence: Creative Commons Attribution 2.0 Belgium (CC BY 2.0 BE) (as stated in the FOSDEM video archive). Attribution: “ML Guided Optimizations in LLVM” by Mircea Trofin, FOSDEM 2024. Original recording: 2024/ub2252a/fosdem-2024-1958-ml-guided-optimizations-in-llvm.av1.webm from the FOSDEM video archive. This channel is an unofficial mirror; FOSDEM is a free and non-commercial event organised by the community for the community. Sun, 25 Oct 2026 01:52:21 GMT https://validator.w3.org/feed/docs/rss2.html PeerTube - https://peertube.watch https://peertube.watch/lazy-static/avatars/3b284627-b75f-409e-ac2e-10f4c2946f43.png https://peertube.watch/videos/watch/ac465580-4293-4372-86fe-bac0cd3f15e3 All rights reserved, unless otherwise specified in the terms specified at https://peertube.watch/about and potential licenses granted by each content's rightholder.