{"type":"post","id":"how-we-built-tlx-a-hardware-native-gpu-31371f","slug":"how-we-built-tlx-a-hardware-native-gpu-31371f","title":"How We Built TLX: A Hardware-Native GPU Compiler for Production ML at Scale","description":"Meta has introduced Triton Low-Level Extensions (TLX), an open-source extension to Triton that exposes explicit warp-group and cluster-level orchestration for GPU workloads.","retrieval_nugget":"Meta has introduced Triton Low-Level Extensions (TLX), an open-source extension to Triton that exposes explicit warp-group and cluster-level orchestration for GPU workloads. Meta says TLX has improved recommendation-system training, recommendation inference, and large-scale LLM pretraining; it may offer a route to newer GPU capabilities without abandoning Triton’s programming model.","published_at":"2026-11-10","updated_at":"2026-11-10","record_date":"2026-11-10","date_kind":"published_at","topics":["orchestration"],"entities":["Meta","TLX","Triton"],"related_patterns":["harness-architecture-outlives-model-choice"],"novelty_label":"new_fact","what_changed":"Meta introduces TLX, a Triton extension adding warp-group and cluster-level GPU orchestration, with reported gains in recommendation-system training and inference and large-scale LLM pretraining.","prior_refs":[],"source_url":"https://engineering.fb.com/2026/11/10/open-source/tlx-a-hardware-native-gpu-compiler-for-production-ml","source_urls":["https://engineering.fb.com/2026/11/10/open-source/tlx-a-hardware-native-gpu-compiler-for-production-ml"],"schema_version":"newruntime-agent-readable-v0.2","stable_id":"post:how-we-built-tlx-a-hardware-native-gpu-31371f","status":"published","visuals":[],"editor":{"role":"editor","name":"Andrey Reshetnikov","profile":"https://newruntime.com/owner-profile.md","profile_json":"https://newruntime.com/owner-profile.json","same_as":["https://www.linkedin.com/in/reshetnikov1/","https://t.me/lapetuse"]},"routes":{"html":"https://newruntime.com/posts/how-we-built-tlx-a-hardware-native-gpu-31371f/","markdown":"https://newruntime.com/posts/how-we-built-tlx-a-hardware-native-gpu-31371f.md","json":"https://newruntime.com/posts/how-we-built-tlx-a-hardware-native-gpu-31371f.json"},"markdown_url":"https://newruntime.com/posts/how-we-built-tlx-a-hardware-native-gpu-31371f.md","next_reads":[{"type":"topic","path":"/topics/orchestration/","url":"https://newruntime.com/topics/orchestration/","title":"Orchestration","reason":"Explore the orchestration topic hub.","media_type":"text/html","markdown_url":"https://newruntime.com/topics/orchestration.md","json_url":"https://newruntime.com/topics/orchestration.json"},{"type":"collection","path":"/posts/","url":"https://newruntime.com/posts/","title":"Field Notes","reason":"Open the latest editorial analysis.","media_type":"text/html","json_url":"https://newruntime.com/posts.json"},{"type":"collection","path":"/sources/","url":"https://newruntime.com/sources/","title":"Sources","reason":"See the original sources this archive cites.","media_type":"text/html"}]}