Inside the Together Compiler: from Megakernels and PTX hacking to agent-driven optimization
simran_s_arora · x · 2026-11-05
Together researcher simransarora shares the personal arc behind the Together Compiler: from ThunderKittens and Megakernels to a year at Together scaling optimization from hand-tuning every customer to peak performance delivered in hours. The compiler took several forms: (1) trying to productionize Megakernels; (2) augmenting the Inductor FX graph by lowering all code to PTX, inserting PDL, and launching in one giant C++ Cubin to avoid PyTorch dispatch overhead; (3) hypothesizing agents would drown in millions of lines of inference engine code and kernel dispatch mechanisms, building a highly instrumented environment (thread continues).
More from coding & agent
- Dev finds pi coding tool defaults to Opus 4.8 on a fresh install — lucasmeijer · 2026-11-05
- Kylon puts whole teams and agents in shared rooms, with a Neon branch per app draft — jefrankle · 2026-11-05
- "Evals > vibes": agent workflow debates are just guessing without eval stats — annetgriffin · 2026-11-05
- OpenAI hackathon winner joins YC, launches Raven, an agent-first phone — ycombinator · 2026-11-05
- Open-source Cicada brings real-time translation and AI meeting notes to any Mac call, free 50h/month — dotey · 2026-11-05
- IBM launches AI coding assistant named "Bob", devs can't stop laughing — zeeg · 2026-11-05