8% pretraining and 20% RL speedup from native 4-bit NVFP4 training, screenshots show
snowclipsed · x · 2026-11-14
The author shares a screenshot quantifying the gains of the NVFP4 recipe: about an 8% wall-clock pretraining speedup and a 20% RL speedup over full MXFP8 training in the largest jobs, with no calibration needed when the quantized checkpoint ships. They also ask whether the experts use MXFP8.
More from Infra
- Atreides sends letter to Nvidia board outlining minimum governance standards — GavinSBaker · 2026-11-14
- OSINT deep-dive: FAA nationwide grounding caused by long-haul ATC comms, radar and GPS fine — eigenrobot · 2026-11-14
- Airtel chairman: Starlink V3 delivers 1 Tbps per launch, matching all of OneWeb — elonmusk · 2026-11-14
- Dev goes all-in on Cloudflare, citing its excellent MCP as the deciding factor — ritakozlov · 2026-11-14
- Brave Leo adds verifiable encrypted LLM inference via TEEs, calls ZDR clauses a pinky promise — gnukeith · 2026-11-14
- 97% of coding-agent tokens are cache reads, but cache writes eat 31% of the cost — Nitjsefnie · 2026-11-14