ThunderKittens ports to NVIDIA Vera Rubin, hitting 22 PFLOPS on NVFP4 GEMMs
HazyResearch · x · 2026-09-11
Together AI's kernels team got early access to NVIDIA's Vera Rubin NVL72 and ported ThunderKittens to the new platform in a few days.
- Reworked kernels implement nvfp4 and fp8 GEMMs on Rubin's new ISA
- Benchmarks exceed 22 and 12 PFLOPS respectively, competitive with cuBLAS + CUTLASS DSL
- The team describes the porting as smooth and published a blog with details
Related event: ThunderKittens Lands on Vera Rubin: NVFP4 GEMM Tops 22 PFLOPS(7 posts)→
More from Infra
- AI Infra Meme: While Countries Need TW-Scale Power for Datacenters, People Argue About Lightbulbs — mgostIH · 2026-10-27
- Legora runs 2,000 parallel agent evals on Modal Sandboxes for legal AI — blelbach · 2026-10-27
- Cloud release velocity scores: AWS 95, Azure 90, Google Cloud 82 — DavidLinthicum · 2026-10-27
- Selling a 5090 to scale local inference: one business weighs 5 hardware paths — Anarchaotic · 2026-10-27
- AWS tutorial: fleet-wide edge CV model lifecycle with SageMaker SDK v3, ~$0.07 per run — AWS ML Blog · 2026-10-27
- Qwen3.8-27B hits 40 tok/s on Jetson Thor with INT4 + MTP, accuracy beats FP16 — No-Dragonfly6246 · 2026-10-27