HazyResearch Open-Sources MoK: A Fused MoE Megakernel for NVL72
HazyResearch · x · 2026-08-05
HazyResearch announced the open-source release of Mixture-of-Kittens (MoK), a Mixture-of-Experts (MoE) megakernel designed specifically for NVL72.
MoK fuses all MoE-related communication and computation into a single, fully deterministic kernel, powering Composer training across tens of thousands of GPUs.
Related event: Cursor Open-Sources MoK Megakernel, Nearly Doubling NVL72 MoE Performance(5 posts)→
More from Infra
- Nvidia's Strategic Shift: From Chip Sales to Cloud Revenue Sharing — GavinSBaker · 2026-08-05
- llama.cpp PR Caches Hot MoE Experts on GPU, Doubling Inference Speed on 8GB VRAM — BTA_Labs · 2026-08-05
- Big Tech Locks in Over $1 Trillion in Future AI Data Center Lease Payments — Polymarket · 2026-08-05
- YC-Backed Lamb Labs Claims 63x Higher Efficiency for AI Inference Chips vs GPUs — ycombinator · 2026-08-05
- DSpark Open-Sources Speculative Decoding Path for Kimi K3 — ying11231 · 2026-08-05
- US Drafts Ban on Chinese Datacenter Components as Europe Pushes for Tech Sovereignty — nordicinst · 2026-08-05