New all-reduce kernels deliver 12% decode speedup for DeepSeek-V4.1 on PCIe-connected GPUs
AlpinDale · x · 2026-10-23
Developer AlpinDale opened a PR optimizing DeepSeek-V4.1 inference on SM120 hardware: new all-reduce kernels targeting small-message latency across PCIe-connected GPUs, plus a native V4.1 backend. On 4x RTX 6000 Pro Server GPUs, end-to-end decode speeds up 12%. The 12.5k-line change also adds paired PCIe all-reduce and distributed dispatch test fixtures.
Related event: Community PR Speeds Up DeepSeek-V4.1 Inference by 12% on PCIe GPUs(2 posts)→
More from Infra
- Nvidia's Berlin GTC spotlights humanoid robots and Vera Rubin AI supercomputing — emmanuelvivier · 2026-10-23
- Lenovo adds 5 AI Library use cases, claims 1-week deployment and up to 70% lower token cost — shashib · 2026-10-23
- Your kernel doesn't need to beat cuBLAS on every shape, argues GPU dev — blelbach · 2026-10-23
- llama.cpp team launches Llama: one-click local AI on your Mac menu bar — Hugging Face · 2026-10-23
- China's daily token use hits 500T, up 4x as average price doubles to ¥13.46 — teortaxesTex · 2026-10-23
- Offshore oil rigs could be converted into AI data centers powered by stranded gas — clemnt · 2026-10-23