Jane Street and NVIDIA speed up Mamba-2 with custom cuTile kernels
blelbach · x · 2026-10-23
Jane Street and NVIDIA published a technical blog on implementing Mamba-2's forward pass in cuTile, NVIDIA's Python-based tile-centric DSL. Three custom kernels were optimized: chunkcumsum (log-space cumulative decay), fwdpassstate (Tensor Core state advancement via ct.mma), and computeout (fused state, decayed masked interaction, and skip connection). Optimizations include latency hints, operand swapping to Tensor Memory, TMA-friendly layouts, and per-shape autotuning.
More from Infra
- 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
- New all-reduce kernels deliver 12% decode speedup for DeepSeek-V4.1 on PCIe-connected GPUs — AlpinDale · 2026-10-23
- Offshore oil rigs could be converted into AI data centers powered by stranded gas — clemnt · 2026-10-23
- Same native UI, two stacks: Swift build is ~80MB Mac-only, Zig build is ~10MB cross-platform — jedisct1 · 2026-10-23
- SurfSense: an open-source local NotebookLM with MCP agents and a prompt-caching lesson — FurtiveMirth · 2026-10-23