Lc Optimize

Optimize LuisaCompute DSL kernels using warp/wave primitives, shared-memory aggregation, and block-level collectives. Use when kernels bottleneck on atomics, reductions, or inter-thread communication.

luisagroup Updated 1.1k repo stars

File contents

luisagroup/luisacompute commit d9f7107311

Frequently asked questions

npx get-skill add luisagroup/lc-optimize