Triton Ascend Case Elemwise Cast

大shape类型转换(int8→fp16)优化:通过二次切分(BLOCK_SIZE+TILE_SIZE)提高UB利用率,核数2048时性能最优,适用于shape较大(百万级元素)的elementwise类型转换场景

wenyi-li Updated 10 repo stars

File contents

wenyi-li/awesome-agent-kernel-skills commit 0265784f7c

Frequently asked questions

npx get-skill add wenyi-li/triton-ascend-case-elemwise-cast