viable/strict/1792837579: Use vectorized kernels for index_copy (#201814)

发生了什么

Linked issue or supporting maintainer Supported by @ngimel BC-breaking? No. Summary (human written only) Use vectorized kernel for scatter and index_copy ops, redispatch index_copy to scatter when profitable Cap the number of blocks because that leads to better perf (revisit later to maybe use CLC) Agent details (AI-generated) Authored with assistance from Codex. Adds a copy specialization to the existing native vec…

摘要按规则整理自下方来源原文

来源