Pull Request已成功合入, 合并人@CANN-robot
（感谢 luwenxiang 的贡献）
Thanks for your pull-request.
The full list of commands accepted by me can be found at here。
You can get sig-info at here
PR Approval Progress
✅ Congratulations! All modules have met the lgtm and approve requirements.
Module Approval Details
| module | lgtm status | approve status |
|---|---|---|
| conversion/broadcast_to | ✅ llimwang, songkai111 (2/2) | ✅ llimwang (1/1) |
💡 Tip:
- Committer can comment
/approveor/lgtm- Commenting
/approveimplies both code review (lgtm) and intent to merge (approve)
CLA Signature Pass
luwenxiang1998, thanks for your pull request. All authors of the commits have signed the CLA. 👍
CANN-robotCANN-robot🔵 source code change are detected, tasks labels is removed in this pull request!
问题/功能描述
本次PR移除了BroadcastTo算子在特定TILING模式下的一个过度优化逻辑，解决了其可能引发的潜在性能问题。同时，通过扩展DMA广播模式的判断条件，优化了特定维度组合下的计算路径选择，旨在提升算子在更广泛场景下的整体性能。
修改方案描述
具体修改包括：1）在UpdateTilingKey函数中，移除了TILING_MODE_LAST_DIM_SMALL_A模式下根据单次搬运数据量与缓存行大小关系来动态调整uLpUnit_的代码。2）在CalcTensorSize函数中，为isDMABrcA_的判断新增了一个分支，当输出张量最后一维尺寸小于8且倒数第四维及之后的维度总大小小于最小张量尺寸时，也启用DMA广播模式，从而优化了计算路径。
luwenxiangluwenxiang🔵 source code change are detected, tasks labels is removed in this pull request!
compile
流水线任务触发成功
任务链接 [fb82badc21cb4ab2b403d49d2236cb08][流水线指导]
| 任务名称 | 状态 | 日志 | 下载链接 |
|---|---|---|---|
| codecheck | ✅ SUCCESS | >>>>> | |
| SCA | ✅ SUCCESS | >>>>> | |
| anti_virus | ✅ SUCCESS | >>>>> | |
| Check_Pr | ✅ SUCCESS | >>>>> | |
| Compile_Ascend_X86 | ✅ SUCCESS | >>>>> | >>>>> |
| Compile_Ascend_X86_mobile_station | ✅ SUCCESS | >>>>> | >>>>> |
| Compile_Ascend_ARM | ✅ SUCCESS | >>>>> | >>>>> |
| Compile_Ascend_X86_experimental | ✅ SUCCESS | >>>>> | >>>>> |
| Compile_Ascend_ARM_experimental | ✅ SUCCESS | >>>>> | >>>>> |
| Compile_Ascend_single | ✅ SUCCESS | >>>>> | >>>>> |
| Compile_Ascend_X86_A5 | ✅ SUCCESS | >>>>> | >>>>> |
| Compile_Ascend_ARM_A5 | ✅ SUCCESS | >>>>> | >>>>> |
| pre_comment | ✅ SUCCESS | >>>>> | |
| Compile_Ascend_X86_ubuntu24 | ✅ SUCCESS | >>>>> | |
| Compile_Ascend_X86_mobile_station_ubuntu24 | ✅ SUCCESS | >>>>> | |
| Compile_Ascend_X86_experimental_ubuntu24 | ✅ SUCCESS | >>>>> | |
| Compile_Ascend_X86_A5_ubuntu24 | ✅ SUCCESS | >>>>> | |
| harmony-infer | ✅ SUCCESS | ||
| Compile_Ascend_ARM_ubuntu24 | ✅ SUCCESS | >>>>> | |
| Compile_Ascend_ARM_experimental_ubuntu24 | ✅ SUCCESS | >>>>> | |
| Compile_Ascend_single_ubuntu24 | ✅ SUCCESS | >>>>> | |
| Compile_Ascend_ARM_A5_ubuntu24 | ✅ SUCCESS | >>>>> | |
| Compile_X86_monitor | ✅ SUCCESS | >>>>> | >>>>> |
| Compile_ARM_monitor | ✅ SUCCESS | >>>>> | >>>>> |
| StaticCheck_codespell_check | ✅ SUCCESS | ||
| StaticCheck_link_validity_check | ✅ SUCCESS | ||
| StaticCheck_resource_existence_check | ✅ SUCCESS | ||
| StaticCheck_tag_closed_check | ✅ SUCCESS | ||
| StaticCheck_markdownlint | ✅ SUCCESS | ||
| UT_Test | ✅ SUCCESS | ||
| UT_Test_experimental | ✅ SUCCESS | ||
| UT_Test_kernel | ✅ SUCCESS | ||
| Smoke_A900_npupool | ✅ SUCCESS | >>>>> | |
| API_Check | ✅ SUCCESS | >>>>> |
[2026-03-28 20:36:07] CI执行结束
/lgtm
Review Guide
This pull-request passes review.
Committers who wrote a comment of /approve are: llimwang.
Reviewers who wrote a comment of /lgtm are: llimwang, songkai111.
描述
broadcast_to算子性能优化: nddma条件优化
关联的Issue
https://gitcode.com/cann/ops-math/issues/1109
测试
文档更新
类型标签