Pull Request已成功合入, 合并人@CANN-robot
(感谢 zl_hw 的贡献)变更摘要
本 PR 针对 arch35(Ascend950)平台的 segsum 算子内核头文件进行改名:将 math/segsum/op_kernel/arch35/ 目录下的头文件 segsum.h 重命名为 segsum_arch35.h,以避免与 A2 平台同名的头文件产生冲突,并同步更新了引用该头文件的源码及文件注释。
主要改动
- 头文件重命名:将
math/segsum/op_kernel/arch35/segsum.h重命名为segsum_arch35.h,通过文件名带arch35后缀区分不同架构实现,避免与 A2 平台同名头文件冲突。 - 更新 include 引用:在
math/segsum/op_kernel/arch35/segsum.cpp中将#include "segsum.h"改为#include "segsum_arch35.h",保证 arch35 内核入口能正确引用改名后的头文件。 - 同步文件注释:在
segsum_arch35.h中将\file segsum.h的注释更新为\file segsum_arch35.h,保持文档描述与实际文件名一致。


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.
You can self-configure the PR merge rules for this repository. For more details, please refer to Here.
For more, you also can visit HICANN.
PR Approval Progress
✅ Congratulations! All modules have met the lgtm and approve requirements.
Module Approval Details
| module | lgtm status | approve status |
|---|---|---|
| math/segsum | ✅ 汤平川, 宋恺 (2/2) | ✅ 汤平川 (1/1) |
💡 Tip:
- Committer can comment
/approveor/lgtm- Commenting
/approveimplies both code review (lgtm) and intent to merge (approve)
CLA Signature Pass
zl_hw, thanks for your pull request. All authors of the commits have signed the CLA. 👍


流水线任务触发成功
任务链接 [6b853081097444fdaa859a3427b11fd7][流水线指导]
| 任务名称 | 状态 | 日志 | 下载链接 |
|---|---|---|---|
| pre_comment | ✅ SUCCESS | >>>>> | |
| Compile_Ascend_X86_ubuntu24 | ✅ SUCCESS | >>>>> | >>>>> |
| Compile_Ascend_X86_mobile_station_ubuntu24 | ✅ SUCCESS | >>>>> | >>>>> |
| Compile_Ascend_X86_experimental_ubuntu24 | ✅ SUCCESS | >>>>> | >>>>> |
| Compile_Ascend_ARM_ubuntu24 | ✅ SUCCESS | >>>>> | >>>>> |
| Compile_Ascend_ARM_experimental_ubuntu24 | ✅ SUCCESS | >>>>> | >>>>> |
| Compile_Ascend_single_ubuntu24 | ✅ SUCCESS | >>>>> | >>>>> |
| Compile_X86_monitor_910b | ✅ SUCCESS | >>>>> | >>>>> |
| Compile_X86_monitor_910c | ✅ SUCCESS | >>>>> | >>>>> |
| Compile_X86_monitor_950 | ✅ SUCCESS | >>>>> | >>>>> |
| Compile_Ascend_X86_A5_ubuntu24 | ✅ SUCCESS | >>>>> | >>>>> |
| Compile_Ascend_ARM_A5_ubuntu24 | ✅ SUCCESS | >>>>> | >>>>> |
| Compile_Ascend_X86_mobile_station_9030_ubuntu24 | ✅ SUCCESS | >>>>> | >>>>> |
| Compile_Ascend_single | ✅ SUCCESS | >>>>> | >>>>> |
| UT_Test | ✅ SUCCESS | ||
| UT_Test_experimental | ✅ SUCCESS | ||
| UT_Test_kernel | ✅ SUCCESS | ||
| PreSmoke_A900 | ✅ SUCCESS | >>>>> | |
| API_Check | ✅ SUCCESS | >>>>> | |
| PreSmoke_ATK_Test_A2 | ✅ SUCCESS | >>>>> |
[2026-08-25 15:41:32] CI执行结束


流水线任务触发成功
任务链接 [cd28f012de6945159e5323ce04b48ae7][流水线指导]
| 任务名称 | 状态 | 日志 | 下载链接 |
|---|---|---|---|
| SCA | ✅ SUCCESS | >>>>> | |
| antipoison | ✅ SUCCESS | >>>>> | |
| codecheck_Pr | ✅ SUCCESS | ||
| StaticCheck_codespell | ✅ SUCCESS | ||
| StaticCheck_link_validity | ✅ SUCCESS | ||
| StaticCheck_resource_existence | ✅ SUCCESS | ||
| StaticCheck_tag_closed | ✅ SUCCESS | ||
| StaticCheck_markdownlint | ✅ SUCCESS | ||
| codecheck_precommit | ✅ SUCCESS | >>>>> |
[2026-08-25 15:36:18] CI执行结束


/lgtm
/approve


/lgtm


描述
将
math/segsum/op_kernel/arch35/segsum.h更名为segsum_arch35.h,并同步arch35/segsum.cpp的 include。改动原因:该头文件与 A2 的
math/segsum/op_kernel/segsum.h基名完全相同,两者都可能出现在内核编译的头文件搜索路径上,#include "segsum.h"的落点依赖 include 顺序;检视与排障时也无法从文件名分辨代次。仓内 248 个含op_kernel/arch35的算子中仅 4 个存在这种同名(mul_addn、pad_v3_grad_replication、exp_segsum_grad、segsum),主流写法是 arch35 侧另起名(如conversion/slice/op_kernel/arch35/slice_arch35.h);本算子 host 侧已经是segsum_tiling_arch35.cpp/.h,本次把内核头文件对齐。改动方法(2 个文件,各 1 行):
op_kernel/arch35/segsum.h→op_kernel/arch35/segsum_arch35.h(git rename,相似度 99%,仅同步\file注释)op_kernel/arch35/segsum.cpp:#include "segsum.h"→#include "segsum_arch35.h"未改动项:
SEGSUM_ARCH35_H、命名空间本就是SegsumArch35,无需调整arch35/segsum.cpp不更名:.cpp不被 include,由CMakeLists.txt的KERNEL_SRC arch35/segsum.cpp按显式路径编译,segsum_def.cpp的ExtendCfgInfo("opFile.value", "segsum")也以算子名为准,同名不产生歧义segsum;A2 侧op_kernel/segsum.{cpp,h}未改动;arch35/segsum_tiling_data.h无同名冲突,未改动影响面核查:全仓检索旧名引用,除本次改掉的一处外无其他引用;产物命名不受影响(
<op>_apt目录与 json 取自 OP_NAME 而非文件名,见cmake/gen_ops_info.cmake);kernel UT 按KERNEL_SRC的值 glob 入口文件(cmake/ut.cmake),入口未更名故不受影响;打包用find -maxdepth 1 -exec cp -r保留arch35/子目录结构。关联的Issue
测试
重新出包:
bash build.sh --pkg --soc=ascend950 --ops=segsum,编译无 error;出包内op_impl/ai_core/tbe/custom_math_impl/ascendc/segsum/arch35/下已只有segsum_arch35.h。真机 Ascend950PR(TTK kernel 模式,
-d false -b release,判定取部署 binary 腿):合计 118/118 PASS;
bin_compile_s全部BINARY_MATCH,越界检查bin_oob_result全OK,OK,memory_oob_status全 PASS,两个 tilingKey 分支均已覆盖。文档更新
无。
类型标签