已开启
feat: add CustomGRUV2 operator #10
nawenyu139创建于 8月19日
feat: add CustomGRUV2 operator #10
已开启
合并受阻
8月19日 创建了 pull request,commit 4609d8d5
8月19日 添加了label:cann-cla/yes
CANN-robot
8月19日 评论:
8月19日 评论:
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
⚠️ This PR does not yet meet the following requirements:lgtm (requires ≥ 2 person(s) per module)、approve (requires ≥ 1 person(s) per module)
Module Approval Details
| module | lgtm status | approve status |
|---|---|---|
| sig-oil-gas-engineering | ❌ (0/2)(You can also ask: 胡艺馨, 张乐乐, 李鹏, 孙浩, 张宝凌) | ❌ (0/1)(You can also ask: 李一鸣, 张乐乐, 孙浩, 张宝凌, 徐澜) |
💡 Tip:
- Committer can comment
/approveor/lgtm- Commenting
/approveimplies both code review (lgtm) and intent to merge (approve)
CLA Signature Pass
nawenyu139, thanks for your pull request. All authors of the commits have signed the CLA. 👍


nawenyu139
8月19日 评论:
8月19日 评论:
已用个人绑定邮箱(slamdunk111@163.com,已签CLA)重新 squash 单一 commit 并解耦无关算子。重新触发审核。


nawenyu139
8月19日 评论:
8月19日 评论:
/check-cla


此处折叠了156条消息 查看更多
2 天前 预合并成功(commit_id: 33befdbedba5536b1864b8a54bc4bb05099d35aa)
CANN-robot
2 天前 评论:
2 天前 评论:
| 🚀 CI 流水线已启动 |
|---|
| 📋 执行详情: 点击查看流水线 |


2 天前 添加了label:ci-pipeline-running
2 天前 删除了label:ci-pipeline-running
2 天前 添加了label:ci-pipeline-passed
变更规模:+5271 行新增 / -1 行删除,共 52 个文件
背景
分支
feat/custom-gru-v2→master,含 1 个 commit。master当前 tip:8037f22 feat: add Triton-Ascend interpolate operator。Commits(各 commit 做什么)
ab7898e404e6abdaa7ec80d0c735aa4c22d5c135feat: add CustomGRUV2 operator单提交重铸+PR 面收敛:剔除与 CustomGRUV2 无关的 triton 系搭车文件
(axis_spectral/real_fft/_local_ai4fwi_ops/whl 冒烟/build_install_ops.sh,与 #57/#58/#59 同标准);
ai4fwi 包 init 对齐 #57 版(all 去 triton);custom_gru_v2 全部代码逐字节不变。
变更面
52 files changed, 5271 insertions(+), 1 deletion(-)M README.mdA ascendc/operators/custom_gru_v2/delivery/README.mdA ascendc/operators/custom_gru_v2/delivery/bindings.cppA ascendc/operators/custom_gru_v2/delivery/custom_gru_v2.pyA ascendc/operators/custom_gru_v2/op_host/custom_gruv2.cppA ascendc/operators/custom_gru_v2/op_host/custom_gruv2_tiling.hA ascendc/operators/custom_gru_v2/op_kernel/custom_gruv2.cppA ascendc/operators/custom_gru_v2/testplan/README.mdA ascendc/operators/custom_gru_v2/testplan/conftest.pyA ascendc/operators/custom_gru_v2/testplan/coverage/run_coverage.pyA ascendc/operators/custom_gru_v2/testplan/integration/test_e2e_inference.pyA ascendc/operators/custom_gru_v2/testplan/integration/test_e2e_training.pyA ascendc/operators/custom_gru_v2/testplan/msopst/_memcheck_run.pyA ascendc/operators/custom_gru_v2/testplan/msopst/_prof_run.pyA ascendc/operators/custom_gru_v2/testplan/msopst/_racecheck_run.pyA ascendc/operators/custom_gru_v2/testplan/msopst/custom_gruv2_test.jsonA ascendc/operators/custom_gru_v2/testplan/msopst/npu_reports/NPU_TEST_REPORT.mdA ascendc/operators/custom_gru_v2/testplan/msopst/npu_reports/raw/memcheck.logA ascendc/operators/custom_gru_v2/testplan/msopst/npu_reports/raw/msopst_run.logA ascendc/operators/custom_gru_v2/testplan/msopst/npu_reports/raw/msprof_run.logA ascendc/operators/custom_gru_v2/testplan/msopst/npu_reports/raw/op_statistic_batch.csvA ascendc/operators/custom_gru_v2/testplan/msopst/npu_reports/raw/op_statistic_single.csvA ascendc/operators/custom_gru_v2/testplan/msopst/npu_reports/raw/op_summary_batch.csvA ascendc/operators/custom_gru_v2/testplan/msopst/npu_reports/raw/op_summary_single.csvA ascendc/operators/custom_gru_v2/testplan/msopst/npu_reports/raw/racecheck.logA ascendc/operators/custom_gru_v2/testplan/msopst/npu_reports/raw/sanitizer_run.logA ascendc/operators/custom_gru_v2/testplan/msopst/npu_reports/raw/st_report.jsonA ascendc/operators/custom_gru_v2/testplan/msopst/run_prof.shA ascendc/operators/custom_gru_v2/testplan/msopst/run_sanitizer.shA ascendc/operators/custom_gru_v2/testplan/pack_submission.shA ascendc/operators/custom_gru_v2/testplan/performance/bench_backward.pyA ascendc/operators/custom_gru_v2/testplan/performance/bench_forward.pyA ascendc/operators/custom_gru_v2/testplan/pytest.iniA ascendc/operators/custom_gru_v2/testplan/requirements.txtA ascendc/operators/custom_gru_v2/testplan/run_all_tests.shA ascendc/operators/custom_gru_v2/testplan/test_evidence.mdA ascendc/operators/custom_gru_v2/testplan/test_plan.mdA ascendc/operators/custom_gru_v2/testplan/unit/test_backward_precision.pyA ascendc/operators/custom_gru_v2/testplan/unit/test_coverage_paths.pyA ascendc/operators/custom_gru_v2/testplan/unit/test_edge_cases.py验证
bash scripts/check_license_header.sh(40 个纳管扩展变更文件)→ passed,0 缺失--dry-run -Werror):5 个 c/cpp/h → 0 违规;trailing-whitespace 0、end-of-file 01024-油气算子PR9-PR10-精度性能测试执行指导书-20261024.md)执行中,回传后补贴最新截图与报告合入
由 maintainer/owner 手动合入;两提交形态如需重铸单提交由 owner 执行(force push)。
精度与性能测试结果(SACT KB
3dc60a7,2026-10-27 真机)CustomGRUV2 — 全部通过:UT 54/55、IT 10/11、memcheck/racecheck PASS、精度 1.19e-07、msProf 2 shapes、稳定性 10×10 一致
完整报告原文见本 PR 讨论区最新评论;原始证据(UT/IT/memcheck/msprof/bench 日志与 CSV)位于
SACT Knowledge Base 仓
oil-gas-ops/raw/(commit3dc60a7)。测试报告附件
本 PR 直接携带 NPU 精度 + 性能测试报告文件(点击下载查看):
tests/reports/custom_gru_v2_precision_perf.mdtests/reports/custom_gru_v2_msopst.md2026-10-28 合入冲突刷新
master 前进至
8037f22(新增 Tritoninterpolate算子)后本分支出现README.md冲突,冲突面仅「接口支持清单」的算子计数行。已按并集解冲突并重建为基于8037f22的单提交:AscendC 4→5、Triton 3→4、Composite 保持 1,表格两侧新增行(interpolate与本 PR 算子)均保留;算子代码零改动。