Skip to content

feat: inject HCCL OOM fault by counter - #1

Draft
yjyang62 wants to merge 3 commits into
pta_oom_2.9.1from
cursor/hccl-oom-fault-2bd5
Draft

feat: inject HCCL OOM fault by counter#1
yjyang62 wants to merge 3 commits into
pta_oom_2.9.1from
cursor/hccl-oom-fault-2bd5

Conversation

@yjyang62

@yjyang62 yjyang62 commented May 29, 2026

Copy link
Copy Markdown
Owner

【合入来源】

(如有)请关联需求文档/issue链接

  • 需求
  • 问题单
  • issue/工单
  • 重构优化
  • 资料更新

【修改方案】

  1. 在 HCCL 错误检查层增加计数器,复用 PTA OOM 的计数触发方式。
  2. 当计数器命中第 60004 次 HCCL 检查时,不调用底层 HCCL 表达式,直接构造并上抛 HCCL_E_OOM
  3. HCCL_E_OOM 统一转为 OutOfMemoryError,保留 HCCL 错误码信息和 OOM snapshot 触发逻辑。

【资料变更】

不涉及

【接口变更】

不涉及

【功能验证】

  • git diff --check pta_oom_2.9.1..HEAD
  • 最终 diff 检查,确认未修改 third_party/HCCL 底层实现,仅在 HCCL 错误检查层构造上抛错误码。

【CheckList】

  • 代码注释完备,正确记录错误日志
  • 代码实现进行了返回值、空指针等校验
  • PR标题正确使用类型标签,如:feat、fix、refactor、docs、test等
  • PR持续集成流水线(CI)执行通过,代码检查无异常
Open in Web Open in Cursor 

cursoragent and others added 3 commits May 29, 2026 03:43
Co-authored-by: yjyang62 <yjyang62@users.noreply.github.com>
Co-authored-by: yjyang62 <yjyang62@users.noreply.github.com>
Co-authored-by: yjyang62 <yjyang62@users.noreply.github.com>
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

2 participants