0
代码
代码
Issues
Pull Requests
讨论
Wiki
分析
0
ascend-robotascend-robotmake watchdog thread a class
3a0bee1c创建于 2025年12月9日历史提交
文件最后提交记录最后更新时间
rpc
!24702 Code security Merge pull request !24702 from SCh-zx/safem 8 个月前
symm_mem
[SHMEM] support npu shmem Co-authored-by: wangchao430<wangchao430@huawei.com> # message auto-generated for no-merge-commit merge: !26027 merge v2.99.0_shmem1 into master [SHMEM] support npu shmem Created-by: wangchao430 Commit-by: wangchao430 Merged-by: ascend-robot Description: <!-- Thanks for sending a pull request! --> **What type of PR is this?** > Uncomment only one /kind <> line, hit enter to put that in a new line, and remove leading whitespaces from that line: > > /kind bug > /kind task > /kind feature **What does this PR do / why do we need it**: **Which issue(s) this PR fixes**: <!-- *Automatically closes linked issue when PR is merged. Usage: Fixes #<issue number>, or Fixes (paste link of issue). --> Fixes # **Special notes for your reviewers**: See merge request: Ascend/pytorch!260276 个月前
CMakeLists.txt
[SHMEM] support npu shmem Co-authored-by: wangchao430<wangchao430@huawei.com> # message auto-generated for no-merge-commit merge: !26027 merge v2.99.0_shmem1 into master [SHMEM] support npu shmem Created-by: wangchao430 Commit-by: wangchao430 Merged-by: ascend-robot Description: <!-- Thanks for sending a pull request! --> **What type of PR is this?** > Uncomment only one /kind <> line, hit enter to put that in a new line, and remove leading whitespaces from that line: > > /kind bug > /kind task > /kind feature **What does this PR do / why do we need it**: **Which issue(s) this PR fixes**: <!-- *Automatically closes linked issue when PR is merged. Usage: Fixes #<issue number>, or Fixes (paste link of issue). --> Fixes # **Special notes for your reviewers**: See merge request: Ascend/pytorch!260276 个月前
HCCLUtils.cpp
check and log hccl error Co-authored-by: zhaoyu<nanzhaogang@qq.com> # message auto-generated for no-merge-commit merge: !26292 merge hccl_error into master check and log hccl error Created-by: zhaoyu65 Commit-by: zhaoyu Merged-by: ascend-robot Description: <!-- Thanks for sending a pull request! --> **What type of PR is this?** > Uncomment only one /kind <> line, hit enter to put that in a new line, and remove leading whitespaces from that line: > /kind feature **What does this PR do / why do we need it**: **Which issue(s) this PR fixes**: <!-- *Automatically closes linked issue when PR is merged. Usage: Fixes #<issue number>, or Fixes (paste link of issue). --> **Special notes for your reviewers**: HCCL错误打屏，PTA透传HCCL的错误码和错误信息。 代码逻辑：先判断是否有hccl错误，如果有错误，通过aclGetRecentErrMsg接口获取最近的错误并打印到日志里。 See merge request: Ascend/pytorch!262926 个月前
HCCLUtils.hpp
check and log hccl error Co-authored-by: zhaoyu<nanzhaogang@qq.com> # message auto-generated for no-merge-commit merge: !26292 merge hccl_error into master check and log hccl error Created-by: zhaoyu65 Commit-by: zhaoyu Merged-by: ascend-robot Description: <!-- Thanks for sending a pull request! --> **What type of PR is this?** > Uncomment only one /kind <> line, hit enter to put that in a new line, and remove leading whitespaces from that line: > /kind feature **What does this PR do / why do we need it**: **Which issue(s) this PR fixes**: <!-- *Automatically closes linked issue when PR is merged. Usage: Fixes #<issue number>, or Fixes (paste link of issue). --> **Special notes for your reviewers**: HCCL错误打屏，PTA透传HCCL的错误码和错误信息。 代码逻辑：先判断是否有hccl错误，如果有错误，通过aclGetRecentErrMsg接口获取最近的错误并打印到日志里。 See merge request: Ascend/pytorch!262926 个月前
HcclCompile.h
[feat] pg register window memory Co-authored-by: XDaoHong<xudaohong@huawei.com> # message auto-generated for no-merge-commit merge: merge master into master [feat] pg register window memory Created-by: XDaoHong Commit-by: XDaoHong Merged-by: ascend-robot Description: <!-- Thanks for sending a pull request! --> **What type of PR is this?** > Uncomment only one /kind <> line, hit enter to put that in a new line, and remove leading whitespaces from that line: > > /kind bug > /kind task > feature **What does this PR do / why do we need it**: processgroup新增window内存注册和申请接口 **Which issue(s) this PR fixes**: <!-- *Automatically closes linked issue when PR is merged. Usage: Fixes #<issue number>, or Fixes (paste link of issue). --> Fixes # **Special notes for your reviewers**: See merge request: Ascend/pytorch!248477 个月前
Init.cpp
[feat] pg register window memory Co-authored-by: XDaoHong<xudaohong@huawei.com> # message auto-generated for no-merge-commit merge: merge master into master [feat] pg register window memory Created-by: XDaoHong Commit-by: XDaoHong Merged-by: ascend-robot Description: <!-- Thanks for sending a pull request! --> **What type of PR is this?** > Uncomment only one /kind <> line, hit enter to put that in a new line, and remove leading whitespaces from that line: > > /kind bug > /kind task > feature **What does this PR do / why do we need it**: processgroup新增window内存注册和申请接口 **Which issue(s) this PR fixes**: <!-- *Automatically closes linked issue when PR is merged. Usage: Fixes #<issue number>, or Fixes (paste link of issue). --> Fixes # **Special notes for your reviewers**: See merge request: Ascend/pytorch!248477 个月前
Init.h
!6511 Set compile_options -fvisibility=hidden Merge pull request !6511 from 姜怡文/master_TORCH_API 2 年前
LCCLUtils.cpp
!19461 cleanCode Fix Merge pull request !19461 from 叶子凡/master_cleancode_0322 1 年前
LCCLUtils.hpp
!17775 add interface for LCCL backend Merge pull request !17775 from 闫鹏全/master_LCCL 1 年前
ParallelStoreProxy.cpp
!19725 fix paralleltcpstore wait bug Merge pull request !19725 from 王超/v2.7.0_tcpstore 1 年前
ParallelStoreProxy.hpp
!19725 fix paralleltcpstore wait bug Merge pull request !19725 from 王超/v2.7.0_tcpstore 1 年前
ParallelTcpServer.cpp
!22137 Added permission verification and prevents memory overflow Merge pull request !22137 from yuhaiyan/master-dev3 11 个月前
ParallelTcpServer.hpp
!21002 Parallel tcpstore support domain name Merge pull request !21002 from 王超/v2.7.0_storeup 1 年前
ParallelTcpStore.cpp
!21462 TORCH MAIN SYNC : link fmt && clone in c10d/Store Merge pull request !21462 from dilililiwhy/main_sync_20250527 11 个月前
ParallelTcpStore.hpp
!21462 TORCH MAIN SYNC : link fmt && clone in c10d/Store Merge pull request !21462 from dilililiwhy/main_sync_20250527 11 个月前
ProcessGroupHCCL.cpp
make watchdog thread a class Co-authored-by: wanghonghao<wanghonghao9@huawei.com> # message auto-generated for no-merge-commit merge: !27136 merge pg_update into master make watchdog thread a class Created-by: wanghonghao Commit-by: wanghonghao Merged-by: ascend-robot Description: <!-- Thanks for sending a pull request! --> **What type of PR is this?** > Uncomment only one /kind <> line, hit enter to put that in a new line, and remove leading whitespaces from that line: > > /kind bug > /kind task > /kind feature **What does this PR do / why do we need it**: **Which issue(s) this PR fixes**: <!-- *Automatically closes linked issue when PR is merged. Usage: Fixes #<issue number>, or Fixes (paste link of issue). --> Fixes # **Special notes for your reviewers**: See merge request: Ascend/pytorch!271365 个月前
ProcessGroupHCCL.hpp
make watchdog thread a class Co-authored-by: wanghonghao<wanghonghao9@huawei.com> # message auto-generated for no-merge-commit merge: !27136 merge pg_update into master make watchdog thread a class Created-by: wanghonghao Commit-by: wanghonghao Merged-by: ascend-robot Description: <!-- Thanks for sending a pull request! --> **What type of PR is this?** > Uncomment only one /kind <> line, hit enter to put that in a new line, and remove leading whitespaces from that line: > > /kind bug > /kind task > /kind feature **What does this PR do / why do we need it**: **Which issue(s) this PR fixes**: <!-- *Automatically closes linked issue when PR is merged. Usage: Fixes #<issue number>, or Fixes (paste link of issue). --> Fixes # **Special notes for your reviewers**: See merge request: Ascend/pytorch!271365 个月前
ProcessGroupLCCL.cpp
add lock for workspaceallocator Co-authored-by: huangyunlong2022<huangyunlong4@h-partners.com> Co-authored-by: zhaoyu65<nanzhaogang@qq.com> # message auto-generated for no-merge-commit merge: !26720 merge 2.10ts into master add lock for workspaceallocator Created-by: huangyunlong2022 Commit-by: zhaoyu65;huangyunlong2022 Merged-by: ascend-robot Description: <!-- Thanks for sending a pull request! --> **What type of PR is this?** > Uncomment only one /kind <> line, hit enter to put that in a new line, and remove leading whitespaces from that line: > > /kind bug > /kind task > /kind feature **What does this PR do / why do we need it**: 1、通过环境变量控制是否开启每个流一个taskqueue，默认关闭 2、开启后，在enqueue的时候初始化taskqueue，避免流创建的时候初始化太多taskqueue线程，同时为了避免多线程多次初始化进行加锁保护 3、初始化的时候默认选择当前流下发，当前计算算子都是下发到当前流上，对于通信算子下发到通信流上，将其下发的流传到enqueue使用传入的通信流下发 4、取流的时候进行清queue，只清对应流上的queue，可以避免不必要的清queue耗时 5、当前event需要先record后wait，在多taskqueue时为了保序，需要wait在enqueue时确保record已经下发（dequeue阶段通过record数量判断，在event复用场景下，wait之后的record会导致record数量判断失效，导致卡死） 6、event销毁需要确保record,wait已经下发后进行，为了不阻塞，当前采用lazy destroy 7、workspaceallocator进行加锁保护，防止多taskqueue出现竞争问题 **Which issue(s) this PR fixes**: <!-- *Automatically closes linked issue when PR is merged. Usage: Fixes #<issue number>, or Fixes (paste link of issue). --> Fixes # **Special notes for your reviewers**: See merge request: Ascend/pytorch!267206 个月前
ProcessGroupLCCL.hpp
!19093 cleancodeFix Merge pull request !19093 from 叶子凡/master_cleancode_0313 1 年前
StoreClient.cpp
!21002 Parallel tcpstore support domain name Merge pull request !21002 from 王超/v2.7.0_storeup 1 年前
StoreClient.hpp
!21002 Parallel tcpstore support domain name Merge pull request !21002 from 王超/v2.7.0_storeup 1 年前
StoreMessagePacker.cpp
!22240 Remove code that doesn't have much effect Merge pull request !22240 from yuhaiyan/master-dev1 11 个月前
StoreMessagePacker.hpp
!16974 parallel tcpstore support tiered optimizer Merge pull request !16974 from wuxiaotong/cherry-pick-1734348942 1 年前
TraceUtils.h
!22218 JSON serialization for hccl status dump Merge pull request !22218 from huangyunlong/2.8js1 11 个月前
default_comm_hooks.cpp
!18875 CleanCode Fix Merge pull request !18875 from 叶子凡/master_cleancode_0307 1 年前
default_comm_hooks.hpp
!17257 [distributed] Adjust space indentation to 4 spaces. Merge pull request !17257 from 王超/master_cleancode 1 年前
reducer.cpp
fix code issue. Variables must be declared before use Co-authored-by: liangsongwei<liangsongwei@huawei.com> # message auto-generated for no-merge-commit merge: !26365 merge cherry-pick-mr-26289-1762593406721-auto into master fix code issue. Variables must be declared before use Created-by: liangsongwei Commit-by: liangsongwei Merged-by: ascend-robot Description: <!-- Thanks for sending a pull request! --> **What type of PR is this?** > Uncomment only one /kind <> line, hit enter to put that in a new line, and remove leading whitespaces from that line: > > /kind bug > /kind task > /kind feature **What does this PR do / why do we need it**: **Which issue(s) this PR fixes**: <!-- *Automatically closes linked issue when PR is merged. Usage: Fixes #<issue number>, or Fixes (paste link of issue). --> Fixes # **Special notes for your reviewers**: See merge request: Ascend/pytorch!263656 个月前
reducer.hpp
!19093 cleancodeFix Merge pull request !19093 from 叶子凡/master_cleancode_0313 1 年前
reducer_npu.cpp
!18875 CleanCode Fix Merge pull request !18875 from 叶子凡/master_cleancode_0307 1 年前