[cherry-pick][Distributed] fix recreate nccl comm bug (#73625)

[tianhaodongbd]

PR Category

Distributed Strategy

PR Types

Bug fixes

Description

Pcard-90602
cherry-pick pr: #73625
在开pp的场景下recreate nccl comm存在hang的问题， 原因是同一个通信组内 tcp通信时unique_key的获取是无序的，导致相互等待。当前通过有序map代替无序map来修复这个问题。

[paddle-bot]

你的PR提交成功，感谢你对开源项目的贡献!
请关注后续CI自动化测试结果，详情请参考Paddle-CI手册。
Your PR has been submitted. Thanks for your contribution!
Please wait for the result of CI firstly. See Paddle CI Manual for details.

[codecov-commenter]

Codecov Report

Attention: Patch coverage is 0% with 2 lines in your changes missing coverage. Please review.

Please upload report for BASE (develop@d5b8b45). Learn more about missing BASE report.

Files with missing lines	Patch %	Lines
...addle/phi/core/distributed/comm_context_manager.cc	0.00%	2 Missing ⚠️

❌ Your patch status has failed because the patch coverage (0.00%) is below the target coverage (90.00%). You can increase the patch coverage or adjust the target coverage.

Additional details and impacted files

@@            Coverage Diff             @@
##             develop   #74168   +/-   ##
==========================================
  Coverage           ?    0.00%           
==========================================
  Files              ?        1           
  Lines              ?        2           
  Branches           ?        0           
==========================================
  Hits               ?        0           
  Misses             ?        2           
  Partials           ?        0           

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

🚀 New features to boost your workflow:

• ❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.

[swgu98]

/re-run cpu

[tianhaodongbd]

/re-run all-failed

[paddle-ci-bot]

Sorry to inform you that 727bd39's CIs have passed for more than 7 days. To prevent PR conflicts, you need to re-run all CIs manually.

[tianhaodongbd]

/re-run all-failed

[gongweibao]

LGTM

[tianhaodongbd]

/re-run Approval

[tianhaodongbd]

/re-run all-failed

[SylarTiaNII]

LGTM