You are viewing a plain text version of this content. The canonical link for it is here.
Posted to issues@celeborn.apache.org by "waitinfuture (via GitHub)" <gi...@apache.org> on 2023/02/25 15:31:41 UTC
[GitHub] [incubator-celeborn] waitinfuture opened a new pull request, #1278: [CELEBORN-301] Refactor PartitionLocationInfo to use ConcurrentHashMap
waitinfuture opened a new pull request, #1278:
URL: https://github.com/apache/incubator-celeborn/pull/1278
<!--
Thanks for sending a pull request! Here are some tips for you:
- Make sure the PR title start w/ a JIRA ticket, e.g. '[CELEBORN-XXXX] Your PR title ...'.
- Be sure to keep the PR description updated to reflect all changes.
- Please write your PR title to summarize what this PR proposes.
- If possible, provide a concise example to reproduce the issue for a faster review.
-->
### What changes were proposed in this pull request?
Refactor PartitionLocationInfo, replace HashMap with ConcurrentHashMap and removed synchronized blocks. Also change the core data structure from ```(shuffleKey -> (partitionId -> List[PartitionLocation]))``` to ```(shuffleKey -> (encodedUniqueId -> partitionLocation))```, where ```encodedUniqueId``` is ```partitionId.toLong << 32 | epoch```
### Why are the changes needed?
For very large test case, say 300T terasort with 30w * 15w parallelism with sort pusher, the synchronization lock becomes the bottleneck. Tried with ReentrantReadWriteLock, task duration decreases from 7min to 5.6min. The PR further uses ConcurrentHashMap to avoid lock.
I tested the following with sort pusher
```
spark.sparkContext.parallelize(1 to 3000).repartition(2000).flatMap( _ => (1 to 15000000).iterator.map(num => num)).repartition(150000).count
```
Before this PR the shuffle write stage takes 6.7 min
After this PR the shuffle write stage takes 5.3 min
### Does this PR introduce _any_ user-facing change?
No.
### How was this patch tested?
Integration test.
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: issues-unsubscribe@celeborn.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
[GitHub] [incubator-celeborn] codecov[bot] commented on pull request #1278: [CELEBORN-301] Refactor PartitionLocationInfo to use ConcurrentHashMap
Posted by "codecov[bot] (via GitHub)" <gi...@apache.org>.
codecov[bot] commented on PR #1278:
URL: https://github.com/apache/incubator-celeborn/pull/1278#issuecomment-1445144858
# [Codecov](https://codecov.io/gh/apache/incubator-celeborn/pull/1278?src=pr&el=h1&utm_medium=referral&utm_source=github&utm_content=comment&utm_campaign=pr+comments&utm_term=The+Apache+Software+Foundation) Report
> Merging [#1278](https://codecov.io/gh/apache/incubator-celeborn/pull/1278?src=pr&el=desc&utm_medium=referral&utm_source=github&utm_content=comment&utm_campaign=pr+comments&utm_term=The+Apache+Software+Foundation) (59bfdbf) into [main](https://codecov.io/gh/apache/incubator-celeborn/commit/f0b9236ff21b52610148444b833e97300ab422db?el=desc&utm_medium=referral&utm_source=github&utm_content=comment&utm_campaign=pr+comments&utm_term=The+Apache+Software+Foundation) (f0b9236) will **increase** coverage by `0.02%`.
> The diff coverage is `0.00%`.
```diff
@@ Coverage Diff @@
## main #1278 +/- ##
============================================
+ Coverage 27.02% 27.04% +0.02%
+ Complexity 814 813 -1
============================================
Files 215 215
Lines 18424 18405 -19
Branches 1983 1981 -2
============================================
- Hits 4978 4976 -2
+ Misses 13123 13104 -19
- Partials 323 325 +2
```
| [Impacted Files](https://codecov.io/gh/apache/incubator-celeborn/pull/1278?src=pr&el=tree&utm_medium=referral&utm_source=github&utm_content=comment&utm_campaign=pr+comments&utm_term=The+Apache+Software+Foundation) | Coverage Δ | |
|---|---|---|
| [...born/common/meta/WorkerPartitionLocationInfo.scala](https://codecov.io/gh/apache/incubator-celeborn/pull/1278?src=pr&el=tree&utm_medium=referral&utm_source=github&utm_content=comment&utm_campaign=pr+comments&utm_term=The+Apache+Software+Foundation#diff-Y29tbW9uL3NyYy9tYWluL3NjYWxhL29yZy9hcGFjaGUvY2VsZWJvcm4vY29tbW9uL21ldGEvV29ya2VyUGFydGl0aW9uTG9jYXRpb25JbmZvLnNjYWxh) | `0.00% <0.00%> (ø)` | |
| [...apache/celeborn/service/deploy/worker/Worker.scala](https://codecov.io/gh/apache/incubator-celeborn/pull/1278?src=pr&el=tree&utm_medium=referral&utm_source=github&utm_content=comment&utm_campaign=pr+comments&utm_term=The+Apache+Software+Foundation#diff-d29ya2VyL3NyYy9tYWluL3NjYWxhL29yZy9hcGFjaGUvY2VsZWJvcm4vc2VydmljZS9kZXBsb3kvd29ya2VyL1dvcmtlci5zY2FsYQ==) | `0.00% <0.00%> (ø)` | |
| [...oy/worker/congestcontrol/CongestionController.java](https://codecov.io/gh/apache/incubator-celeborn/pull/1278?src=pr&el=tree&utm_medium=referral&utm_source=github&utm_content=comment&utm_campaign=pr+comments&utm_term=The+Apache+Software+Foundation#diff-d29ya2VyL3NyYy9tYWluL2phdmEvb3JnL2FwYWNoZS9jZWxlYm9ybi9zZXJ2aWNlL2RlcGxveS93b3JrZXIvY29uZ2VzdGNvbnRyb2wvQ29uZ2VzdGlvbkNvbnRyb2xsZXIuamF2YQ==) | `76.86% <0.00%> (-0.92%)` | :arrow_down: |
| [...cala/org/apache/celeborn/common/CelebornConf.scala](https://codecov.io/gh/apache/incubator-celeborn/pull/1278?src=pr&el=tree&utm_medium=referral&utm_source=github&utm_content=comment&utm_campaign=pr+comments&utm_term=The+Apache+Software+Foundation#diff-Y29tbW9uL3NyYy9tYWluL3NjYWxhL29yZy9hcGFjaGUvY2VsZWJvcm4vY29tbW9uL0NlbGVib3JuQ29uZi5zY2FsYQ==) | `81.14% <0.00%> (-0.05%)` | :arrow_down: |
:mega: We’re building smart automated test selection to slash your CI/CD build times. [Learn more](https://about.codecov.io/iterative-testing/?utm_medium=referral&utm_source=github&utm_content=comment&utm_campaign=pr+comments&utm_term=The+Apache+Software+Foundation)
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: issues-unsubscribe@celeborn.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
[GitHub] [incubator-celeborn] waitinfuture commented on pull request #1278: [CELEBORN-301] Refactor PartitionLocationInfo to use ConcurrentHashMap
Posted by "waitinfuture (via GitHub)" <gi...@apache.org>.
waitinfuture commented on PR #1278:
URL: https://github.com/apache/incubator-celeborn/pull/1278#issuecomment-1445271287
Tested regression and results are good.
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: issues-unsubscribe@celeborn.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
[GitHub] [incubator-celeborn] waitinfuture merged pull request #1278: [CELEBORN-301] Refactor PartitionLocationInfo to use ConcurrentHashMap
Posted by "waitinfuture (via GitHub)" <gi...@apache.org>.
waitinfuture merged PR #1278:
URL: https://github.com/apache/incubator-celeborn/pull/1278
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: issues-unsubscribe@celeborn.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org