[Improvement](shuffle) Use a knob to decide whether a serial exchange should be used #43199
Conversation
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
|
Thank you for your contribution to Apache Doris. Since 2024-03-18, the Document has been moved to doris-website. |
|
run buildall |
TPC-H: Total hot run time: 41784 ms |
|
run buildall |
TPC-H: Total hot run time: 41704 ms |
TPC-DS: Total hot run time: 192400 ms |
ClickBench: Total hot run time: 31.8 s |
|
run buildall |
TPC-H: Total hot run time: 41205 ms |
TPC-DS: Total hot run time: 194742 ms |
ClickBench: Total hot run time: 32.17 s |
|
run buildall |
TPC-H: Total hot run time: 41838 ms |
TPC-DS: Total hot run time: 193966 ms |
ClickBench: Total hot run time: 32.73 s |
924060929
approved these changes
Nov 6, 2024
github-actions
bot
added
the
approved
|
PR approved by at least one committer and no changes requested. |
|
PR approved by anyone and no changes requested. |
BiteTheDDDDt
approved these changes
Nov 6, 2024
Gabriel39
added a commit
to Gabriel39/incubator-doris
that referenced
this pull request
Nov 6, 2024
… should be used (apache#43199) A new knob to control a exchange node should be serial or not. For example, a partitioned hash join should be executed like below: ``` ┌────────────────────────────┐ ┌────────────────────────────┐ │ │ │ │ │Exchange(HASH PARTITIONED N)│ │Exchange(HASH PARTITIONED N)│ │ │ │ │ └────────────────────────────┴─────────┬────────┴────────────────────────────┘ │ │ │ │ │ │ ┌──────▼──────┐ │ │ │ HASH JOIN │ │ │ └─────────────┘ ``` After turning on this knob, the real plan should be: ``` ┌──────────────────────────────┐ ┌──────────────────────────────┐ │ │ │ │ │ Exchange (HASH PARTITIONED 1)│ │ Exchange (HASH PARTITIONED 1)│ │ │ │ │ └────────────┬─────────────────┘ └────────────┬─────────────────┘ │ │ │ │ │ │ │ │ │ │ ┌──────────────▼─────────────────────┐ ┌──────────────▼─────────────────────┐ │ │ │ │ │ Local Exchange(HASH PARTITIONED N)│ │ Local Exchange(HASH PARTITIONED N)│ │ 1 -> N │ │ 1 -> N │ └────────────────────────────────────┴─────────┬────────┴────────────────────────────────────┴ │ │ │ │ │ │ ┌──────▼──────┐ │ │ │ HASH JOIN │ │ │ └─────────────┘ ``` For large cluster, X (mappers) * Y (reducers) rpc channels can be reduced to X (mappers) * Z (BEs).
Gabriel39
added
dev/3.0.3-merged
and removed
dev/3.0.x
cherry-pick-conflict-in-3.0
labels
yiguolei
added a commit
to yiguolei/incubator-doris
that referenced
this pull request
Nov 16, 2024
…exchange should be used (apache#43199)" This reverts commit b2e1c08.
This was referenced Nov 16, 2024
yiguolei
added a commit
to yiguolei/incubator-doris
that referenced
this pull request
Nov 16, 2024
…exchange should be used (apache#43199)" This reverts commit b2e1c08.
16 tasks
Gabriel39
added a commit
that referenced
this pull request
Nov 28, 2024
#44676) … should be used This improvement was completed in #43199 and reverted by #44075 due to performance fallback. After fixing it, this improvement is re-submited. A new knob to control a exchange node should be serial or not. For example, a partitioned hash join should be executed like below: ``` ┌────────────────────────────┐ ┌────────────────────────────┐ │ │ │ │ │Exchange(HASH PARTITIONED N)│ │Exchange(HASH PARTITIONED N)│ │ │ │ │ └────────────────────────────┴─────────┬────────┴────────────────────────────┘ │ │ │ │ │ │ ┌──────▼──────┐ │ │ │ HASH JOIN │ │ │ └─────────────┘ ``` After turning on this knob, the real plan should be: ``` ┌──────────────────────────────┐ ┌──────────────────────────────┐ │ │ │ │ │ Exchange (HASH PARTITIONED 1)│ │ Exchange (HASH PARTITIONED 1)│ │ │ │ │ └────────────┬─────────────────┘ └────────────┬─────────────────┘ │ │ │ │ │ │ │ │ │ │ ┌──────────────▼─────────────────────┐ ┌──────────────▼─────────────────────┐ │ │ │ │ │ Local Exchange(HASH PARTITIONED N)│ │ Local Exchange(HASH PARTITIONED N)│ │ 1 -> N │ │ 1 -> N │ └────────────────────────────────────┴─────────┬────────┴────────────────────────────────────┴ │ │ │ │ │ │ ┌──────▼──────┐ │ │ │ HASH JOIN │ │ │ └─────────────┘ ``` For large cluster, X (mappers) * Y (reducers) rpc channels can be reduced to X (mappers) * Z (BEs).
github-actions bot
pushed a commit
that referenced
this pull request
Nov 28, 2024
#44676) … should be used This improvement was completed in #43199 and reverted by #44075 due to performance fallback. After fixing it, this improvement is re-submited. A new knob to control a exchange node should be serial or not. For example, a partitioned hash join should be executed like below: ``` ┌────────────────────────────┐ ┌────────────────────────────┐ │ │ │ │ │Exchange(HASH PARTITIONED N)│ │Exchange(HASH PARTITIONED N)│ │ │ │ │ └────────────────────────────┴─────────┬────────┴────────────────────────────┘ │ │ │ │ │ │ ┌──────▼──────┐ │ │ │ HASH JOIN │ │ │ └─────────────┘ ``` After turning on this knob, the real plan should be: ``` ┌──────────────────────────────┐ ┌──────────────────────────────┐ │ │ │ │ │ Exchange (HASH PARTITIONED 1)│ │ Exchange (HASH PARTITIONED 1)│ │ │ │ │ └────────────┬─────────────────┘ └────────────┬─────────────────┘ │ │ │ │ │ │ │ │ │ │ ┌──────────────▼─────────────────────┐ ┌──────────────▼─────────────────────┐ │ │ │ │ │ Local Exchange(HASH PARTITIONED N)│ │ Local Exchange(HASH PARTITIONED N)│ │ 1 -> N │ │ 1 -> N │ └────────────────────────────────────┴─────────┬────────┴────────────────────────────────────┴ │ │ │ │ │ │ ┌──────▼──────┐ │ │ │ HASH JOIN │ │ │ └─────────────┘ ``` For large cluster, X (mappers) * Y (reducers) rpc channels can be reduced to X (mappers) * Z (BEs).
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
What problem does this PR solve?
A new knob to control a exchange node should be serial or not.
For example, a partitioned hash join should be executed like below:
After turning on this knob, the real plan should be:
For large cluster, X (mappers) * Y (reducers) rpc channels can be reduced to X (mappers) * Z (BEs).
Check List (For Committer)
Test
Behavior changed:
Does this need documentation?
Release note
None
Check List (For Reviewer who merge this PR)