feat: Support prefetch in index lookup join #12611
Open
+410
−131
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
facebook-github-bot
added
the
CLA Signed
|
This pull request was exported from Phabricator. Differential Revision: D70909786 |
✅ Deploy Preview for meta-velox canceled.
|
xiaoxmeng
added a commit
to xiaoxmeng/velox
that referenced
this pull request
Mar 12, 2025
Summary: This PR adds prefetch for index join with query config. The index join operator can prefetch up to the configured prefetch limit to enable: (1) parallel prefetches at backend for parallel execution (1-1) in case of multiple backend shards or (1-2) enable backend to batch multiple requests to improve throughput; (2) pipeline the table scan and index lookup execution in the same driver pipeline. The table scan is sync executed while index lookup is async. This achieve pipelining without relying exchange which might cause non-deterministic execution and Meta internal use case needs deterministic execution for checkpointing. With Meta internal testing, this can achieve 2x throughput improvement (measured in rows per second) with 33% memory overhead with up to 4 batches prefetch. The follow is to add memory based prefetch throttling to integrate with Meta internal ML use case and memory pool wiring to ease performance (memory overhead) analysis Reviewed By: wenqiwooo Differential Revision: D70909786
xiaoxmeng
force-pushed
the
export-D70909786
branch
from
a6ae1ec to
8798a2f
Compare
|
This pull request was exported from Phabricator. Differential Revision: D70909786 |
xiaoxmeng
added a commit
to xiaoxmeng/velox
that referenced
this pull request
Mar 12, 2025
Summary: This PR adds prefetch for index join with query config. The index join operator can prefetch up to the configured prefetch limit to enable: (1) parallel prefetches at backend for parallel execution (1-1) in case of multiple backend shards or (1-2) enable backend to batch multiple requests to improve throughput; (2) pipeline the table scan and index lookup execution in the same driver pipeline. The table scan is sync executed while index lookup is async. This achieve pipelining without relying exchange which might cause non-deterministic execution and Meta internal use case needs deterministic execution for checkpointing. With Meta internal testing, this can achieve 2x throughput improvement (measured in rows per second) with 33% memory overhead with up to 4 batches prefetch. The follow is to add memory based prefetch throttling to integrate with Meta internal ML use case and memory pool wiring to ease performance (memory overhead) analysis Reviewed By: wenqiwooo Differential Revision: D70909786
Summary: This PR adds prefetch for index join with query config. The index join operator can prefetch up to the configured prefetch limit to enable: (1) parallel prefetches at backend for parallel execution (1-1) in case of multiple backend shards or (1-2) enable backend to batch multiple requests to improve throughput; (2) pipeline the table scan and index lookup execution in the same driver pipeline. The table scan is sync executed while index lookup is async. This achieve pipelining without relying exchange which might cause non-deterministic execution and Meta internal use case needs deterministic execution for checkpointing. With Meta internal testing, this can achieve 2x throughput improvement (measured in rows per second) with 33% memory overhead with up to 4 batches prefetch. The follow is to add memory based prefetch throttling to integrate with Meta internal ML use case and memory pool wiring to ease performance (memory overhead) analysis Reviewed By: wenqiwooo Differential Revision: D70909786
xiaoxmeng
force-pushed
the
export-D70909786
branch
from
8798a2f to
71a1913
Compare
|
This pull request was exported from Phabricator. Differential Revision: D70909786 |
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.
Summary:
This PR adds prefetch for index join with query config. The index join operator can prefetch up to the
configured prefetch limit to enable: (1) parallel prefetches at backend for parallel execution (1-1) in case
of multiple backend shards or (1-2) enable backend to batch multiple requests to improve throughput;
(2) pipeline the table scan and index lookup execution in the same driver pipeline. The table scan is sync
executed while index lookup is async. This achieve pipelining without relying exchange which might cause
non-deterministic execution and Meta internal use case needs deterministic execution for checkpointing.
With Meta internal testing, this can achieve 2x throughput improvement (measured in rows per second)
with 33% memory overhead with up to 4 batches prefetch.
The follow is to add memory based prefetch throttling to integrate with Meta internal
ML use case and memory pool wiring to ease performance (memory overhead) analysis
Differential Revision: D70909786