Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

[Gluten-1.2] Port #10782 to Branch-1.2 for Release excessively reserved memory in HashBuild even if non-reclaimable (#10782) #496

Merged
merged 1 commit into from
Sep 18, 2024

Conversation

zsmj2017
Copy link

Summary:
When hash build is under table building stage, we reserve excessive amount of memory to account for the worst case scenario duplicate rows for NextRowVector (i.e. we assume every row in build table has duplicates, which in most cases is not true). Other than let the query fail because the current stage is unreclaimable, we can perform a desperate try to release the unused reserved memory, giving the query a chance to succeed.

Pull Request resolved: facebookincubator#10782

Reviewed By: xiaoxmeng

Differential Revision: D61510068

Pulled By: tanjialiang

fbshipit-source-id: 1d62804d22ab11d08080e7cf872da0656cbd1010 (cherry picked from commit 55888da)

…ble (facebookincubator#10782)

Summary:
When hash build is under table building stage, we reserve excessive amount of memory to account for the worst case scenario duplicate rows for NextRowVector (i.e. we assume every row in build table has duplicates, which in most cases is not true). Other than let the query fail because the current stage is unreclaimable, we can perform a desperate try to release the unused reserved memory, giving the query a chance to succeed.

Pull Request resolved: facebookincubator#10782

Reviewed By: xiaoxmeng

Differential Revision: D61510068

Pulled By: tanjialiang

fbshipit-source-id: 1d62804d22ab11d08080e7cf872da0656cbd1010
(cherry picked from commit 55888da)
@kecookier
Copy link

@zhouyuan Can you help review this?

@weiting-chen
Copy link
Collaborator

The PR has run pass 333 velox tests with fail 1 test related to ParquetTableScanTest.timestampFilter test report empty parquet file.
The issue is related to facebookincubator#4680 and not related to this PR.
Approval to merge it.

@weiting-chen weiting-chen merged commit eefc229 into oap-project:branch-1.2 Sep 18, 2024
@zsmj2017 zsmj2017 deleted the branch-1.2-10782 branch September 19, 2024 03:12
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
None yet
Projects
None yet
Development

Successfully merging this pull request may close these issues.

5 participants