Ifpack2: BlockTriDi fix for large blocks #13792
Merged
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
In ExtractAndFactorize kernels that use scratch, fall back to level 1 when there isn't sufficient level 0. Signed-off-by: Brian Kelley <[email protected]>
Use max block size of 32 and a small grid. Test standard BTD, BTD with Schur line splitting, and Block Jacobi. Signed-off-by: Brian Kelley <[email protected]>
brian-kelley
added
type: bug
|
Status Flag 'Pre-Test Inspection' - Auto Inspected - Inspection is Not Necessary for this Pull Request. |
|
Status Flag 'Pull Request AutoTester' - Testing Jenkins Projects: Pull Request Auto Testing STARTING (click to expand)Build InformationTest Name: PR_gcc-openmpi-openmp
Jenkins Parameters
Build InformationTest Name: PR_gcc
Jenkins Parameters
Build InformationTest Name: PR_gcc-openmpi_debug
Jenkins Parameters
Build InformationTest Name: PR_clang
Jenkins Parameters
Build InformationTest Name: PR_cuda
Jenkins Parameters
Build InformationTest Name: PR_intel
Jenkins Parameters
Build InformationTest Name: PR_cuda-uvm
Jenkins Parameters
Using Repos:
Pull Request Author: brian-kelley |
|
lgtm, thanks! I'll give it a try today. |
|
Status Flag 'Pull Request AutoTester' - Jenkins Testing: all Jobs PASSED Pull Request Auto Testing has PASSED (click to expand)Build InformationTest Name: PR_gcc-openmpi-openmp
Jenkins Parameters
Build InformationTest Name: PR_gcc
Jenkins Parameters
Build InformationTest Name: PR_gcc-openmpi_debug
Jenkins Parameters
Build InformationTest Name: PR_clang
Jenkins Parameters
Build InformationTest Name: PR_cuda
Jenkins Parameters
Build InformationTest Name: PR_intel
Jenkins Parameters
Build InformationTest Name: PR_cuda-uvm
Jenkins Parameters
|
|
Status Flag 'Pre-Merge Inspection' - - This Pull Request Requires Inspection... The code must be inspected by a member of the Team before Testing/Merging |
|
All Jobs Finished; status = PASSED, However Inspection must be performed before merge can occur... |
jewatkins
approved these changes
Feb 7, 2025
|
Status Flag 'Pre-Merge Inspection' - SUCCESS: The last commit to this Pull Request has been INSPECTED AND APPROVED by [ jewatkins ]! |
|
Status Flag 'Pull Request AutoTester' - AutoMerge IS ENABLED, but the Label AT: AUTOMERGE is not set. Either set Label AT: AUTOMERGE or manually merge the PR... |
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.
In BlockTriDi's ExtractAndFactorize kernels that use scratch, fall back to level 1 when there isn't sufficient level 0.
@trilinos/ifpack2
Motivation
Ifpack2's block TriDi and block Jacobi should work for block sizes to and including 32. For relatively large block sizes like 27, many GPUs didn't have enough shared memory per block for the numeric setup (extract+factorize). When there isn't enough shared, this PR falls back to level 1 scratch.
The level 0 and level 1 cases are instantiated separately to make sure there is no performance hit to the original level 0 case.
Related Issues
No issue for this, was reported by email
Stakeholder Feedback
Issue reported directly by @jewatkins for SPARC.
Testing
Added testing with the maximum block size of 32. Verified that the level 1 scratch code path is taken on nvidia and amd gpus in this test.