Fix: cusolverMP BLACS grid mapping order causing errors#6885
Merged
mohanchen merged 3 commits intodeepmodeling:developfrom Jan 23, 2026
Merged
Fix: cusolverMP BLACS grid mapping order causing errors#6885mohanchen merged 3 commits intodeepmodeling:developfrom
mohanchen merged 3 commits intodeepmodeling:developfrom
Conversation
mohanchen
added
Bugs
mohanchen
approved these changes
Jan 23, 2026
Collaborator
|
Excellent team work from ABACUS community! |
Collaborator
|
It seems to make sense, but I'm a little confused as to why it worked before. I remember 4-card and 8-card setups are tested. |
Collaborator
|
And, great job. |
Collaborator
Since #4207 , char order = 'R'. Maybe char order = 'C' before this PR? |
Flying-dragon-boxing
pushed a commit
to Flying-dragon-boxing/abacus-develop
that referenced
this pull request
Jan 26, 2026
This file contains hidden or 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
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
4 participants
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
Description
This PR fixes a critical bug in the cusolverMP solver that caused crashes when the number of GPUs is not a prime number.
Problem
When using the cusolverMP eigensolver, the code would fail with specific GPU counts due to a mismatch between the BLACS grid mapping order used in ABACUS and the grid mapping order specified for cusolverMP. This mismatch caused the process grid topology to be inconsistent between BLACS and cusolverMP, leading to runtime errors.
Solution
Changed the cusolverMP device grid initialization from
CUSOLVERMP_GRID_MAPPING_COL_MAJORtoCUSOLVERMP_GRID_MAPPING_ROW_MAJORto match the BLACS grid initialization order.Acknowledgments
Thanks to @Entropy-Enthalpy、@Stardust0831 and @ZhouXY-PKU for testing and helping identify this bug!