Fix MPI.Reduce! operations in some test cases when running on ARM chips#446
Open
johnomotani wants to merge 5 commits into
Open
Fix MPI.Reduce! operations in some test cases when running on ARM chips#446johnomotani wants to merge 5 commits into
johnomotani wants to merge 5 commits into
Conversation
Avoids errors when running on ARM chips.
No longer needed, and causes MPI.jl to be restricted to an old version, when we need a newer one for MPI.@RegisterOp.
This is needed for the docs build, as Something is preventing MPI.jl from using the latest version in that workflow. Updating Julia apparently updates whatever was blocking MPI.jl from updating.
johnomotani
force-pushed
the
fix-macos-MPI.Reduce-error
branch
from
b45e00f to
214abbc
Compare
Update to Julia-1.12 has increased rounding errors slightly in CI.
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
1 participant
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
MPI.jl has some restrictions on what operations can be used for MPI reductions when running on ARM (or other non-x86) chips (https://juliaparallel.org/MPI.jl/stable/knownissues/#Custom-reduction-operators). In some test cases, we use Float128 from Quadmath.jl, which triggers a 'custom' reduction operation, which is unsupported on ARM. The workaround is to 'register' an MPI.Op for addition of Float128, using
MPI.@RegisterOp.