v4.0.x: Update the PML selection/check logic to avoid direct modex "storms" #8408
Conversation
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
With this patch the best PML is selected earlier, before finalizing the others PML. This provides a simpler mechanism to intercept and highjack the PML (as done in the monitoring PML) Signed-off-by: George Bosilca <[email protected]> (cherry picked from commit 668aa15) (cherry picked from commit 65fbffa)
For direct modex, all procs publish the selected pml module and then at add_procs pml module for each proc is checked against every other proc in the add_proc call. For full modex, there is no change in functionality. Only Rank0 publishes its selected pml, all other procs in the add_proc call check their selected pml against Rank0. If pml's do not match, throw error and exit. Signed-off-by: Dipti Kothari <[email protected]> (cherry picked from commit 5418cc5) (cherry picked from commit 5de4423)
Signed-off-by: Ralph Castain <[email protected]> (cherry picked from commit 56eb572)
|
I have moved this to "draft" status because I believe we need to revisit the PML selection check scheme. Please see #8404 (comment) for an explanation |
|
After conversation, this is good to go! |
|
bot:aws:retest |
bwbarrett
approved these changes
Jan 21, 2021
Closed
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Outdated suggestions cannot be applied.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.
As currently written in this release branch, the PML selection "check" logic doesn't guarantee that the caller's PML choice will be checked against only that from MPI_COMM_WORLD rank=0 when a full modex has been performed. This can lead to every process calling "dmodex" to obtain the PML selection of every other process in the job, causing major delay in wireup on first call to communicate.
These cherry-picks contain the updates developed/committed to master after the code in this release branch was brought over to it. One additional cherry-pick was required to cleanly port the code.