-
Notifications
You must be signed in to change notification settings - Fork 936
v4.0.x: Correctly process 0 slots with -host option. #9221
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
Conversation
a135574 to
64a7aad
Compare
orte/mca/ess/hnp/ess_hnp_module.c
Outdated
| node->name = strdup(orte_process_info.nodename); | ||
| node->index = ORTE_PROC_MY_NAME->vpid; | ||
| node->slots = 0; | ||
| ORTE_FLAG_SET(node, ORTE_NODE_FLAG_SLOTS_GIVEN); |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
I don't think you really want to do this - setting that flag means we won't sense the number of available slots by counting the number of cores. So basically you are "locking" mpirun itself to 0 slots. IIRC, the object constructor sets slots to zero - otherwise, this initializer is fine.
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Thanks - I think you're right and these lines aren't needed for the patch. Verifying..
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Updated.
64a7aad to
21418ee
Compare
The following command: mpirun -host hostA:0,hostB:6 ./x was launching: (num_cores on hostA) + (6 hostB ranks) instead of only the expected 6 ranks on hostB. This works correctly with -hostfile with "hostA slots=0". This patch just makes the behaviors consistent. Co-authored-by: Austen Lauria <[email protected]> Signed-off-by: Austen Lauria <[email protected]>
21418ee to
7b6f478
Compare
|
bot:aws:retest more java. |
|
FWIW: I ported this over to PRRTE |
|
@rhc54 thanks! I can port the others as well if you don't have the time to do it. |
|
Does #9222 apply as well? |
No - PRRTE doesn't modify |
|
Does PR #9234 need to be updated? |
The following command:
mpirun -host hostA:0,hostB:6 ./x
was launching: (num_cores on hostA) + (6 hostB ranks)
instead of only the expected 6 ranks on hostB.
This works correctly with -hostfile with "hostA slots=0".
This patch just makes the behaviors consistent.
bot:notacherrypick
Co-authored-by: Austen Lauria [email protected]
Signed-off-by: Austen Lauria [email protected]