Skip to content

Conversation

@rhc54
Copy link
Contributor

@rhc54 rhc54 commented Jan 21, 2017

Rework the threading in oob/tcp so that daemons (including mpirun) use multiple progress threads to get messages out to their children, and so that the oob/base uses a separate one to setup sends. This allows the daemon cmd processor to execute in parallel with relay of messages, which significantly reduces launch times at scale

Improve packing efficiency by raising the initial buffer size and modifying the extension code. Flag if a job map has had its nodes added so we don't have to loop repeatedly to check it.

Signed-off-by: Ralph Castain [email protected]

Ralph Castain added 2 commits January 21, 2017 13:26
…e multiple progress threads to get messages out to their children, and so that the oob/base uses a separate one to setup sends. This allows the daemon cmd processor to execute in parallel with relay of messages, which significantly reduces launch times at scale

Signed-off-by: Ralph Castain <[email protected]>
…ifying the extension code. Flag if a job map has had its nodes added so we don't have to loop repeatedly to check it.

Signed-off-by: Ralph Castain <[email protected]>
@rhc54
Copy link
Contributor Author

rhc54 commented Jan 21, 2017

bot:mellanox:retest

…rt them while retaining the larger initial and threshold sizes

Signed-off-by: Ralph Castain <[email protected]>
@rhc54 rhc54 merged commit 164fc64 into open-mpi:master Jan 21, 2017
@rhc54 rhc54 deleted the topic/oob3 branch January 21, 2017 23:46
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

1 participant