New feature: intra-node write aggregation #156
Merged
Conversation
This file contains 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
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Outdated suggestions cannot be applied.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.
This new feature is designed to reduce the possible congestion of all-to-many
personalized communication inside of MPI-IO when the number of allocated MPI
processes per compute node is large, for example larger than 64 per node, and
the non-contiguous requests per process spread over a wide range of file space.
Such congestion have been observed when running the E3SM-IO F case using 21600
MPI processes on DOE parallel computers, Perlmutter at NRSC, Frontier at OLCF,
and Aurora at ALCF.
When this feature enabled, a subset of MPI processes running on the same
compute node will be selected to be the aggregator. The non-aggregators send
their write requests to the assigned aggregators and then only the aggregators
make the MPI file write calls to write to the file.
To enable this new feature, a PnetCDF I/O hint 'nc_num_aggrs_per_node' can set
by the users. When setting its value to a positive integral value, PnetCDF will
select this number of processes per node to be the aggregators. Setting it to 0
disables the aggregation (the default mode.)