The direct modex operation is slow, especially at scale for even modestly-connected applications. Likewise, blocking in MPI_Init while we wait for a full modex to complete takes too long. However, as George pointed out, there is a middle ground here. We could kickoff the modex operation in the background, and then trap any modex_recv's until the modex completes and the data is delivered. For most non-benchmark apps, this may prove to be the best of the available options as they are likely to perform other (non-communicating) setup operations after MPI_Init, and so there is a reasonable chance that the modex will actually be done before the first modex_recv gets called.
Once we get instant-on-enabled hardware, this won't be necessary. Clearly, zero time will always out-perform the time spent doing a modex. However, this provides a decent compromise in the interim.
This PR changes the default settings of a few relevant params to make "background modex" the default behavior:
* pmix_base_async_modex -> defaults to true
* pmix_base_collect_data -> continues to default to true (no change)
* async_mpi_init - defaults to true. Note that the prior code attempted to base the default setting of this value on the setting of pmix_base_async_modex. Unfortunately, the pmix value isn't set prior to setting async_mpi_init, and so that attempt failed to accomplish anything.
The logic in MPI_Init is:
* if async_modex AND collect_data are set, AND we have a non-blocking fence available, then we execute the background modex operation
* if async_modex is set, but collect_data is false, then we simply skip the modex entirely - no fence is performed
* if async_modex is not set, then we block until the fence completes (regardless of collecting data or not)
* if we do NOT have a non-blocking fence (e.g., we are not using PMIx), then we always perform the full blocking modex operation.
* if we do perform the background modex, and the user requested the barrier be performed at the end of MPI_Init, then we check to see if the modex has completed when we reach that point. If it has, then we execute the barrier. However, if the modex has NOT completed, then we block until the modex does complete and skip the extra barrier. So we never perform two barriers in that case.
HTH
Ralph
Signed-off-by: Ralph Castain <rhc@open-mpi.org>
Set the daemons' state to "running" and mark them as "alive" by default when constructing the nidmap
Get the DVM running again
Fix direct modex by eliminating race condition caused by releasing data while sending it
Up the size limit before compressing
Signed-off-by: Ralph Castain <rhc@open-mpi.org>
Cleanup a race condition segfault during finalize by ensuring the PMIx progress thread is stopped prior to starting to tear down the messaging components
Signed-off-by: Ralph Castain <rhc@open-mpi.org>
Do not use opal_output_verbose inside O(n) loops. This was causing us
to make O(n) calls to snprintf which was greatly slowing launch at
scale.
Signed-off-by: Nathan Hjelm <hjelmn@lanl.gov>
It's possible that we can have zlib.h but still not have zlib support.
Use the correct macro to protect the usage of calling zlib functions.
This fixes 32-bit MTT builds at Cisco (e.g.,
https://mtt.open-mpi.org/index.php?do_redir=2389).
Submitted upstream to PMIX: https://github.com/pmix/master/pull/290
Signed-off-by: Jeff Squyres <jsquyres@cisco.com>
PMIx_server_register_nspace() is an asynchronous operation, so
the pmix glue wait for it completes before returning.
Signed-off-by: Gilles Gouaillardet <gilles@rist.or.jp>
the argc field from the opal_pmix_app_t struct was removed,
so adjust the pmix/ext11 glue accordingly.
Signed-off-by: Gilles Gouaillardet <gilles@rist.or.jp>
