deactivated by default. It is activated by setting either of the
following two MCA parameters to values greater than 0:
* coll_sync_barrier_before
* coll_sync_barrier_after
If !_before is >0, then the sync coll collective will insert itself
before the underlying collective operations and invoke a barrier
before every Nth barrier (N == coll_sync_barrier_before). Similar for
!_after. Note that N is a _per communicator_ value; not global to the
MPI process.
If both are 0 (which is the default), this component returns NULL for
the comm query, meaning that it is not insertted into the coll module
stack.
The intent of this component is to provide a a workaround for
applications with large numbers of collectives of short messages that
can cause unbounded unexpected messages. Specifically, it is possible
for some iterative collective communication patterns to cause
unbounded unexpected messages. Forcing a barrier before or after
every Nth collective operation would prevent that behavior by forcing
applications to synchronize (and thereby consume any outstanding
unexpected messages caused by collectives on the same communicator).
Open MPI still needs to bound unexpected messages resource consumption
at the receiver, but this is a viable workaround for at least some
symptoms of the problem.
Additionally, there has been anecdotal evidence of some applications
that "perfom better" when they put barriers after other collective
operations. This could be due to many factors -- including shortening
the unexpected message queue. Putting this component in Open MPI
allows people to try this with their own applications and give real
world feedback on this kind of behavior.
This commit was SVN r20584.
Often, orte/util/show_help.h is included, although no functionality
is required -- instead, most often opal_output.h, or
orte/mca/rml/rml_types.h
Please see orte_show_help_replacement.sh commited next.
- Local compilation (Linux/x86_64) w/ -Wimplicit-function-declaration
actually showed two *missing* #include "orte/util/show_help.h"
in orte/mca/odls/base/odls_base_default_fns.c and
in orte/tools/orte-top/orte-top.c
Manually added these.
Let's have MTT the last word.
This commit was SVN r20557.
by r20496 for the sm BTL, openib BTL on iWarp, and the sm & sm2 coll modules.
This commit was SVN r20515.
The following SVN revision numbers were found above:
r20496 --> open-mpi/ompi@4cdf91a8d4
The prior ompi_proc_t structure had a uint8_t flag field in it, where only one
bit was used to flag that a proc was "local". In that context, "local" was
constrained to mean "local to this node".
This commit provides a greater degree of granularity on the term "local", to include tests
to see if the proc is on the same socket, PC board, node, switch, CU (computing
unit), and cluster.
Add #define's to designate which bits stand for which local condition. This
was added to the OPAL layer to avoid conflicting with the proposed movement of
the BTLs. To make it easier to use, a set of macros have been defined - e.g.,
OPAL_PROC_ON_LOCAL_SOCKET - that test the specific bit. These can be used in
the code base to clearly indicate which sense of locality is being considered.
All locations in the code base that looked at the current proc_t field have
been changed to use the new macros.
Also modify the orte_ess modules so that each returns a uint8_t (to match the
ompi_proc_t field) that contains a complete description of the locality of this
proc. Obviously, not all environments will be capable of providing such detailed
info. Thus, getting a "false" from a test for "on_local_socket" may simply
indicate a lack of knowledge.
This commit was SVN r20496.
* New "op" MPI layer framework
* Addition of the MPI_REDUCE_LOCAL proposed function (for MPI-2.2)
= Op framework =
Add new "op" framework in the ompi layer. This framework replaces the
hard-coded MPI_Op back-end functions for (MPI_Op, MPI_Datatype) tuples
for pre-defined MPI_Ops, allowing components and modules to provide
the back-end functions. The intent is that components can be written
to take advantage of hardware acceleration (GPU, FPGA, specialized CPU
instructions, etc.). Similar to other frameworks, components are
intended to be able to discover at run-time if they can be used, and
if so, elect themselves to be selected (or disqualify themselves from
selection if they cannot run). If specialized hardware is not
available, there is a default set of functions that will automatically
be used.
This framework is ''not'' used for user-defined MPI_Ops.
The new op framework is similar to the existing coll framework, in
that the final set of function pointers that are used on any given
intrinsic MPI_Op can be a mixed bag of function pointers, potentially
coming from multiple different op modules. This allows for hardware
that only supports some of the operations, not all of them (e.g., a
GPU that only supports single-precision operations).
All the hard-coded back-end MPI_Op functions for (MPI_Op,
MPI_Datatype) tuples still exist, but unlike coll, they're in the
framework base (vs. being in a separate "basic" component) and are
automatically used if no component is found at runtime that provides a
module with the necessary function pointers.
There is an "example" op component that will hopefully be useful to
those writing meaningful op components. It is currently
.ompi_ignore'd so that it doesn't impinge on other developers (it's
somewhat chatty in terms of opal_output() so that you can tell when
its functions have been invoked). See the README file in the example
op component directory. Developers of new op components are
encouraged to look at the following wiki pages:
https://svn.open-mpi.org/trac/ompi/wiki/devel/Autogenhttps://svn.open-mpi.org/trac/ompi/wiki/devel/CreateComponenthttps://svn.open-mpi.org/trac/ompi/wiki/devel/CreateFramework
= MPI_REDUCE_LOCAL =
Part of the MPI-2.2 proposal listed here:
https://svn.mpi-forum.org/trac/mpi-forum-web/ticket/24
is to add a new function named MPI_REDUCE_LOCAL. It is very easy to
implement, so I added it (also because it makes testing the op
framework pretty easy -- you can do it in serial rather than via
parallel reductions). There's even a man page!
This commit was SVN r20280.
components to use. This code was rendered obsolete (albiet harmless)
by the MCA base improvements that only open the components that were
specified by each framework's MCA parameter.
This commit was SVN r20176.
We still do an interreduce but it is now followed by an intrascatterv.
This fixes trac:1554.
This commit was SVN r19723.
The following Trac tickets were found above:
Ticket 1554 --> https://svn.open-mpi.org/trac/ompi/ticket/1554
There was an argument that was barely used, and on return at the PML
level it contained nothing usable. It has been removed, so now we're
using less memory ...
This commit was SVN r19657.
2. fix a bug in the initial communicator creation of llcomm
3. fix a bug which showed up as the result of fixing issue number 2: we have
to check now whether llcomm has really be created before freeing the
according llcomm in hierarch_destruct.
This commit was SVN r19361.
generate an overlap between the different layers. Why four versions? Because
there is right now always the trade-off between using non-blocking operations
on a layer with a trivial, linear algorithm and using the more sophisticaed
algorithms in a blocking manner.
- bcast_intra_seg used the bcast of lcomm and llcomm, similarly
to original algorithm in hierarch. However, it can segment
the message, such that we might get an overlap between the two
layers. This overlap is based on the assumption, that a process
might be done early with a bcast and can start the next one.
- bcast_intra_seg1: replaces the llcomm->bcast by isend/irecvs
to increase the overlap, keeps the lcomm->bcast however
- bcast_intra_seg2: replaced lcomm->bcast by isend/irecvs
to increase the overlap, keeps however llcomm->bcast
- bcast_intra_seg3: replaced both lcomm->bcast and llcomm->bcast
by isend/irecvs
The code is lightly tested, more testing to follow right now.
This commit was SVN r19358.
* add "register" function to mca_base_component_t
* converted coll:basic and paffinity:linux and paffinity:solaris to
use this function
* we'll convert the rest over time (I'll file a ticket once all
this is committed)
* add 32 bytes of "reserved" space to the end of mca_base_component_t
and mca_base_component_data_2_0_0_t to make future upgrades
[slightly] easier
* new mca_base_component_t size: 196 bytes
* new mca_base_component_data_2_0_0_t size: 36 bytes
* MCA base version bumped to v2.0
* '''We now refuse to load components that are not MCA v2.0.x'''
* all MCA frameworks versions bumped to v2.0
* be a little more explicit about version numbers in the MCA base
* add big comment in mca.h about versioning philosophy
This commit was SVN r19073.
The following Trac tickets were found above:
Ticket 1392 --> https://svn.open-mpi.org/trac/ompi/ticket/1392
hierarch disables itself now if the pml module used is *not* ob1. The reason
is, that the multi-level hierarchy detection algorithm checks the names of the
btl modules used. In case there are no btl's, we would segfault.
Furthermore, three minor changes:
- the 2-level hierarchy detection is now the default (sm vs. everything else
in the world).
- add udapl to the list of protocols checked for by the multi-level hierarch detection
- some of the verbose statements of hierarch were inaccurate. Fixed those comments/messages.
This commit was SVN r18817.
After much work by Jeff and myself, and quite a lot of discussion, it has become clear that we simply cannot resolve the infinite loops caused by RML-involved subsystems calling orte_output. The original rationale for the change to orte_output has also been reduced by shifting the output of XML-formatted vs human readable messages to an alternative approach.
I have globally replaced the orte_output/ORTE_OUTPUT calls in the code base, as well as the corresponding .h file name. I have test compiled and run this on the various environments within my reach, so hopefully this will prove minimally disruptive.
This commit was SVN r18619.
