PGI (20.4) compiler do not define this intrinsic, so only build
AVX512 support if _mm512_mullo_epi64() intrisic is defined.
Signed-off-by: Gilles Gouaillardet <gilles@rist.or.jp>
If PMIX_PACKAGE_RANK is available, uses this value to select between multiple
NIC of equal distance between the current process. If this value is not
available, try to calculate it by getting the locality string from each local
process and assign a package_rank. If everything fails, fall back to using
process_id.rank to select the NIC. This last case is not ideal, but has a small
chance of occuring, and causes an output to be displayed to notify that this is
occuring.
Signed-off-by: Nikola Dancejic <dancejic@amazon.com>
Among many other things:
- Fix an imbalance bug in MPI_allgather
- Accept more human readable configuration files. We can now specify
the collective by name instead of a magic number, and the component
we want to use also by name.
- Add the capability to have optional arguments in the collective
communication configuration file. Right now the capability exists
for segment lengths, but is yet to be connected with the algorithms.
- Redo the initialization of all HAN collectives.
Cleanup the fallback collective support.
- In case the module is unable to deliver the expected result, it will fallback
executing the collective operation on another collective component. This change
make the support for this fallback simpler to use.
- Implement a fallback allowing a HAN module to remove itself as
potential active collective module, and instead fallback to the
next module in line.
- Completely disable the HAN modules on error. From the moment an error is
encountered they remove themselves from the communicator, and in case some
other modules calls them simply behave as a pass-through.
Communicator: provide ompi_comm_split_with_info to split and provide info at the same time
Add ompi_comm_coll_preference info key to control collective component selection
COLL HAN: use info keys instead of component-level variable to communicate topology level between abstraction layers
- The info value is a comma-separated list of entries, which are chosen with
decreasing priorities. This overrides the priority of the component,
unless the component has disqualified itself.
An entry prefixed with ^ starts the ignore-list. Any entry following this
character will be ingnored during the collective component selection for the
communicator.
Example: "sm,libnbc,^han,adapt" gives sm the highest preference, followed
by libnbc. The components han and adapt are ignored in the selection process.
- Allocate a temporary buffer for all lower-level leaders (length 2 segments)
- Fix the handling of MPI_IN_PLACE for gather and scatter.
COLL HAN: Fix topology handling
- HAN should not rely on node names to determine the ordering of ranks.
Instead, use the node leaders as identifiers and short-cut if the
node-leaders agree that ranks are consecutive. Also, error out if
the rank distribution is imbalanced for now.
Signed-off-by: Xi Luo <xluo12@vols.utk.edu>
Signed-off-by: Joseph Schuchart <schuchart@icl.utk.edu>
Signed-off-by: George Bosilca <bosilca@icl.utk.edu>
The current iprobe/improbe implementations merely checks the return
code on the posted receive operation to tell if there is a match or
not. This commit moves the check to the probe's error callback
instead. Per the semantics defined in libfabric, the peek operation is
asynchronous and the results are to be fetched from the completion
queue. If no message is found matching the tags specified in the peek
request, then a completion queue error entry with err field set to
FI_ENOMSG will be available.
Signed-off-by: Raghu Raja <craghun@amazon.com>
In multi-threaded scenarios, any thread that attempts to read a CQ
when there's a pending error CQ entry gets an -FI_EAVAIL. Without
any serialization here (which is okay, since libfabric will protect
access to critical CQ objects), all threads proceed to read from the
error CQ, but only one thread fetches the entry while others get
-FI_EAGAIN indicating an empty queue, which is not erroneous.
Signed-off-by: Raghu Raja <craghun@amazon.com>
* first import of Bull specific modifications to HAN
* Cleaning, renaming and compilation fixing Changed all future into han.
* Import BULL specific modifications in coll/tuned and coll/base
* Fixed compilation issues in Han
* Changed han_output to directly point to coll framework output.
* The verbosity MCA parameter was removed as a duplicated of coll verbosity
* Add fallback in han reduce when op cannot commute and ppn are imbalanced
* Added fallback wfor han bcast when nodes do not have the same number of process
* Add fallback in han scatter when ppn are imbalanced
+ fixed missing scatter_fn pointer in the module interface
Signed-off-by: Brelle Emmanuel <emmanuel.brelle@atos.net>
Co-authored-by: a700850 <pierre.lemarinier@atos.net>
Co-authored-by: germainf <florent.germain@atos.net>
a hierarchical, architecture-aware collective communication module.
Add Reduce and remove up_seg_size and low_seg_size in Bcast
Increase HAN's priority
Signed-off-by: Xi Luo <xluo12@vols.utk.edu>
Signed-off-by: George Bosilca <bosilca@icl.utk.edu>
Made a couple vars static if they didn't look like they were used
more than one place, and added prefixes to a few.
Signed-off-by: Mark Allen <markalle@us.ibm.com>
With Open MPI 5.0, the decision was made to stop building
3rd-party packages, such as Libevent, HWLOC, PMIx, and PRRTE as
MCA components and instead 1) start relying on external libraries
whenever possible and 2) Open MPI builds the 3rd party
libraries (if needed) as independent libraries, rather than
linked into libopen-pal.
This patch moves the hwloc library bundled with Open MPI from a
MCA framework to a stand-alone library built outside of OPAL. Due
to the amount of code in the MCA base (and its assumptions about
being part of an MCA framework), the framework is left with no
active components. Any pre-installed version of HWLOC 1.6 or
newer is preferred over the internal version.
Signed-off-by: Brian Barrett <bbarrett@amazon.com>
With Open MPI 5.0, the decision was made to stop building
3rd-party packages, such as Libevent, HWLOC, PMIx, and PRRTE as
MCA components and instead 1) start relying on external libraries
whenever possible and 2) Open MPI builds the 3rd party
libraries (if needed) as independent libraries, rather than
linked into libopen-pal.
This patch moves libevent from an MCA framework to a stand-alone
library built outside of OPAL. A wrapper in opal/util is provided
to minimize the unnecessary changes in the rest of the code. When
using the internal Libevent, it will be installed as a stand-alone
libevent.a, instead of bundled in OPAL. Any pre-installed version
of Libevent at or after 2.0.21 is preferred over the internal
version.
Signed-off-by: Brian Barrett <bbarrett@amazon.com>
mca_pml_ob1_recv_request_put_frag is used to request a put from the peer if get fails
mca_pml_ob1_recv_request_ack_send_btl is used to send an acknowledgement, not data
Signed-off-by: Joseph Schuchart <schuchart@icl.utk.edu>
- Add support for fallback to previous coll module on non-commutative operations (#30)
- Replace mutexes by atomic operations.
- Use the correct nbc request type (for both ibcast and ireduce)
* coll/base: document type casts in ompi_coll_base_retain_*
- add module-wide topology cache
- use standard instead of synchronous send and add mca parameter to control mode of initial send in ireduce/ibcast
- reduce number of memory allocations
- call the default request completion.
- Remove the requests from the Fortran lookup conversion tables before completing
and free it.
Signed-off-by: George Bosilca <bosilca@icl.utk.edu>
Signed-off-by: Joseph Schuchart <schuchart@hlrs.de>
Co-authored-by: Joseph Schuchart <schuchart@hlrs.de>
The upper 2 bits of an ompi tag encode the synchronize send and
synchronize send ack.
Because the mtl_ofi_create_recv_tag_CQD and mtl_ofi_create_recv_tag
functions both use ompi_mtl_ofi.sync_proto_mask instead of
ompi_mtl_ofi.sync_send when generating their "ignore" masks, they hide
the ack bit, turning the tag into an "any tag receive"
This is an issue because ssend is implemented by doing a send and
receive internally. So if there happens to be an outstanding posted
receive posted before the ssend, that receive will end up consuming the
internal message intended for the ssend's internal receive.
Updating mtl_ofi_create_recv_tag_CQD and mtl_ofi_create_recv_tag functions
to use ompi_mtl_ofi.sync_send fixes this.
Authored-by: John L. Byrne <john.l.byrne@hpe.com>
Signed-off-by: Harumi Kuno <harumi.kuno@hpe.com>
Reduce scatter block and reduce scatter algorithms were hitting
correctness issues for non commutative strided tests. We will revert to
the original default algorithms for those two collectives (basic linear
and non overlapping respectively) in the non commutative op case.
See #8010
Signed-off-by: William Zhang <wilzhang@amazon.com>
As it is possible to have multiple outstanding non-blocking collectives
provided by different collective modules, we need a consistent
mechanism to allow them to select unique tags for each instance of a
collective.
Signed-off-by: George Bosilca <bosilca@icl.utk.edu>
* piggybacking Bull functionalities
* coll/adapt: Fix naming conventions and C11 atomic use
This commit fixes some naming convention issues, such as function names
which should follow the naming ompi_coll_adapt instead of
mca_coll_adapt, reserved for component and module naming (cf. tuned
collective component);
It also fixes the use of _Atomic construct, which is only valid in C11.
OPAL constructs have already been adapted to that use, so use
opal_atomic_* types instead.
* coll/adapt: Remove unused component field in module
This commit removes an unneeded field referencing the component in the
module of adapt, as it is already available through the
mca_coll_adapt_component global variable.
Signed-off-by: Marc Sergent <marc.sergent@atos.net>
Co-authored-by: Lemarinier, Pierre <pierre.lemarinier@atos.net>
Co-authored-by: pierrele <31764860+pierrele@users.noreply.github.com>
the ofi mtl mrecv was not properly setting the message in/out
arg to MPI_MRECV to MPI_MESSAGE_NULL.
Signed-off-by: Howard Pritchard <hppritcha@gmail.com>
MPI-4 is finally cleaning up its language: an MPI "exception" does not
actually exist. The only thing that exists is an MPI "error" (and
associated handlers). This commit replaces all relevant uses of the
word "exception" with "error". Note that this is still applicable in
versions of the MPI standard less than MPI-4.0 (indeed, nearly all the
cases fixed in this commit are just changes to comments, anyway).
One exception to this is the Java bindings, where there's an
MPIException class. In hindsight, it probably should have been named
MPIError, but changing it now would break anyone who is using the Java
bindings.
Signed-off-by: Jeff Squyres <jsquyres@cisco.com>