1
1
Граф коммитов

421 Коммитов

Автор SHA1 Сообщение Дата
Jeff Squyres
567c9e3a5b mtl_ofi_component.c: add missing argv.h header 2015-09-17 10:05:05 -07:00
Nathan Hjelm
987e865c99 mtl/psm2: add support for dynamic add_procs
Add an accessor for the proc_endpoints[OMPI_PROC_ENDPOINT_TAG_MTL]
member of the ompi_proc_t structure. This accessort calls add_procs
with the ompi_proc_t if the member is NULL.

Signed-off-by: Nathan Hjelm <hjelmn@me.com>
2015-09-10 08:55:55 -06:00
Nathan Hjelm
8df9b1d40d mtl/psm: add support for dynamic add_procs
Add an accessor for the proc_endpoints[OMPI_PROC_ENDPOINT_TAG_MTL]
member of the ompi_proc_t structure. This accessort calls add_procs
with the ompi_proc_t if the member is NULL. Tested on an infinipath
system with InfiniPath_QLE7340 HCAs.

Signed-off-by: Nathan Hjelm <hjelmn@me.com>
2015-09-10 08:55:55 -06:00
yohann
404393b9d7 mtl/ofi: Minor code cleanup. 2015-09-03 15:04:55 -07:00
yohann
a8cac09769 mtl/ofi: Renamed macro to prevent clash with FI_ namespace. 2015-09-03 14:42:45 -07:00
yohann
7adb9b7ab4 mtl/ofi: Handle -FI_EAGAIN on send and recv operations. 2015-09-03 10:47:00 -07:00
Ralph Castain
cf6137b530 Integrate PMIx 1.0 with OMPI.
Bring Slurm PMI-1 component online
Bring the s2 component online

Little cleanup - let the various PMIx modules set the process name during init, and then just raise it up to the ORTE level. Required as the different PMI environments all pass the jobid in different ways.

Bring the OMPI pubsub/pmi component online

Get comm_spawn working again

Ensure we always provide a cpuset, even if it is NULL

pmix/cray: adjust cray pmix component for pmix

Make changes so cray pmix can work within the integrated
ompi/pmix framework.

Bring singletons back online. Implement the comm_spawn operation using pmix - not tested yet

Cleanup comm_spawn - procs now starting, error in connect_accept

Complete integration
2015-08-29 16:04:10 -07:00
--quiet
1e9227765a ofi mtl: also link in mtl_ofi_LIBS in the static case 2015-08-20 10:40:46 -07:00
yohann
bcc10fbcd4 mtl/ofi: remove redundant code. 2015-08-19 13:13:59 -07:00
yohann
98b300e1bb mtl/ofi: Require proper ordering by OFI provider. 2015-08-14 16:36:10 -07:00
yohann
27520b99b8 mtl/ofi: add include/exclude list MCA vars.
mtl_ofi_provider_include (resp. mtl_ofi_provider_exclude) can be used
to specify which provider(s) the OFI MTL can select (resp. ignore).

e.g. --mca mtl_ofi_provider_include "psm,sockets"

By default, mtl_ofi_provider_exclude is set to "sockets,mxm".

This deprecates the old MCA var named "mtl_ofi_provider".
2015-08-12 13:52:04 -07:00
yohann
88038b5261 mtl/ofi: Deprioritize some OFI providers.
Some OFI providers such as "sockets" are used for debugging
purposes mostly. For these providers, other components usually
offer better performance -- e.g. for sockets, the BTL/TCP would
be a better choice.
Thus, we chose to ignore some providers unless explicitly asked
by the user on the command line:

e.g. --mca mtl_ofi_provider sockets
2015-08-07 16:09:51 -07:00
Jithin Jose
bc4e8b7e73 Fix warnings in direct (pml-cm,mtl-ofi) build
Signed-off-by: Jithin Jose <jithin.jose@intel.com>
2015-07-29 15:49:37 -07:00
yohann
6eba52a121 mtl/ofi: add missing return. 2015-07-29 14:14:34 -07:00
Howard Pritchard
f5c43c1185 mtl/ofi: retain inline progress function
Retain inline progress function for ofi
mtl, but have a non-inlined progress function
which is registered with the opal progress
mechanism.

 @jithinjosepkl

I've bad news about the psm provider.  I still notice
segfaults - not always - but frequently at finalize
when using the psm provider.  I don't notice this
when using the sockets provider.

Signed-off-by: Howard Pritchard <howardp@lanl.gov>
2015-07-27 09:16:52 -06:00
Ralph Castain
265cd14f60 Purge whitespace 2015-06-25 13:27:56 -07:00
rhc54
8b62b63786 Merge pull request #666 from rhc54/topic/libfab
Add a common/libfabric component
2015-06-25 12:46:34 -07:00
Yohann Burette
7fd5ded327 mtl/ofi: message truncation is now indicated by FI_ETRUNC. 2015-06-25 11:06:41 -07:00
Yohann Burette
483ff23db1 mtl/ofi: cancels are now tracked by an error entry. 2015-06-25 11:06:41 -07:00
Ralph Castain
ea0e21bb06 Add a common/libfabric component to the opal layer where we can place common functions 2015-06-25 11:04:00 -07:00
Ralph Castain
869041f770 Purge whitespace from the repo 2015-06-23 20:59:57 -07:00
Andrew Friedley
2c9be59b37 Add new PSM2 MTL.
This new MTL runs over PSM2 for Omni Path.  PSM2 is a descendant of PSM
with changes to support more ranks and some MPI-3 features like mprobe.

PSM2 will only support Omni Path networks; PSM only supports True Scale.
Likewise, the existing PSM MTL will continue to be maintained for True
Scale, while the PSM2 MTL is developed and maintained for Omni Path.
2015-06-22 07:55:46 -07:00
rhc54
9a8bda0b72 Merge pull request #637 from jithinjosepkl/pr/pml-cm-opt
pml-cm bug fixes
2015-06-15 19:25:09 -07:00
Jithin Jose
7ccde09a09 Do opal_convertor_copy_and_prepare_for_send for buffered send mode as
MCA_PML_CM_HVY_SEND_REQUEST_BSEND_ALLOC calls opal_convertor_pack
directly.

Signed-off-by: Jithin Jose <jithin.jose@intel.com>
2015-06-15 17:12:50 -07:00
Jeff Squyres
cc66745e7a mtl/ofi: convert to use external libfabric
Use the new OPAL_CHECK_LIBFABRIC macro.
2015-06-15 15:17:06 -07:00
Todd Kordenbrock
b725186768 mtl-portals4: Verify the result of PtlPTAlloc()
The Portals4 MTL allocates two Portals IDs requesting specific
well-known IDs and assumes that those IDs are allocated.  If those IDs
are in use, PtlPTAlloc() will allocate a different ID.  This commit
verifies that the requested IDs were allocated.
2015-06-09 14:43:50 -05:00
Jithin Jose
c745854d9b Avoid opal_convertor_pack for contigous data types in MXM mtl
Signed-off-by: Jithin Jose <jithin.jose@intel.com>
2015-05-27 11:09:25 -07:00
Jithin Jose
50089977ac Inline PML-CM
Signed-off-by: Jithin Jose <jithin.jose@intel.com>
2015-05-26 21:42:41 -07:00
Jithin Jose
56869bff38 Avoid datatype pack/unpack for contiguous data on homogenous systems.
Signed-off-by: Jithin Jose <jithin.jose@intel.com>
2015-05-26 21:42:41 -07:00
Jeff Squyres
fb12572438 OFI: make v1.10 and v2.0 fit in version checking scheme
v1.10 is now in the same compatibility level as v1.7/v1.8 (there
is/will be no v1.9 series).  v2.0 now takes over for what used to be
called v1.9.
2015-05-26 18:58:28 -07:00
Todd Kordenbrock
c42e277385 mtl-portals4: thread multiple updates
When activating short receive blocks on the overflow list, remove
the PTL_ME_EVENT_LINK_DISABLE flag so the event gets generated.
Without PTL_EVENT_LINK, the block status can't reach the activated
state.

Replace #ifdef with #if for Open MPI configure booleans, because
Open MPI configure booleans are always defined and the value must
be checked.
2015-05-13 17:06:18 -05:00
Yohann Burette
27f1884cf8 mtl/ofi: Reworked header files. Added compat to ease maintenance. 2015-05-12 15:47:50 -07:00
Ryan Grant
bbeaf41a52 Merge pull request #580 from tkordenbrock/topic/mtl.add.status.to.short.recv.blocks
mtl-portals4: add status to short recv blocks to coordinate out of or…
2015-05-11 13:44:45 -06:00
Todd Kordenbrock
9df163f116 portals4: use a single Memory Descriptor to cover all of memory
In days past, some implementations of Portals4 could not cover all
of memory with a single Memory Descriptor so multiple large
overlapping Memory Descriptors were created.  Because none of the
current implementations have this limitation (and no future
implementations should either), this commit removes the overlapping
Memory Descriptors code.
2015-05-11 11:49:41 -05:00
Todd Kordenbrock
074583060d mtl-portals4: add status to short recv blocks to coordinate out of order events
If OMPI is initialized as thread multiple, then it is possible for
Portals events to be processed out of order by different threads.
Out of order events could lead to reactivation of the block
(PTL_EVENT_AUTO_FREE) before the block is removed from the active
list (PTL_EVENT_AUTO_UNLINK).  This commit adds a status field to
ompi_mtl_portals4_recv_short_block_t that coordinates these events.
2015-05-11 11:48:25 -05:00
Ryan Grant
6ab91a6781 Merge pull request #561 from tkordenbrock/topic/mtl.fix.datatype.overflow
mtl-portals4: fix datatype overflow in ompi_mtl_portals4_long_isend()
2015-04-28 15:15:59 -06:00
Todd Kordenbrock
8a4616f724 mtl-portals4: fix datatype overflow in ompi_mtl_portals4_long_isend()
The length parameter of ompi_mtl_portals4_long_isend() was declared
as "int", which may not be big enough depending on the platform and
compiler options used.  This commit changes the type to size_t to
prevent overflow.
2015-04-28 14:40:25 -05:00
Todd Kordenbrock
3e437f6184 mtl-portals4: expand the source field of the match bits to 24 bits
The source field was 16 bits which is not sufficient for many
current and future machines.  This commit expands the source field
to 24 bits and reduces the tag field from 32 bits to 24 bits.
2015-04-28 14:25:30 -05:00
Yohann Burette
1be185ed87 mtl/ofi: Remove use of MR. 2015-04-24 15:55:21 -07:00
Todd Kordenbrock
8e56002ec7 mtl-portals4: add missing return to portals4_init_interface() 2015-04-21 11:30:33 -05:00
Todd Kordenbrock
34c50fa963 mtl-portals4: move MD cleanup closer to failure
PtlMDRelease() was called if read_msg() returned a failure code.
This commit moves the PtlMDRelease() inside read_msg() so that it
doesn't get called in cases where the failure happens before or at
the PtlMDBind().
2015-04-21 11:30:33 -05:00
Todd Kordenbrock
422be76770 mtl-portals4: add a debug message for thread multiple mode 2015-04-21 11:30:33 -05:00
Todd Kordenbrock
35e5ffd001 mtl-portals4: add the option to use the Portals4 logical to physical table
This commit adds an MCA variable to select Portals4 logical
addressing, populates the logical-to-physical mapping table and
initializes the NI in this mode.
2015-04-21 11:30:33 -05:00
Yohann Burette
19607d2ce7 mtl/ofi: Remove memset() from progress path. 2015-04-20 14:12:39 -07:00
Yohann Burette
d2eda04801 mtl/ofi: Use fi_tinject for small messages. 2015-04-20 14:12:39 -07:00
Nathan Hjelm
033894b493 Merge pull request #541 from hjelmn/c99_components
C99 component initialization
2015-04-20 10:45:39 -06:00
Yohann Burette
ba1bc00df1 mtl/ofi: Remove FI_CANCEL. 2015-04-20 09:40:37 -07:00
Nathan Hjelm
df75d0382f ompi: use C99 subobject naming for component initialization
This commit helps future-proof ompi components by initializing each
component member by name.

Signed-off-by: Nathan Hjelm <hjelmn@lanl.gov>
2015-04-18 10:29:58 -06:00
Yohann Burette
9392bb5ede mtl/ofi: Implement Probe/Mprobe/Mrecv using FI_PEEK/FI_CLAIM. 2015-04-17 16:42:13 -07:00
Nathan Hjelm
3436f2917d Merge pull request #449 from hjelmn/mca_base_update
mca/base update
2015-04-16 08:41:48 -06:00