1
1
Граф коммитов

646 Коммитов

Автор SHA1 Сообщение Дата
Nathan Hjelm
f68c315188 pml/ob1: add missing ompi_request_wait_completion for buffered sends
This commit adds a call to ompi_request_wait_completion for buffered
sends. Without this line it is possible to get into a state where the
data is never sent.

Fixes open-mpi/ompi#1185

Signed-off-by: Nathan Hjelm <hjelmn@lanl.gov>
2015-12-07 22:28:07 -07:00
Rolf vandeVaart
f2ff6e03ab Make CUDA 4.1 a requirement for CUDA-aware support.
Remove all related preprocessor conditionals.
2015-10-29 11:24:02 -04:00
Nathan Hjelm
08e267b811 add_procs: add threading protection for dynamic add_procs
This commit add protection to the group, ob1, and bml endpoint lookup
code. For ob1 and the bml a lock has been added. For performance
reasons the lock is only held if a bml or ob1 endpoint does not
exist. ompi_group_dense_lookup no uses opal_atomic_cmpset to ensure
the proc is only retained by the thread that actually updates the
group.

Signed-off-by: Nathan Hjelm <hjelmn@lanl.gov>
2015-10-21 16:13:41 -06:00
Nathan Hjelm
bedd80214e pml/ob1: remove priority check
This commit removes code that checks the ob1 priority vs the previous
priority. The previous priority is meaningless here and may only cause
ob1 to disable itself when it shouldn't.

Signed-off-by: Nathan Hjelm <hjelmn@lanl.gov>
2015-10-19 12:32:41 -06:00
KAWASHIMA Takahiro
4e56505202 pml/ob1: Fix a memory leak regarding pending FIN control messages.
Once a FIN control message is appended to the pending list,
the ob1 PML attempts to send the FIN again in the                               `mca_pml_ob1_process_pending_packets` function.
But if the PML failed to sent the FIN again, the `mca_pml_ob1_send_fin`
function creates a new `mca_pml_ob1_pckt_pending_t` object and the
old object is not retured to the free list.
2015-10-15 11:15:03 +09:00
Nathan Hjelm
12bd300c40 Merge pull request #929 from hjelmn/add_procs
Update add_procs support
2015-09-28 17:29:13 -06:00
Nathan Hjelm
6611c000c9 Fix coverity warnings
Fix CID 1315271: Constant expression result

The intent of this conditional is to not produce a peruse event for
probe or mprobe requests. Coverity is correct that the expression is
always true. Changed the || to && to fix. Also moved the conditional
within an OMPI_WANT_PERUSE to ensure the conditional is not evaluated
if peruse is disabled.

Signed-off-by: Nathan Hjelm <hjelmn@lanl.gov>
2015-09-28 15:35:25 -06:00
George Bosilca
01d8e23ccc Fix the random errors related to the recursive sends and receives
identified by Fujitsu.
2015-09-26 00:44:51 +02:00
Nathan Hjelm
54a4061d88 Add support for detecting when dynamic add_procs is not possible
This commit adds support to the pml, mtl, and btl frameworks for
components to indicate at runtime that they do not support the new
dynamic add_procs behavior. At the high end the lack of dynamic
add_procs support is signalled by the pml using the new pml_flags
member to the pml module structure. If the
MCA_PML_BASE_FLAG_REQUIRE_WORLD flag is set MPI_Init will generate the
ompi_proc_t array passed to add_proc from ompi_proc_world () instead
of ompi_proc_get_allocated ().

Both cm and ob1 have been updated to detect if the underlying mtl and
btl components support dynamic add_procs.

Signed-off-by: Nathan Hjelm <hjelmn@lanl.gov>
2015-09-23 16:22:05 -06:00
Gilles Gouaillardet
a611274704 pml: fix commit open-mpi/ompi@6e6a3e965c
do not use the const modifier for allocator nor recv buffers
2015-09-18 09:54:18 +09:00
Nathan Hjelm
b4a0d40915 pml/ob1: Add support for dynamically calling add_procs
This commit contains the following changes:

 - pml/ob1: use the bml accessor function when requesting a bml
   endpoint. this will ensure that bml endpoints are only created when
   needed. for example, a bml endpoint is not requested and not
   allocated when receiving an eager message from a peer.

 - pml/ob1: change the pml_procs array in the ob1 communicator to a
   proc pointer array. at the cost of a single level of extra
   redirection this will allow us to allocate pml procs on demand.

 - pml/ob1: add an accessor function to access the pml proc structure
   for a given peer. this function will allocate the proc if it
   doesn't already exist.

Signed-off-by: Nathan Hjelm <hjelmn@lanl.gov>
2015-09-10 08:55:54 -06:00
Gilles Gouaillardet
6e6a3e965c pml: do not cast way the const modifier when this is not necessary
update the pml framework and mpi c bindings
2015-09-09 09:18:57 +09:00
Rolf vandeVaart
30a872b478 Add the ability to send host buffers through one sized staging buffers and CUDA buffers through different sized buffers. Fixes performance issues 2015-07-02 11:11:15 -04:00
bosilca
1b8556f926 Merge pull request #653 from hjelmn/moar_ob1_fixes
pml/ob1: fix bugs in static request objects
2015-06-24 14:28:11 -07:00
Ralph Castain
869041f770 Purge whitespace from the repo 2015-06-23 20:59:57 -07:00
Nathan Hjelm
9a8a87611e pml/ob1: fix bugs in static request objects
This commit fixes several bugs in the static request objects used by
ob1 for blocking send/receive operations.

 - Fix memory leak when using MPI_THREAD_MULTIPLE. Requests were
   allocated off the free list but were destructed and NOT returned.

 - Fix double-destruct of static objects. There is no reason to
   CONSTRUCT/DESTUCT the static object for each send/receive
   operation. This adds overhead and no benefit. To keep the code
   clean helper functions have been added to finalize ob1 send/receive
   requests.

 - Remove now unnecessary include of alloca.h.

Signed-off-by: Nathan Hjelm <hjelmn@me.com>
2015-06-23 11:00:45 -06:00
Nathan Hjelm
284dd6babe pml/ob1: do not use OPAL_ENABLE_MULTI_THREADS to determine thread multiple support
OPAL_ENABLE_MULTI_THREADS is always on. The correct value to check is
OMPI_ENABLE_THREAD_MULTIPLE.

Signed-off-by: Nathan Hjelm <hjelmn@me.com>
2015-06-22 19:17:23 -06:00
Gilles Gouaillardet
ee3a1da28a pml/ob1:mca_pml_ob1_recv_request_put_frag silence a warning
proc local variable is used only in heterogeneous mode
2015-06-15 10:00:53 +09:00
George Bosilca
67b70bb47a Add multi-threaded support. 2015-06-12 14:22:17 -07:00
George Bosilca
b2cf74cabc A first cut at a possible solution for the missing requests
from the message queues (a debugging feature). With this approach
all blocking (single threaded) requests are allocated from the main
freelist, so they will be accounted for during the message queues
investigation).
2015-06-12 14:22:17 -07:00
Gilles Gouaillardet
e980958ad4 pml/ob1: silence a warning 2015-05-26 15:05:44 +09:00
Gilles Gouaillardet
85c45e2275 pml/ob1: fix mca_pml_ob1_recv_request_put_frag(...) in heterogeneous mode 2015-05-22 15:48:45 +09:00
Nathan Hjelm
ce48eabd84 pml/ob1: use c99 flexible array members instead of size 1 arrays
This commit updates several ob1 structures to take advantage of C99's
flexible array member.

Signed-off-by: Nathan Hjelm <hjelmn@lanl.gov>
2015-05-20 10:31:35 -06:00
Nathan Hjelm
033894b493 Merge pull request #541 from hjelmn/c99_components
C99 component initialization
2015-04-20 10:45:39 -06:00
Nathan Hjelm
d251fa1525 pml/ob1: fix heterogenous build
Signed-off-by: Nathan Hjelm <hjelmn@lanl.gov>
2015-04-20 09:27:00 -06:00
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
adrianreber
714d9aa67e Merge pull request #348 from adrianreber/topic/orte_cr_continue_like_restart
Topic/orte cr continue like restart
2015-03-12 14:54:02 +01:00
Adrian Reber
c08e234af7 FT: fix compilation using --with-ft (5/5)
Enabling the FT code breaks compilation (again). This series
tries to fix the compiler errors. This is again only fixing
the compiler errors without any warranty that the result
might actually support FT again.

With the changes introduced in the previous patches in this series
some goto constructs for cleanup are no longer necessary and removed.
2015-03-11 14:23:33 +01:00
Adrian Reber
1c5a8df724 FT: fix compilation using --with-ft (2/5)
Enabling the FT code breaks compilation (again). This series
tries to fix the compiler errors. This is again only fixing
the compiler errors without any warranty that the result
might actually support FT again.

The FT code used barrier mechanisms which have been removed
with aec5cd08bd. This patch replaces
all those different barriers with opal_pmix.fence(NULL, 0);
I am not sure this is completely correct but at least a starting
point for a review.
2015-03-11 14:23:33 +01:00
Adrian Reber
f45dd069bd FT: fix compilation using --with-ft (1/5)
Enabling the FT code breaks compilation (again). This series
tries to fix the compiler errors. This is again only fixing
the compiler errors without any warranty that the result
might actually support FT again.

This first patch moves orte_cr_continue_like_restart from ORTE
to opal_cr_continue_like_restart in OPAL. This only leaves three
calls from OPAL to ORTE in the FT code. As it is not yet 100%
clear how to handle these calls the code orte_sstore.set_attr()
has been #ifdef'd out for now.
2015-03-11 14:23:33 +01:00
Nathan Hjelm
3d32dbd793 btl/openib: cuda: fix CUDA-aware support with async copy
This commit should resolve an issue seen with CUDA-aware support. The
problem came in with BTL 3.0. Before 3.0 the size of the copy was
stored in the incoming segment's des_remote_count field. This field
does not exist in BTL 3.0 so I stored the value in the
des_segment_count field. This caused problems with the cuda support
code. To fix the issue the endpoint pointer is now stored in the in
fragment's endpoint pointer which free's up the segment's des_cbdata
pointer for storing the transfer size.

Signed-off-by: Nathan Hjelm <hjelmn@lanl.gov>
2015-03-10 14:38:12 -06:00
Rolf vandeVaart
30e9dd5066 Look in extra rdma array to find bml. This is needed with recent BML changes. Only affects CUDA-aware code. 2015-02-27 09:02:21 -05:00
George Bosilca
3fd8dc099d Revert "This function is now useless."
This reverts commit 0871c5c489.
2015-02-26 17:54:46 -05:00
George Bosilca
7f90cedf23 Revert "Fix the logic for computing the different weights for each BTLs. This"
This reverts commit de118609ec.
2015-02-26 17:54:31 -05:00
George Bosilca
f3b58006c8 Merge branch 'master' of github.com:open-mpi/ompi 2015-02-25 12:01:35 -05:00
Jeff Squyres
c3381150de ob1: fix another PERUSE compile error 2015-02-25 05:53:12 -08:00
Nathan Hjelm
0ac2f08460 pml/ob1: fix peruse compile error
Fixes #416
2015-02-24 15:39:46 -07:00
Nathan Hjelm
5f1254d710 Update code base to use the new opal_free_list_t
Use of the old ompi_free_list_t and ompi_free_list_item_t is
deprecated. These classes will be removed in a future commit.

This commit updates the entire code base to use opal_free_list_t and
opal_free_list_item_t.

Notes:

OMPI_FREE_LIST_*_MT -> opal_free_list_* (uses opal_using_threads ())

Signed-off-by: Nathan Hjelm <hjelmn@lanl.gov>
2015-02-24 10:05:45 -07:00
Howard Pritchard
bf89131f9e add owner files to opa/ompi/orte mca directories
This commit adds an owner file in each of the component directories
for each framework.  This allows for a simple script to parse
the contents of the files and generate, among other things, tables
to be used on the project's wiki page.  Currently there are two
"fields" in the file, an owner and a status.  A tool to parse
the files and generate tables for the wiki page will be added
in a subsequent commit.
2015-02-22 15:10:23 -07:00
George Bosilca
0871c5c489 This function is now useless. 2015-02-21 16:38:17 -05:00
George Bosilca
de118609ec Fix the logic for computing the different weights for each BTLs. This
removes the call to qsort, as the BTLs are already sorted based on
their respective bandwidth.
2015-02-21 16:37:18 -05:00
Rolf vandeVaart
dbd0064713 Fix bug in CUDA-aware and GDR introduced by refactoring 2015-02-18 17:44:28 -05:00
Nathan Hjelm
3847025540 pml/ob1: when using btl_get try to register the entire region before attempting to break the get into multiple rdma fragments
A little background. Historically ob1 always registered the entire memory
region when the RGET protocol was in use. This changed when Mellanox
added support to fragment RGET using the btl_prepare_dst function. Now
that the BTL layer has changed to split out the limits of get/put there
is explicit fragmentation code in ob1. Before this commit the registration
was still done per RGET fragment.

This commit will attempt to register the entire region before creating
RGET fragments. If the registration is successfull then all RGET
fragments will use this registration otherwise they will each attempt
to register their own segment of the receive buffer. If that fails
enough times each fragment will give up and fall back on send/recv.

Signed-off-by: Nathan Hjelm <hjelmn@lanl.gov>
2015-02-13 11:46:37 -07:00
Nathan Hjelm
c4a0e02261 pml/ob1: update for BTL 3.0 interface
Signed-off-by: Nathan Hjelm <hjelmn@lanl.gov>
2015-02-13 11:46:37 -07:00
George Bosilca
df0512550e The extent of the datatype is irrelevant for deciding to do an immediate
send as long as we have to pack.
2015-01-19 02:23:12 -05:00
Gilles Gouaillardet
d14daf40d0 ob1: correctly handle types in which size > extent
do not send inline if extent*count *OR* size*count are greater than 256
2015-01-19 14:07:23 +09:00
Howard Pritchard
3fc7b389ff initial async progress changes for gni 2014-12-24 11:50:23 -07:00
Nathan Hjelm
1b564f62bd Revert "Merge pull request #275 from hjelmn/btlmod"
This reverts commit ccaecf0fd6, reversing
changes made to 6a19bf85dd.
2014-11-19 23:22:43 -07:00
Nathan Hjelm
0110603782 ob1 warning fix 2014-11-19 11:33:04 -07:00
Nathan Hjelm
24427639b6 Fix ob1 warnings 2014-11-19 11:33:03 -07:00