ports/openmpi
ports/openmpi
1
1
Форкнуть 0
Код Релизы Активность
4 780 коммитов 3 Ветки 0 Теги 130 MiB
f7d2b120ef
Граф коммитов

4780 Коммитов

This Branch
This Branch
Все ветки
Автор SHA1 Сообщение Дата
Galen Shipman
507730a2ca Added with-bmi-ib instead of with-ptl-ib 
Begun adding support for RDMA_WRITE_WITH_IMM

This commit was SVN r6081.
 2005-06-17 19:15:03 +00:00
Tim Woodall
2bc9d9c657 added timer profiling to trace rdma protocol 
This commit was SVN r6080.
 2005-06-17 18:25:44 +00:00
Galen Shipman
f886160f7c Check rr on send.. 
This commit was SVN r6079.
 2005-06-17 17:14:45 +00:00
Jeff Squyres
8a0e7dfbb5 Arf. Need to explicitly add _DATA macros to EXTRA_DIST. Grumble. 
This commit was SVN r6077.
 2005-06-17 09:40:16 +00:00
Jeff Squyres
93ba5e0a84 Oops -- the prettyprint label was too wide. 
This commit was SVN r6076.
 2005-06-17 01:11:08 +00:00
Jeff Squyres
d78ecb98d8 - Make ompi_info show reasonable help messages 
- Take Ben Allan's suggestion to have ompi_info show the absolute path
  name of the compilers that were used to compile OMPI

This commit was SVN r6075.
 2005-06-17 01:06:36 +00:00
Jeff Squyres
d1a362e5de I never liked the ****'s -- change to -------'s. 
This commit was SVN r6074.
 2005-06-17 01:04:21 +00:00
Jeff Squyres
d9b0aa9654 Temporarily comment out the test_rds2 test because all it does is test 
the RDS selection logic, which is, unfortunately, not yet well
supported by the testing infrastructure (it causes false failures in
the nightly build).

This commit was SVN r6073.
 2005-06-16 11:25:27 +00:00
Galen Shipman
2cb52d028e Fixed rdma protocol for small fragments from the free list, The opcode 
VAPI_SEND was not being set in mca_bmi_ib_endpoint_post_send so when the frag
had been used for an RMDA the opcode was incorrect. 

Various other misc fixes related to small frag rdma. 

This commit was SVN r6072.
 2005-06-16 03:49:16 +00:00
Tim Woodall
cddaa0a292 remove threshold parameter 
This commit was SVN r6071.
 2005-06-15 19:54:16 +00:00
Tim Woodall
9c3dcdc7a2 loopback bmi 
This commit was SVN r6069.
 2005-06-15 19:10:26 +00:00
Galen Shipman
6d6510d147 Added a second queue pair for high priority messages 
This commit was SVN r6068.
 2005-06-15 19:09:51 +00:00
Tim Woodall
2b756307af add flag to indicate memory is already pinned 
This commit was SVN r6066.
 2005-06-15 17:29:41 +00:00
Tim Woodall
406679cb20 run progress after sending schedule 
This commit was SVN r6065.
 2005-06-15 15:25:19 +00:00
Tim Woodall
b59c4368dd send control message with higher priority 
This commit was SVN r6064.
 2005-06-15 15:23:20 +00:00
Tim Woodall
22a946d789 add priority bit to descriptor flags fields for high priority (e.g. control) messages 
the bmi should attempt to deliver these at a higher priority than other traffic 

This commit was SVN r6063.
 2005-06-15 15:08:58 +00:00
Galen Shipman
d6716465d3 Performance tuning.. 
This commit was SVN r6062.
 2005-06-15 14:11:42 +00:00
Brian Barrett
21a3974b73 * as requested by Ben Allan, include mpicxx in addition to mpic++ and (maybe) 
mpiCC

This commit was SVN r6061.
 2005-06-14 22:41:54 +00:00
Ralph Castain
62854316fb Fix a bug that caused proxies to duplicate subscriptions. Make sure to regenerate non-blocking receive after processing callbacks for timing reasons. 
This commit was SVN r6059.
 2005-06-14 20:10:00 +00:00
Tim Woodall
a216a056c3 update current offset prior to using the convertor - may have been changed by rdma protocol 
This commit was SVN r6058.
 2005-06-14 12:56:36 +00:00
Ralph Castain
092f14f2c5 Fix the timing between turning notifies on when trigger fires so that only one notification is sent. Reverse the order of processing callbacks so that the xcast comes out after all messages sent. 
This commit was SVN r6057.
 2005-06-14 12:37:38 +00:00
Josh Hursey
98bf06baa6 minor bit of wrapping 
This commit was SVN r6056.
 2005-06-13 22:10:07 +00:00
Galen Shipman
7f5296bcfc Changed variable names .. 
This commit was SVN r6055.
 2005-06-13 21:55:20 +00:00
Josh Hursey
1587b01a19 Some changes that allow the mca directory to compile for Windows. 
We should refrain from using the variable 'new' since it is a 
keyword in C++, and some compilers are really picky about it,
namely cl :)

Did a bit of funny wrapping in convertor.h so that the 
ompi_convertor_prepare function is only exported when
doing a windows build. Talked with George about this
change and we both agreed that the windows build is 
messy, but shouldn't corrupt the main line precausions.

Also wrapped convertor.h in the __cplusplus defines
to help with exporting the inline functions under 
windows.

This commit was SVN r6054.
 2005-06-13 21:41:25 +00:00
Tim Woodall
6e919f9fb1 - fix scheduling logic 
- use cached descriptor for control messages

This commit was SVN r6053.
 2005-06-13 20:52:13 +00:00
Ralph Castain
83cba7f7cf Checkpoint. Fixed a logic problem that removed one-shot subscriptions even though the notifiers were supposed to stay. 
This commit was SVN r6052.
 2005-06-13 20:43:05 +00:00
Josh Hursey
729dc65935 Some small changes to get libmpi building on Windows. Also fixed a typo in convertor.h 
This commit was SVN r6051.
 2005-06-13 20:28:44 +00:00
Brian Barrett
43f3a7ee19 * re-add check to see if sched_yield should be called or not. This had been 
accidently removed along the way...

This commit was SVN r6050.
 2005-06-13 19:32:52 +00:00
Jeff Squyres
4fc09dd261 Fix up faulty Makefile.am so that it builds DSO's properly 
This commit was SVN r6049.
 2005-06-13 18:34:00 +00:00
Galen Shipman
0bca761350 Use prepinned memory via MPI_alloc_mem when appropriate 
This commit was SVN r6048.
 2005-06-13 18:24:48 +00:00
Ralph Castain
445f8a802e Insert a flag into the registry replica to indicate that we are already processing callbacks, so don't call process_callbacks again. Helps to limit the stack size. 
This commit was SVN r6047.
 2005-06-13 17:00:57 +00:00
Galen Shipman
42d6bf48b6 clean up 
This commit was SVN r6046.
 2005-06-13 15:20:36 +00:00
Tim Woodall
68aefc12cf unpack addresses individually - as they are packed that way 
This commit was SVN r6045.
 2005-06-13 14:50:59 +00:00
Ralph Castain
3dd08bf0c6 Fix another subscription in the OOB to be quiet until after trigger fires. 
This commit was SVN r6044.
 2005-06-13 14:49:21 +00:00
Ralph Castain
098cc8cf3a Bring the rest of the notification modes online. Update the unit test to cover notify-on-change. 
This commit was SVN r6043.
 2005-06-13 14:37:02 +00:00
Galen Shipman
e580c98365 flag for rdma/directed send in IB is now a mca param 
clear the descriptors on completion.. 

This commit was SVN r6042.
 2005-06-13 14:36:09 +00:00
Ralph Castain
25f0075ac3 Believe this should fix Friday's concerns over callback stacks. The registry was working just fine - the problem was that the modex exchange was asking to be notified any time an entry was added or changed, which the registry dutifully did. However, the initial "put" of information takes place during the compound command portion of mpi_init, during which time we do not allow messages to come back to the process. 
All that was required was to change the subscription request in module_exchange to specify "don't start notifying me until after the trigger". The trigger is set to fire when all processes have completed their initial stage_gate, so that means the process will receive notifications only of any changes from the base starting point.

Specifically:

1. I changed the "don't start" flag to ORTE_GPR_NOTIFY_STARTS_AFTER_TRIG to make it more obvious as to its function

2. Added that flag to the module_exchange subscription, and changed the name elsewhere it was used (I had already fixed this problem for the OOB)

3. Found an unrelated compiler warning of an unused variable in ns_proxy, so I fixed it too.

This commit was SVN r6041.
 2005-06-13 13:30:12 +00:00
Galen Shipman
66e0fb28dc bugfix.. 
This commit was SVN r6040.
 2005-06-12 22:55:00 +00:00
Galen Shipman
48677e4489 Changed recv to check des_dst instead od des_src 
other misc fixes. 

This commit was SVN r6039.
 2005-06-11 00:44:45 +00:00
Jeff Squyres
6477eb3c14 Need to have a virtual destructor since we have one virtual function 
(g++ 4.0 complains).  This function was commented out from many years
ago in the "bad old days" of new-not-conformant compilers.

This commit was SVN r6038.
 2005-06-10 21:02:52 +00:00
Galen Shipman
b3c8d47f62 More rdma work.. 
This commit was SVN r6036.
 2005-06-10 20:40:46 +00:00
Galen Shipman
215ba87116 Aligned the send and receive descriptors 
This commit was SVN r6034.
 2005-06-10 16:46:22 +00:00
Galen Shipman
0bdabafd9b Added RDMA interfaces to ib bmi. 
This commit was SVN r6033.
 2005-06-10 16:02:34 +00:00
Tim Woodall
dea537c9f7 only use rdma protocol for contigous data 
This commit was SVN r6032.
 2005-06-10 14:09:34 +00:00
Jeff Squyres
a1dd4e3173 Be a bit more social while spinning during init to setup the shared 
memory.

This commit was SVN r6031.
 2005-06-10 13:14:32 +00:00
Thara Angskun
5a7b0b598d prevent compilation error when compiled with xlc. 
This commit was SVN r6029.
 2005-06-10 08:06:07 +00:00
Jeff Squyres
9014ab2e86 mpio.h was incorrectly listed twice, and was incorrectly included in 
the distribution tarball (mpio.h is automatically generated and should
not be in the tarball).

This commit was SVN r6028.
 2005-06-10 04:20:57 +00:00
Ralph Castain
5a2c6a0b87 Fix the trunk so it runs again: 
1. Tell the oob subscription not to report on changes until AFTER the trigger fires (signifying that all procs have started)

2. Make sure tokens returned to the subscriber are from the container that housed the data being returned (previous mistake - tokens were from subscription, not from container)

Tested on the Mac running up to 10 processes

This commit was SVN r6025.
 2005-06-10 03:35:36 +00:00
Galen Shipman
ec1c1bae13 Fixed up sending large fragments and preposting of receive descriptors.. 
This commit was SVN r6023.
 2005-06-10 00:01:05 +00:00
George Bosilca
700ec8f999 Add one arguement to the memory allocation function. 
This commit was SVN r6022.
 2005-06-09 22:26:22 +00:00