ports/openmpi
ports/openmpi
1
1
Форкнуть 0
Код Релизы Активность
8 946 коммитов 3 Ветки 0 Теги
Граф коммитов

12 Коммитов

Автор SHA1 Сообщение Дата
Brian Barrett
bfbc281e93 Fix slow startup issue with the MX MTL. The problem is caused by mx_connect() being a one-sided operation from the API level, but not being an interrupting call when the target is not entering the MX library. So if most of the processes exit mtl_mx_add_procs() and enter the stage gate 2 barrier, the other processes can only progress their mx_connect() calls when the targets enter the mx library. Because the event library is in EV_ONELOOP mode, this only happens once a second. The mx progress thread (hidden in the MX library) also only wakes up once a second, so mx_connect calls can take a second to complete. 
The temporary solution is to switch into EV_NONBLOCK mode earlier (right after the mx_connect loop) so that there isn't a giant slowdown when processes enter the stage gate 2 barrier before other proesses.  They will now not block in the event library for any period of time, which appears to have a 50% speedup when running at > 64 procs.

Refs trac:645

This commit was SVN r12713.

The following Trac tickets were found above:
  Ticket 645 --> https://svn.open-mpi.org/trac/ompi/ticket/645
 2006-12-01 02:49:01 +00:00
George Bosilca
63462331c9 Reduce the number of branches. Keep the fast path as short as possible. 
Remove some useless error checking. Add OPAL_UNLIKELY directives.

This commit was SVN r12477.
 2006-11-07 23:59:32 +00:00
George Bosilca
108ea4dbe9 When the MX MTL complete a request, force a return from the progress function. 
Decrease the latency by about 0.3 microseconds.

This commit was SVN r12454.
 2006-11-06 23:13:07 +00:00
Brian Barrett
c3306f7073 * don't abort if we get a status error - just pass it on to the next layer up 
This commit was SVN r11791.
 2006-09-25 17:28:24 +00:00
Brian Barrett
6d414f2d44 * use the MTL-specific output stream for all error messages 
* use OPAL_OUTPUT_VERBOSE rather than printfs for debugging messages

This commit was SVN r11227.
 2006-08-16 16:28:58 +00:00
Brian Barrett
d3c6035ea9 * allow direct calling to work with the MX MTL. Had to move some types 
around so that the myriexpress.h header wasn't included in the same
  header as the interface declarations

This commit was SVN r10817.
 2006-07-14 21:32:03 +00:00
Galen Shipman
6ed255f114 Substantial changes to the CM PML, allows us to have a very thin request for 
all but buffered and persistent requests. Unfortunately we were note able to
reuse the pml_base_request_t as it was just too heavy for our needs. Lots of
code for 2/10 usec ;-) 

This commit was SVN r10810.
 2006-07-14 19:32:26 +00:00
Galen Shipman
9a1221bf7d fix buffered sends (don't use blocking sends!) 
removed inaccurate comment.. 

This commit was SVN r10703.
 2006-07-10 16:11:14 +00:00
Galen Shipman
5085061475 don't call unpack when we received directly into the user buffer.. the 
convertor doesn't handle it properly
continue peeking until we don't get anything else.. 
close the endpoint before closing the library.. 
add a blocking send that uses mx_test .. 

This commit was SVN r10684.
 2006-07-06 19:54:13 +00:00
Brian Barrett
ef6b7e170f * make mtl datatype wrapper code inline functions 
This commit was SVN r10678.
 2006-07-06 15:58:07 +00:00
Galen Shipman
c933c0f65f unpack the length actually received, not the length posted.. 
This commit was SVN r10668.
 2006-07-05 22:16:46 +00:00
Brian Barrett
47725c9b02 * Add new PML (CM) and network drivers (MTL) for high speed 
interconnects that provide matching logic in the library.
  Currently includes support for MX and some support for
  Portals
* Fix overuse of proc_pml pointer on the ompi_proc structuer, 
  splitting into proc_pml for pml data and proc_bml for
  the BML endpoint data
* bug fixes in bsend init code, which wasn't being used by
  the OB1 or DR PMLs...

This commit was SVN r10642.
 2006-07-04 01:20:20 +00:00