1
1
Граф коммитов

5401 Коммитов

Автор SHA1 Сообщение Дата
Brian Barrett
acd652a7ac * have rsh setup opal_progress so that call_yield is only called if the nodes
are oversubscribed (based on information from ras and current data in gpr)

This commit was SVN r6941.
2005-08-19 18:56:44 +00:00
Brian Barrett
0a07341c40 * disconnect if an error occurs after we connected
This commit was SVN r6940.
2005-08-19 18:10:37 +00:00
Brian Barrett
e737bba753 * version of the tm pls that uses the proxy orteds, avoiding all the nasty
multi-client issues the old version had.  Also, ignore the NULL iof
   component, since we shouldn't use it when using the proxy orteds

This commit was SVN r6939.
2005-08-19 16:49:59 +00:00
Brian Barrett
80f27b5d87 * fix some bit rot in tm pls/ras
* remove src/ directory for tm pls/ras

This commit was SVN r6937.
2005-08-19 14:46:11 +00:00
Jeff Squyres
387dacef66 Clarify BProc notes
This commit was SVN r6936.
2005-08-19 14:30:38 +00:00
Jeff Squyres
1f89200c67 Properly cast and remove compiler warning.
This commit was SVN r6935.
2005-08-19 12:20:24 +00:00
Rainer Keller
145cb07562 - Helpful debug code by George.
This commit was SVN r6934.
2005-08-19 10:39:34 +00:00
Rainer Keller
605e13baf4 - Just a comment to circumvent warning in check.
This commit was SVN r6933.
2005-08-19 10:38:08 +00:00
Rainer Keller
202dac7ace - First check for a generated atomic-asm.s file in the Builddir,
iff not availble use the top_srcdir-one.

This commit was SVN r6932.
2005-08-19 10:20:33 +00:00
Brian Barrett
4476212aa7 * check for no components found case and return selection error in that
situation

This commit was SVN r6931.
2005-08-18 21:46:15 +00:00
Tim Woodall
ab4aac2c14 stubs
This commit was SVN r6930.
2005-08-18 19:46:42 +00:00
Jeff Squyres
9b7c1ca97a Forgot to include the check for 0 in the Allreduce case (already got
Reduce and Reduce_scatter)

This commit was SVN r6929.
2005-08-18 19:39:45 +00:00
Galen Shipman
afdfa70f73 Added support for openib RDMA READ.. note that performance is currently an
issue so PUT is default.. We are determining if this is an openib issue or a
btl issue as we have seen performance increases on mvapi. 

This commit was SVN r6928.
2005-08-18 17:08:27 +00:00
Tim Woodall
166ecc9544 removed dependency on ob1
This commit was SVN r6927.
2005-08-18 17:06:59 +00:00
Tim Woodall
95bf77fc8d corrections for rdma pipeline
This commit was SVN r6926.
2005-08-18 17:06:35 +00:00
Brian Barrett
10339416e4 * add C++ mangling protection
This commit was SVN r6925.
2005-08-18 15:48:59 +00:00
Brian Barrett
4da11a4851 * add simple C++ test to make sure that new/delete trigger the correct
callbacks

This commit was SVN r6924.
2005-08-18 15:45:09 +00:00
Brian Barrett
331a971b13 * fix mis-named constant
This commit was SVN r6923.
2005-08-18 15:19:17 +00:00
Brian Barrett
dfdb5dc12a * high resolution, low latency timers for a number of platforms, plus mods
to opal_progress() to use the timers instead of a tick count for deciding
  whether to call the event loop or not.  Currently supported platforms are:

     - solaris (x86 / sparc)
     - Linux (x86 / x86_64 / IA64)
     - Mac OS X (x86 / Power PC)

This commit was SVN r6922.
2005-08-18 05:34:22 +00:00
Brian Barrett
1134d9b7d7 * remove old, broken, horrible hack for doing memory intercepts on Darwin
* Add memory intercept routines for Darwin using the official Darwin
  API (thanks to Drew Gallatin from Myricom for pointing me to some
  information from Apple engineers about how to make this work)
* add debugging output to functionality test

This commit was SVN r6920.
2005-08-18 02:59:02 +00:00
Jeff Squyres
a9777f902a Move all the ptl sm source code out of src (the Makefile.am treated it
that way anway) so that we can properly "make distcheck" easier.  Fix
the rule for making the sym link to ptl_sm_send_alternate.c so that
"distcheck" works.

This commit was SVN r6919.
2005-08-18 00:09:51 +00:00
Galen Shipman
589d53e828 only poll high priority qp 1 per progress.
This commit was SVN r6918.
2005-08-17 22:52:56 +00:00
Brian Barrett
4b3969d2ab * add header file so that this compiles again
This commit was SVN r6917.
2005-08-17 22:44:13 +00:00
Tim Woodall
f274f524ab - added get based protocol (if supported by btl) for pre-registered memory
- removed 8 bytes from the majority of the pml headers 

This commit was SVN r6916.
2005-08-17 18:23:38 +00:00
Jeff Squyres
e8c103ac1f Make MPI_REDUCE and MPI_REDUCE_SCATTER return MPI_SUCCESS immediately
if the count/sum of counts is 0.  This is technically in violation of
the MPI-1 standard, but...  :-(

This commit was SVN r6914.
2005-08-17 15:34:33 +00:00
Rainer Keller
6a7bd04dec - Allow patches to be installed prior to the config.
Currently necessary to install the --prefix-patches for ORTE.
  Used in ompi_cronjob.sh

This commit was SVN r6913.
2005-08-17 14:57:47 +00:00
Jeff Squyres
574a666fbe - Rename the test compile directory to be "test-compile", just to make
it plainly obvious
- Unlink source files when done with them
- Close the perl file handles after writing test source files so that
  stale NFS handles aren't left around when building on networked file
  systems (duh!)

This commit was SVN r6912.
2005-08-17 12:41:21 +00:00
Jeff Squyres
da4d968c1b Because the data structures are so confusing -- add some quick-n-dirty
slides showing what is pointing to what.

This commit was SVN r6910.
2005-08-16 21:56:48 +00:00
Jeff Squyres
a7058bbe43 Be friendly -- if paffinity is requested and we are unable to set it
(for any reason), print a friendly warning message.

This commit was SVN r6907.
2005-08-16 17:18:56 +00:00
Jeff Squyres
5e5fd5a8f2 The fork pls now checks the total number of processes to be launched
against the total number of processors.  If not oversubscribing, emit
the MCA environment variable mpi_paffinity_processor with the
processor number to bind the process to.  This parameter is picked up
during MPI_Init (i.e., ompi_mpi_init()) and used to bind the process,
but currently iif the MCA param mpi_paffinity_alone is set to a
nonzero value (i.e., the user asks for it).

This commit was SVN r6906.
2005-08-16 16:23:20 +00:00
Jeff Squyres
409b9e73b2 Remove debugging printf. Doh!
This commit was SVN r6905.
2005-08-16 16:20:55 +00:00
Jeff Squyres
f5cc86fa07 - Add MCA param mpi_paffinity_alone, which, when nonzero, will assume
that this ORTE job is the only one on the nodes involved, and if
  told what processors to assign the processes to, will bind MPI
  processes to specific processors.
- Convert #include's to new style
- Convert some <tab>'s to spaces

This commit was SVN r6904.
2005-08-16 16:17:52 +00:00
Jeff Squyres
12e6a10197 Enable the paffinity framework during opal_init()
This commit was SVN r6903.
2005-08-16 16:15:53 +00:00
Jeff Squyres
a00e37f516 - Clarification of docs
- Minor fixes in the wrappers

This commit was SVN r6902.
2005-08-16 16:11:32 +00:00
Brian Barrett
f6a64706ad Fix some poorly choosen constants in the XGrid PLS
This commit was SVN r6901.
2005-08-16 16:07:29 +00:00
Jeff Squyres
933fcfe4fd Make the function names obey the prefix rule (whoops)
This commit was SVN r6900.
2005-08-16 14:40:54 +00:00
Galen Shipman
ee6999fa90 typo in threaded build..
This commit was SVN r6898.
2005-08-16 13:22:08 +00:00
Jeff Squyres
dfb6c854ab - Convert to new #include file scheme
- Ensure opal_output*() is properly declared

This commit was SVN r6896.
2005-08-16 10:50:09 +00:00
Brian Barrett
a32a64f0a6 * add the ability to only allow one component from a given framework
be compiled, as well as the ability for components to prioritize
  the order in which they should be compiled so that the "right"
  first one is selected.  Make autogen.sh do all the hard sorting
  work, so that the m4_defined lists of components are in the
  prioritized order.  From there, it's just forcing components
  to fail after the first one succeeds (if we are in the "only one
  compnent can build" mode).

This commit was SVN r6895.
2005-08-16 05:21:34 +00:00
Jeff Squyres
d9e68c1f36 - Fixes for linux paffinity component, to be able to correctly handle
all the different cases (curses to RedHat for continually changing
  the glibc API!)
- Minor fixes for the Solaris paffinity component

This commit was SVN r6894.
2005-08-16 00:11:41 +00:00
Tim Woodall
dc3e8bf041 ack is not required if we are going to do rdma remainder of the message
This commit was SVN r6893.
2005-08-15 22:30:52 +00:00
Tim Woodall
adce576d5e corrections for eager send of data in place
This commit was SVN r6892.
2005-08-15 21:31:52 +00:00
Tim Woodall
9a094ee3b4 - return corrected size
- set send inplace for eager send

This commit was SVN r6891.
2005-08-15 21:30:47 +00:00
Galen Shipman
f248db3789 misc fixes, changes to support multiple mvapi btl's
This commit was SVN r6890.
2005-08-15 19:39:56 +00:00
Jeff Squyres
0dd5b5514c Rename opal_constants.h -> constants.h
This commit was SVN r6889.
2005-08-15 18:55:37 +00:00
Jeff Squyres
cce0950df7 - change a bunch of OMPI_* constants or ORTE_* equivalents
- change the framework opens to [mostly] use the new MCA param API
- properly pass in framework debug output streams to the
  mca_base_component_open() function

This commit was SVN r6888.
2005-08-15 18:25:35 +00:00
George Bosilca
cdee9045c1 ISO C90 forbids mixed declarations and code ...
This commit was SVN r6887.
2005-08-15 17:56:04 +00:00
Jeff Squyres
b080d7360c Remove useless compiler warning
This commit was SVN r6886.
2005-08-15 16:48:43 +00:00
Jeff Squyres
2c7d5009cc Tiny fix (is it enough for Tim?) -- part of a larger set that will be
coming soon (probably tonight).

This commit was SVN r6885.
2005-08-15 16:24:09 +00:00
Jeff Squyres
dc66360faa This component is now open for business.
This commit was SVN r6884.
2005-08-15 15:36:07 +00:00