ports/openmpi
ports/openmpi
1
1
Форкнуть 0
Код Релизы Активность
6 292 коммитов 3 Ветки 0 Теги 130 MiB
bc2e80cbc6
Граф коммитов

799 Коммитов

Автор SHA1 Сообщение Дата
Edgar Gabriel
84c070fc0f get rid of the different modes how to store the colorarray for now. Might be 
reintroduced later as an optimization.

This commit was SVN r7762.
 2005-10-14 18:11:21 +00:00
Edgar Gabriel
6d14440972 checkpoint for moving again to another machine. major rewrite to clean 
up internal interfaces in progress.

This commit was SVN r7761.
 2005-10-14 17:41:44 +00:00
Edgar Gabriel
770aeaf97b modifications towards adding new local-leader communicators. 
This commit was SVN r7760.
 2005-10-14 12:18:29 +00:00
Graham Fagg
636b42afff handle non existant recv buf in reduce for non root processes 
(basic allreduce does this for mpi_in_place case)

This commit was SVN r7759.
 2005-10-14 00:00:37 +00:00
Graham Fagg
61b8218d76 MPI_IN_PLACE fix for reduce. 
(actually a work around for an optimisation in the reduce for not saving ops on the first recv of each segment)
Minor change in topo.

This commit was SVN r7758.
 2005-10-13 23:38:21 +00:00
Edgar Gabriel
48f2563b4c checkpoint. Moving to another machine. 
This commit was SVN r7757.
 2005-10-13 20:04:26 +00:00
Edgar Gabriel
4b05359b16 minor fixes when freeing the component 
This commit was SVN r7756.
 2005-10-13 18:22:16 +00:00
Edgar Gabriel
0a5a346bbb first cut on the reduce operation. 
This commit was SVN r7755.
 2005-10-13 17:58:13 +00:00
Edgar Gabriel
30af775d40 further fixes. The first hierarchical MPI_Bcast works! Its just ~ 100 times slower then basic at the moment :-) 
This commit was SVN r7754.
 2005-10-13 17:34:42 +00:00
Edgar Gabriel
460b5cb840 further corrections to the hierarchy detection algorithms. It seems to work now as far as my tests show... 
This commit was SVN r7753.
 2005-10-13 16:21:13 +00:00
Edgar Gabriel
f5d16419b2 fix in the logic regarding protocol detection. 
This commit was SVN r7749.
 2005-10-13 15:07:35 +00:00
Edgar Gabriel
5d7fbd9d2e minor change in bml_r2_add_procs: the memory for the bml_endpoints structure 
has to be allocated outside of the routine. Thus, the update version of pml/ob1/oml_ob1.c

This commit was SVN r7739.
 2005-10-12 20:59:25 +00:00
Edgar Gabriel
3e5ad3e681 Updates 
This commit was SVN r7738.
 2005-10-12 20:56:29 +00:00
Tim Woodall
22f460bdc5 merge in changes from release branch 
This commit was SVN r7737.
 2005-10-12 20:24:43 +00:00
Tim Woodall
6da9561ea8 merge in correction from v1.0 
This commit was SVN r7732.
 2005-10-12 16:40:52 +00:00
Tim Woodall
d859855dea merge in changes from 1.0 
This commit was SVN r7728.
 2005-10-12 15:54:35 +00:00
Jeff Squyres
727a2cf8b2 Correct a few #if issues that George identified in a code review 
This commit was SVN r7724.
 2005-10-12 13:19:46 +00:00
Jeff Squyres
c568936a7c Add support for MPI_OP_SUM, PROD, REPLACE with MPI_DOUBLE_COMPLEX. 
Need to consult with George -- might also need to add support for
complex types on floats or long doubles...

This commit was SVN r7716.
 2005-10-12 02:31:28 +00:00
George Bosilca
cb7b401ca8 Correct the send-recv operation. 
This commit was SVN r7709.
 2005-10-11 22:41:08 +00:00
Edgar Gabriel
25518b63c5 first version of coll_hierarch which does not crash the rest of the 
library as long as its not selected :-)

This commit was SVN r7707.
 2005-10-11 22:05:24 +00:00
Josh Hursey
ef51608a81 fix compiler warning 
This commit was SVN r7706.
 2005-10-11 22:03:21 +00:00
Edgar Gabriel
0675c22dab updating with Jeff's help to the recent autogen/configure system 
This commit was SVN r7705.
 2005-10-11 21:50:16 +00:00
Edgar Gabriel
7b07dbc163 another round of fixes. Unfortunatly, I also have to provide a trivial 
version of reduce and gather to make all this work....

This commit was SVN r7702.
 2005-10-11 21:26:07 +00:00
Tim Woodall
4a71621410 merge in scheduling changes from release branch 
This commit was SVN r7699.
 2005-10-11 20:41:51 +00:00
Edgar Gabriel
c8adc2e65e coding around the collective operations 
This commit was SVN r7698.
 2005-10-11 20:34:17 +00:00
Edgar Gabriel
083d0b9630 Checkpoint: most of the coding should be done for the basic 
infrastructure.

This commit was SVN r7696.
 2005-10-11 19:45:21 +00:00
Graham Fagg
607bdf51b6 Last Cleanup BEFORE adding last two methods and final cross over points. 
- new mca param calls
- move printfs to OPAL_OUTPUT

This commit was SVN r7692.
 2005-10-11 18:51:03 +00:00
Edgar Gabriel
b42d4ac780 Checkpoint: 
- update the hierarch stuff to use btl's instead of ptl's
- start the new logic regarding how to handle local leader communicators

This commit was SVN r7691.
 2005-10-11 17:29:59 +00:00
Galen Shipman
23cbac25c8 lower default free list sizes.. 
This commit was SVN r7676.
 2005-10-09 18:15:12 +00:00
Galen Shipman
fb19cc4177 compiler warning fixes.. 
This commit was SVN r7661.
 2005-10-07 17:38:34 +00:00
Jeff Squyres
b22fab2826 Fix for a bug Galen noticed yesterday -- make the shared memory only 
be allocated the first time a sm coll is selected for a communicator,
not before.

This commit was SVN r7647.
 2005-10-06 13:17:27 +00:00
George Bosilca
1fe18814da Decrease the default length for the first fragment. 
This commit was SVN r7643.
 2005-10-06 00:05:01 +00:00
George Bosilca
0f04132b13 mx_connect in the MX documentation is supposed to take a timeout in seconds. However, in real life it seems that the timeout should be in micro-second. 
This commit was SVN r7642.
 2005-10-06 00:04:27 +00:00
Brian Barrett
b7ef094766 * the cid in the header is only 16 bits, so limit our max cid to what can fit in there. 
This commit was SVN r7639.
 2005-10-05 15:43:28 +00:00
Jeff Squyres
83b5a675f9 Don't automatically take the first entry off the selected component 
list; be sure to check its priority against the basic component and
take the one with the higher priority.

This commit was SVN r7621.
 2005-10-04 17:09:45 +00:00
George Bosilca
967cd1be32 Make the datatype compile on solaris. 
Remove some warnings ...

This commit was SVN r7619.
 2005-10-04 15:45:18 +00:00
Jeff Squyres
b17c4334c4 - Remove all vestigates of using the built-in mcb_tree from the 
reduce_inorder() function -- we don't use the tree at all.
- Add more relevant "volatile"'s for the control buffers in the
  fragment mpool (and associated casts where necessary)

This commit was SVN r7616.
 2005-10-04 14:52:59 +00:00
George Bosilca
9a67831ba3 Alway call the memory allocation function with the correct type for the first argument. The problem is 
that on some OSes the iovec struct is not POSIX complian, the iov_len is not a size_t but simply an int.
This patch, add a local variable (type size_t) to use with the memory allocation function, and then put
back the value in the iov_len field.

This commit was SVN r7615.
 2005-10-04 14:44:59 +00:00
George Bosilca
059c802094 Correct a small typo 
This commit was SVN r7614.
 2005-10-04 14:42:37 +00:00
Jeff Squyres
94bab558dd Put in a check to ensure the root is valid (all other rooted 
operations have this; we somehow missed this for intracomms on reduce,
and it bit me this morning ;-) )

This commit was SVN r7612.
 2005-10-04 14:38:17 +00:00
Tim Woodall
3b4a134a24 - removed unused define 
- correct free to release registration rather than retain it

This commit was SVN r7611.
 2005-10-04 14:33:26 +00:00
George Bosilca
f8355ec104 Cast the right side member to void* before assignment. 
This commit was SVN r7608.
 2005-10-04 12:37:23 +00:00
George Bosilca
6b3d02b514 Warning cleanups. On some OSes the iov_base member of the iovec structure is defined as an void * when 
on others as an char*. Thus the right side of all assignment should be explicitly casted to an void* in
order to avoid any casting complaints from the compilers.

This commit was SVN r7607.
 2005-10-04 12:36:07 +00:00
George Bosilca
3453a6c0e9 Remove some compiler warnings about unused variables 
Correctly define the 64 bits constants.
Some minor cleanups.

This commit was SVN r7606.
 2005-10-04 12:29:51 +00:00
George Bosilca
492c0e59dc Correct the casting type and remove some useless output (already commented out). 
This commit was SVN r7605.
 2005-10-04 12:28:47 +00:00
Tim Woodall
c05ef28f6e - added routine to ompi_pointer_array to remove array contents 
- corrected memory hook callback to catch all allocations (need to optimize this)
- don't attempt to consolidate allocations

This commit was SVN r7600.
 2005-10-03 23:29:26 +00:00
Jeff Squyres
c7fe54ba44 - Remove some silly compiler warnings 
- Move the "process 0" logic out of the main loop in reduce to make
  the code a bit less complex (at the price of slight code
  duplication, but it iss now significantly easier to read)
- Fix problem with uniquenes guarantee in the bootstrap mpool -- using
  the CID alone was not sufficient enough to guarantee uniquenes; now
  use (CID, rank 0 process name) tuple to check for uniqueness
- Made a few debugging help changes in coll_sm.h; especially helps
  debugging on uniprocessors

This commit was SVN r7599.
 2005-10-03 21:34:58 +00:00
Jeff Squyres
2cedfeec53 - Eliminate some unused base globals 
- Move one base global to the basic component and make it an MCA
  parameter 
- Convert the basic component to use the new MCA param API

This commit was SVN r7598.
 2005-10-03 21:07:42 +00:00
Jeff Squyres
57fb96b018 Clarification of a help message 
This commit was SVN r7597.
 2005-10-03 21:06:13 +00:00
Jeff Squyres
ab099fa8cb Re-indent; real commit with some changes coming shortly. 
This commit was SVN r7596.
 2005-10-03 19:56:39 +00:00
Galen Shipman
eefe0fd04a fix threaded compile 
fix misc warnings 
cleanup posting of receive descriptors 
comment why we retain before deregister in rcache_rb_mru.c 

This commit was SVN r7595.
 2005-10-03 16:35:12 +00:00
Galen Shipman
f46548e691 Add SRQ support to OpenIB btl, removed old mca param - not used.. 
This commit was SVN r7585.
 2005-10-02 18:58:57 +00:00
Jeff Squyres
10064df0e9 Remove compiler warning 
This commit was SVN r7578.
 2005-10-02 10:43:53 +00:00
Jeff Squyres
84feccd3d5 This is something I forgot to commit from long ago -- already 
discussed and cleared with Edgar.

Ensure that only processes who will be in the new communicator call
the coll selection function.  It is pointless (and Bad in some cases)
for processes who are not in the new communicator to try to select a
coll module for the new communicator.

This commit was SVN r7573.
 2005-10-01 11:57:17 +00:00
Jeff Squyres
37fc944b01 Use the right number of segments per in-use flag when calculating 
offsets.

This commit was SVN r7571.
 2005-09-30 23:12:23 +00:00
Galen Shipman
67d38b7896 Add multi-nic support to openib 
Fix connection establishment race in openib 
Other misc 

This commit was SVN r7570.
 2005-09-30 22:58:09 +00:00
Jeff Squyres
934caaf449 Fix at least one segv; use the right number of segments (i.e., the 
number o segments in the fragment pool, not in the bootstrap pool)

This commit was SVN r7565.
 2005-09-30 18:01:15 +00:00
Brian Barrett
db872a0fbb * check that return from ibv_get_devices isn't NULL before calling dlist_start(). 
On thor, if IB is down, we get NULL back from ibv_get_devices(), which then
  caused segfaults in dlist_start().
* Pretty-print error message if no HCAs found

This commit was SVN r7557.
 2005-09-30 14:58:59 +00:00
Jeff Squyres
fcef1774d5 Per advice from Ralf W., change the pkgdata declarations in 
Makefile.am's to be a *slightly* more correct (and, more importantly,
less error-prone) construct.

This commit was SVN r7554.
 2005-09-30 13:32:39 +00:00
Jeff Squyres
80b7deb4d7 Add in EXTRA_DIST to get helpfile in tarballs 
This commit was SVN r7553.
 2005-09-30 10:25:04 +00:00
Brian Barrett
7b20370306 * pretty-print an error message if a btl component loads but can't find 
any NICs to use
* Make mvapi, gm, and mx components all publish information, even if there
  are no NICs available so that modex_recv doesn't hang.  If there are no
  NICs available, don't set the reachable bit, but don't do anything
  to fail.  This unfortunately doesn't cover the hangs that will result if
  different procs load different sets of components, but it's a start

This commit was SVN r7550.
 2005-09-30 04:39:44 +00:00
Brian Barrett
a77c908496 * the last of the tuning params for portals 
This commit was SVN r7548.
 2005-09-30 04:05:31 +00:00
Galen Shipman
8239e635b9 fix misc warnings, cleanup macro.. 
This commit was SVN r7547.
 2005-09-30 03:13:51 +00:00
Galen Shipman
05e6e51fec re-reg from min of bases and max of bounds 
add byte counting for total registered memory 

This commit was SVN r7546.
 2005-09-29 21:28:54 +00:00
Jeff Squyres
bc181d7130 Remove the .ompi_ignore so that everyone starts compiling this, but 
lower the default priority to 0 so that it's not active unless you
specifically ask for it (this component needs more testing by people
other than me before we unleash it on the public).

This commit was SVN r7545.
 2005-09-29 18:05:47 +00:00
Jeff Squyres
d4b7618db7 Comment out what seems to be a debugging output. Will confirm with 
George.

This commit was SVN r7544.
 2005-09-29 16:39:27 +00:00
Brian Barrett
997644af31 * There are now two forms of ibv_create_cq, one with 3 params and one with 5. 
Try to detect which form this version of Open IB uses, defaulting to the 5
  version if we can't figure it out (the new version has 5 params)
* Only add -lcm if it exists on the system - some versions of Open IB
  apparently don't need it.

This commit was SVN r7542.
 2005-09-29 13:35:57 +00:00
Josh Hursey
e825b4522f Upon further investigation the fix in r7537 was an anomoly of zero'ing out the 
bits to expose the low bits being set. We were casting from a size_t to a void*
which is not good when working with big endian machines.

This fix makes MPI 2 dynamics work on PPC 64 (tested with a Linux OS).

This commit was SVN r7538.

The following SVN revision numbers were found above:
  r7537 --> open-mpi/ompi@fd45714c03
 2005-09-28 23:50:42 +00:00
Josh Hursey
fd45714c03 For some reason we have to initialize this variable or bad things happen in the 
comm->c_coll.coll_bcast of the rnamebuflen.

This fixes the threaded MPI 2 Dynamics stuff. Should be working great now! Yay!

This commit was SVN r7537.
 2005-09-28 22:30:41 +00:00
Galen Shipman
3ded88a3c0 use addr +size -1 instead of base->addr as base->addr is down_aligned. 
This commit was SVN r7536.
 2005-09-28 20:19:33 +00:00
Galen Shipman
26a74d42fa release, not retain on gm_free 
This commit was SVN r7535.
 2005-09-28 20:18:52 +00:00
Edgar Gabriel
67dd52efb1 making the allreduce and reduce_scatter tests pass as well 
This commit was SVN r7532.
 2005-09-28 15:12:05 +00:00
Josh Hursey
75419313f7 check the return code and do something reasonable, instead of progressing and hanging on error 
This commit was SVN r7531.
 2005-09-28 06:13:51 +00:00
Galen Shipman
c1f5543f62 need to call mpool_release on all registrations obtained in the pml. 
sanity checks 

This commit was SVN r7530.
 2005-09-28 04:49:40 +00:00
Galen Shipman
b9b78f8f5d modify rcache_rb to find registrations in the middle of a base and bound 
This commit was SVN r7528.
 2005-09-28 02:11:35 +00:00
Edgar Gabriel
dbbbd416df fixing MPI_IN_PLACE for the log-reduce algorithm. 
This commit was SVN r7526.
 2005-09-27 21:51:55 +00:00
Galen Shipman
0fc17cedee change order of ops on register 
This commit was SVN r7525.
 2005-09-27 21:43:41 +00:00
Jeff Squyres
285ded5655 - Ensure to have !initialized || finalized test *first* 
- If we have an NS error, don't return an error -- this function's
  purpose is to abort :-)
- s/abort()/exit(1)/ so that we don't drop massive corefiles

This commit was SVN r7524.
 2005-09-27 20:26:38 +00:00
Galen Shipman
09e67ce4fd fix off by one on up_align_addr, use base and bound instead of base_align and 
bound_align.. 

This commit was SVN r7521.
 2005-09-27 18:10:44 +00:00
Galen Shipman
af04b3e1ab fix warnings.. 
This commit was SVN r7515.
 2005-09-27 14:23:51 +00:00
Brian Barrett
80ac5c2efd * there are now two upcoming points where we want to release a version with 
a random string of characters as part of the version number (the really
  soon to happen 1.0lanl release and the 1.1sc2005 release that we've
  talked about).  So rather than having alpha and beta fields that must
  be numeric values, have a general field that can be any alphanumeric
  value.

This commit was SVN r7511.
 2005-09-27 02:06:05 +00:00
Galen Shipman
3c97b3f722 Modified the registration to include a base_align and bound_align for 
searching the tree. Modified the memory callback to search the tree at each
page boundary for registrations. This is necessary as an application may
malloc memory and send out of any portion of that memory, even discontiguous
regions. 

This commit was SVN r7510.
 2005-09-27 02:01:21 +00:00
Brian Barrett
d9e80d8f2a * increase size of event queue for receives - it was too small to be useful 
on a reasonably sized machine
* if no mpool exists, don't try to malloc out an array of 0 bytes

This commit was SVN r7507.
 2005-09-25 17:04:03 +00:00
Galen Shipman
384c472c94 reset ompi_pointer_array in mca_rcache_rb_find otherwise you might use an old 
registration by accident.. 

This commit was SVN r7506.
 2005-09-24 20:48:14 +00:00
Galen Shipman
02ce7a176e had that backwards.. 
This commit was SVN r7504.
 2005-09-24 16:58:07 +00:00
Galen Shipman
d1246be47e should be strictly > or < 
This commit was SVN r7503.
 2005-09-24 16:45:34 +00:00
Galen Shipman
c53d51778a fix for warnings.. 
This commit was SVN r7501.
 2005-09-24 15:08:28 +00:00
Galen Shipman
9fe5844071 decrement ref count on removal of registration from mru and tree. 
add misc asserts to check for proper reference counting. 

ugly hack 1 -- use mallopt to never release memory ala sbrk - this is
commented out in mca_btl_mvapi_component_init

ugly hack 2 -- test registrations comming out of the tree via rcache_find, for
an unknown reason the tree is returning registrations where the address is not
within the base or bound of the registration. If this happens, we return
NULL. 

comment out code to enable mem hooks if leave_pinned is set, note we can do
this via an mca param and will default it to leave_pinned with mem_hooks when
we iron out these issues. 

I am adding a unit test for the rcache. Note that we have a unit test for the
rb tree but the compare function is significantly different than that used for
registrations. After we have tracked down the issues with rcache_rb we will
remove the above hacks. 

This commit was SVN r7499.
 2005-09-24 00:24:49 +00:00
Brian Barrett
50dc5499b4 * fix some remaining --with-btl-portals configure issues 
This commit was SVN r7498.
 2005-09-24 00:11:40 +00:00
Brian Barrett
0d68728b94 * add some more debugging output for send fragment issue to figure out why 
Red Storm is complaining about invalid memory pointer (need to go back
  to Linux and look at this with valgrind)
* Turn off send in place for now, so I can run the tests on RS and see if
  everything else is ok

This commit was SVN r7497.
 2005-09-23 19:30:54 +00:00
Brian Barrett
07b0b8c943 * add some useful debugging output 
* fix dumb bug in btl_portals_get where I using the dest descriptor key instead
  of the source descriptor key for the match bits, resulting in a PtlGet() with
  the wrong match bits

This commit was SVN r7496.
 2005-09-23 15:30:18 +00:00
Tim Woodall
604c9d1002 bump the default cache size 
This commit was SVN r7492.
 2005-09-22 19:37:42 +00:00
Tim Woodall
848f12e7fd if mpi_leave_pinned is enabled - force malloc hooks 
This commit was SVN r7491.
 2005-09-22 17:27:56 +00:00
Tim Woodall
aceab46c5f use MPI_Alloc_mem/MPI_Free_mem for internally allocated buffers 
This commit was SVN r7487.
 2005-09-22 16:43:17 +00:00
Tim Woodall
147716c249 added hostname to error output 
This commit was SVN r7486.
 2005-09-22 16:41:34 +00:00
Tim Woodall
b404581293 local variables/objects (regs) must be initialized/constructed 
This commit was SVN r7485.
 2005-09-22 16:18:23 +00:00
Tim Woodall
7acf0a6bdb corrections for MPI_Free_mem 
This commit was SVN r7481.
 2005-09-22 15:47:33 +00:00
Andrew Friedley
555ae37255 Add lib{opal,orte,mpi}.la to appropriate LIBADD's, some whitespace cleanup as well. 
This commit was SVN r7477.
 2005-09-22 12:28:54 +00:00
Tim Woodall
da1e4a1292 - since we append new registrations to the end of the list, 
need to remove old ones from the front
- call deregister to actually remove items from the cache/mru
list and deregister the memory (if not being used)

This commit was SVN r7476.
 2005-09-21 23:25:16 +00:00
Tim Woodall
9791c066e8 dont attempt to pin the receive buffer if data has 
already been received

This commit was SVN r7475.
 2005-09-21 23:23:47 +00:00
Tim Woodall
1b7b220089 cleanup refcount 
This commit was SVN r7462.
 2005-09-21 20:04:56 +00:00
Tim Woodall
a74ca0062a reductions to initial memory footprint 
This commit was SVN r7455.
 2005-09-21 19:10:56 +00:00
Galen Shipman
4296e723c9 default free_lists to smaller size.. 
This commit was SVN r7454.
 2005-09-21 18:55:07 +00:00
Galen Shipman
96ab5a6bd3 we can be in WAITING_ACK state without a race if the OOB ack is "slower" than 
the scheduling of queued IB send operations. 

This commit was SVN r7452.
 2005-09-21 16:47:08 +00:00
Tim Woodall
782e5b21cc cleanup 
This commit was SVN r7451.
 2005-09-21 15:34:45 +00:00
Tim Woodall
a49a442fe4 cleanup refcount logic 
This commit was SVN r7450.
 2005-09-21 15:32:27 +00:00
Tim Woodall
0ee34051f8 debug asserts 
This commit was SVN r7449.
 2005-09-21 15:30:17 +00:00
Tim Woodall
1b73d3856e possible race condition - set endpoint state before sending connect ack 
This commit was SVN r7448.
 2005-09-20 21:03:55 +00:00
David Daniel
e4985c2a07 Moving totalview spin to the very end of mpi_init 
This commit was SVN r7444.
 2005-09-20 15:22:15 +00:00
Brian Barrett
fd9901f683 * shell of a portals PML, properly ompi_ignored for most of the world... 
This commit was SVN r7437.
 2005-09-20 08:07:08 +00:00
Brian Barrett
d81726833e * Add memory barriers for shared memory. Rich and I think we got them 
all and the Intel tests pass slightly oversubscribed.

This commit was SVN r7431.
 2005-09-19 16:28:25 +00:00
Tim Woodall
aeb5bc3f57 still need to cleanup/revise the template for mpool changes 
This commit was SVN r7425.
 2005-09-19 14:34:24 +00:00
George Bosilca
b70230858b Correct the misnaming problem in the GM PTL. 
This commit was SVN r7424.
 2005-09-19 10:34:06 +00:00
Galen Shipman
6499eb3976 init the return code.. 
This commit was SVN r7423.
 2005-09-18 14:25:30 +00:00
George Bosilca
97673b45d1 Remove the last bad symbol from the GM PTL. 
This commit was SVN r7422.
 2005-09-18 12:52:37 +00:00
George Bosilca
b5cb27c006 The self should use self named files. 
This commit was SVN r7421.
 2005-09-18 12:37:15 +00:00
George Bosilca
a7db1763e2 Cleanups ... 
This commit was SVN r7420.
 2005-09-18 12:34:29 +00:00
Jeff Squyres
d67c31f238 Remove useless compiler warnings. 
This commit was SVN r7418.
 2005-09-17 10:54:48 +00:00
Jeff Squyres
f9a1e14f65 Per suggestion from our friendly Libtool developer friends, add proper 
dependencies for liborte and libompi (i.e., make liborte depend on
libopal, and make libmpi depend on liborte)

This commit was SVN r7417.
 2005-09-17 10:45:46 +00:00
Galen Shipman
b8cb6e1c64 modified mpool module to contain flags - used to determine if the mpool will 
be used in MPI_Alloc_mem operations. Note that we found an interesting bug in
which if memory was allocated by the sm mpool (via mmap) and then registered
via the mvapi mpool, the registration would fail on certain systems. 

Added mca param mpool_base_use_mem_hooks, set to 1 to enable the memory hooks
so that memory is deregistered if the user frees it behind our back. This is
only useful if the mca param mpi_leave_pinned is also set to 1. Otherwise all
registrations are deregistered within the MPI library or via
MPI_Free_buf. After testing we should probably set both mpi_leave_pinned and
mpool_base_use_mem_hooks to default to 1. 

This commit was SVN r7415.
 2005-09-16 22:22:03 +00:00
Brian Barrett
c87babb565 * if start_rank == end_rank, there doesn't seem to be a requirement that 
stride == 1, and the Intel tests explicitly test the case with
  strides != 1, expecting them to work

This commit was SVN r7411.
 2005-09-16 18:44:21 +00:00
Galen Shipman
808b2c1c53 threaded build fix for btl_gm.. 
This commit was SVN r7409.
 2005-09-16 17:18:15 +00:00
Jeff Squyres
2b82224a4f Remove superflous ; (actually causes an error in some cases) 
This commit was SVN r7405.
 2005-09-16 12:27:25 +00:00
Jeff Squyres
10d02b2110 Make sure to copy the right amount out of the temp buffer. 
This commit was SVN r7400.
 2005-09-15 22:06:36 +00:00
Galen Shipman
22dac4983c fix mca_mpool_base_free to use new rcache 
This commit was SVN r7398.
 2005-09-15 21:09:07 +00:00
Jeff Squyres
15d0a95202 - Remove extra whitespace from Makefile.am's from when we removed 
Makefile.options
- Sample in each of the three projects of how to link againt the
  relevant libraries so that when components are loaded into a parent
  process' space, we don't rely on the libopal/liborte/libmpi symbols
  being in the parent's public symbol namespace -- instead,
  dynamically link to the relevant libraries, allowing the dynamic
  linker to pull those libraries in at run-time, if needed

This commit was SVN r7397.
 2005-09-15 20:56:18 +00:00
Jeff Squyres
46bed32fc6 While double-checking the order of SUBDIRS (ensuring dynamic-mca is 
after "."), remove some whitespace left over from when we removed
Makefile.options.

This commit was SVN r7393.
 2005-09-15 20:46:29 +00:00
Tim Woodall
9b96ecf9b1 correction for rdma case - currently rdma entire message 
This commit was SVN r7392.
 2005-09-15 19:36:56 +00:00
Jeff Squyres
3ecfe02b83 - Properly handle MPI_IN_PLACE 
- Return MPI_ERR_ARG, not EINVAL

This commit was SVN r7391.
 2005-09-15 19:33:54 +00:00
Tim Woodall
6ae2ae4d1a - code cleanup 
- correct mpi semantics

This commit was SVN r7390.
 2005-09-15 18:47:59 +00:00
Jeff Squyres
2c1186cd19 Fix up the offsets for the non-root gatherv in the IN_PLACE case. 
This commit was SVN r7389.
 2005-09-15 18:21:18 +00:00
George Bosilca
789d49d613 Activate thread support ... at least at the MPI level. 
This commit was SVN r7386.
 2005-09-15 15:44:12 +00:00
George Bosilca
32cdc551af No more warnings in the GM PTL. 
This commit was SVN r7385.
 2005-09-15 15:29:31 +00:00
Jeff Squyres
7ca22d9416 - Correct to use the right offsets 
- Copy back to the right location in the non-rank-0-IN_PLACE case

This commit was SVN r7384.
 2005-09-15 15:15:23 +00:00
Jeff Squyres
406f0575eb - Remove useless error check 
- Ensure err is set to MPI_SUCCESS on the IN_PLACE case

This commit was SVN r7383.
 2005-09-15 15:14:00 +00:00
Jeff Squyres
cbfb062a7d Fix silly mistake for IN_PLACE handling in scan 
This commit was SVN r7380.
 2005-09-15 12:47:17 +00:00
Jeff Squyres
068b9c72a2 Bunches of changes 
- remove redundant OBJ_CONSTRUCT in bcast
- fix up some macros in coll_sm.h
- check to ensure that if there are too many processes in the
  communicator (i.e., if we couldn't fit a flag for each of them in
  the control segment), then fail selection
- setup the in_use flags properly
- adapt to new mpool API
- first working copy of reduce -- not tree-baed (but still
  NUMA-aware), and only processes in order from process 0 to process
  N-1 -- do not have a tree-based and/or commutative version yet
  (i.e., process the results in whatever order they arrive)

Reduce now passes the new ibm reduce_big.c test.  Woo hoo!  Time to
declare success for the evening (and run the intel test tomorrow).

This commit was SVN r7379.
 2005-09-15 02:18:16 +00:00
George Bosilca
ec621c2145 That was a bad*bad typo. To compute the flags I have to use & not |. 
In the case where the user want to roll-back a convertor we should first rewind the
convertor to the beginning and then move it to the requested position. Otherwise the
logic do roll-back a convertor in an optimized way is really difficult. BTW, in terms of
performances one would prefer to have 2 convertors and play with position inside, than
just having one and roll it back.

This commit was SVN r7378.
 2005-09-14 23:32:18 +00:00
George Bosilca
fdfd966aad On the last position on the conversion stack in the displacement field we should have the 
relative displacement not the absolute one. This relative displacement will be addesd to the
absolute one in order to compute the correct position on the source (or destination) buffer.

This commit was SVN r7377.
 2005-09-14 23:29:27 +00:00
Tim Woodall
31d392af95 correct name 
This commit was SVN r7376.
 2005-09-14 22:35:58 +00:00
Tim Woodall
6ffa947a85 dont try and use convertor if not initialized 
This commit was SVN r7375.
 2005-09-14 22:03:52 +00:00
Tim Woodall
9279e4f882 use sync send to ensure message is received before exiting 
This commit was SVN r7374.
 2005-09-14 21:28:17 +00:00
Tim Woodall
d190e6a315 handle losing a connection 
This commit was SVN r7373.
 2005-09-14 21:27:30 +00:00
Tim Woodall
c25fb5dab0 - fixed issue w/ btl send-in-place option that was affecting tcp 
- reduced size of match header by an additional 4 bytes to 16 bytes
- corrections for buffered send (work in progress)

This commit was SVN r7371.
 2005-09-14 17:08:08 +00:00
Brian Barrett
e98415eb7b * make tree compile on OS X 
This commit was SVN r7370.
 2005-09-14 15:52:42 +00:00
Galen Shipman
8b146cec13 Callback shouldn't be static as it is used in more than one .c file. 
This commit was SVN r7369.
 2005-09-14 13:27:09 +00:00
Jeff Squyres
5365ae84b9 Remove extra variable. Still working with George / Edgar on 
reduce_log_intra().

This commit was SVN r7368.
 2005-09-14 11:52:20 +00:00
Jeff Squyres
5456d3444f - Add missing header files 
- Use new #include file scheme

This commit was SVN r7367.
 2005-09-14 09:37:20 +00:00
Galen Shipman
f0b1ea52bc if all else fails in prepare_src,, pack 
init the rdma_pending list in ob1

This commit was SVN r7366.
 2005-09-14 04:41:33 +00:00
Galen Shipman
940a2494a2 init the find, release, and retain funcs on the mpool, otherwise segfault.. 
This commit was SVN r7365.
 2005-09-14 02:59:31 +00:00
Jeff Squyres
e0c47dd0bc Fix for allreduce in IN_PLACE cases 
This commit was SVN r7364.
 2005-09-14 02:42:32 +00:00
Jeff Squyres
0fcd682c4c MPI-2 7.3.3 description of MPI_Allgatherv is wrong -- can't just have 
all processes call MPI_Gatherv(MPI_IN_PLACE...) because IN_PLACE is
only allowed to be used at the root.  Non-root processes must use
their receive buf as the send buf.

This commit was SVN r7363.
 2005-09-14 02:21:33 +00:00
Galen Shipman
b3db6fc2d1 Add code for mem_free hooks, commented out right now as there are performance 
issues. 
In PML-OB1 - fix for sync-sends of zero length. 

This commit was SVN r7362.
 2005-09-14 02:17:04 +00:00
Jeff Squyres
b8f4f88ccc Ignore sendtype/sendcount if MPI_IN_PLACE 
This commit was SVN r7361.
 2005-09-14 01:44:55 +00:00
Galen Shipman
272153cd47 fix threaded build issues with rcache. 
This commit was SVN r7358.
 2005-09-13 22:06:44 +00:00
George Bosilca
3e35dbee30 Add a helper function to find out if a couple datatype & count represent a contiguous 
memory layout.

This commit was SVN r7357.
 2005-09-13 21:59:34 +00:00
Graham Fagg
0f75381e56 Added various barrier routines: recursive doubling, bruck, double ring, 2proc etc 
all pass tests

This commit was SVN r7355.
 2005-09-13 20:58:42 +00:00
Brian Barrett
1290b8eed2 * some debugging to figure out why get isn't working on RS 
This commit was SVN r7354.
 2005-09-13 20:52:56 +00:00
George Bosilca
a5871ffc23 Report back the correct position when the caller use ompi_convertor_set_position. That should 
fix Tim's problems.

This commit was SVN r7353.
 2005-09-13 20:29:50 +00:00
George Bosilca
ad0c0cdc03 Make the GM btl compile again. There were just some typos. 
This commit was SVN r7352.
 2005-09-13 20:19:21 +00:00
Jeff Squyres
5dca18f903 First cut of handling MPI_IN_PLACE: 
- added relevant logic for everything except
  mca_coll_basic_reduce_log_intra() -- need some help from George /
  Edgar on this one...
- replaced ompi_ddt_sndrcv() with ompi_ddt_copy_content_same_ddt()
  where relevant
- removed some "if (size > 1)" conditionals, because the self coll
  module will always be chosen for collectives where size==1

Waiting for BA's tests to check the validity of this IN_PLACE stuff.
We'll see how it goes!

This commit was SVN r7351.
 2005-09-13 20:06:54 +00:00
Jeff Squyres
bd95f5d474 Arrgh -- check the right argument for IN_PLACE. 
This commit was SVN r7350.
 2005-09-13 19:56:43 +00:00
Jeff Squyres
7c09923751 Updates: 
- Handle MPI_IN_PLACE
- Use ompi_ddt_copy_content_same_ddt() where relevant

This commit was SVN r7349.
 2005-09-13 19:39:49 +00:00
Jeff Squyres
15f5294da4 Ignore another param in the IN_PLACE case 
This commit was SVN r7348.
 2005-09-13 19:15:01 +00:00
Jeff Squyres
c89e8c3712 Fixes for error checking -- properly ignore arguments as they should 
be

This commit was SVN r7347.
 2005-09-13 19:07:13 +00:00
Graham Fagg
9053790973 fixed bruck alltoall bug. now passes ibm tests 
This commit was SVN r7346.
 2005-09-13 18:35:45 +00:00
Jeff Squyres
47a1a2b7ec Arrgh. Compile *before* commit. 
Sorry folks -- stupid typo fixed.

This commit was SVN r7345.
 2005-09-13 18:12:10 +00:00
Jeff Squyres
da87169d17 Add support for MPI_IN_PLACE for the easy operations. 
This commit was SVN r7344.
 2005-09-13 18:02:36 +00:00
Jeff Squyres
ad647b105d Error checking for MPI_IN_PLACE 
This commit was SVN r7343.
 2005-09-13 17:07:03 +00:00
Tim Woodall
ee58631c82 corrections for probe/iprobe 
This commit was SVN r7342.
 2005-09-13 16:45:41 +00:00
Jeff Squyres
bbae6c3b1a Add missing header file 
This commit was SVN r7338.
 2005-09-13 12:19:34 +00:00
Graham Fagg
f354140587 Multiple changes 
-added some alltoall calls (pairwise checked ok, bruck testing)
-changes in use of data hung of communicator
-making sendrecv call a true inline function
-more use ompi_ddt routines

This commit was SVN r7337.
 2005-09-13 04:28:18 +00:00
Tim Woodall
bacc1b9122 hack to request all existing values for jobids other 
than our own - required for mpi2 dynamic processes

This commit was SVN r7336.
 2005-09-13 03:53:53 +00:00
Galen Shipman
39f25428da missing includes, perhaps related to george's work? 
This commit was SVN r7332.
 2005-09-13 02:00:28 +00:00
Galen Shipman
d932cfd342 merge of rcache work into the trunk.. lotsa fun ;-).. 
I regression tested before the merge, I will regression test tonight and
correct issues that might have crept in. 

This commit was SVN r7329.
 2005-09-12 22:28:23 +00:00
Brian Barrett
4c62c356c7 * more missing header file recovery 
This commit was SVN r7328.
 2005-09-12 22:13:09 +00:00
George Bosilca
8308ab42e9 GM depend on the proc.h now. 
This commit was SVN r7327.
 2005-09-12 21:52:44 +00:00
George Bosilca
e3a8489dd0 Replace ompi_proc_t by struct ompi_proc_t to remove all dependencies to proc.h 
This commit was SVN r7326.
 2005-09-12 21:51:56 +00:00
Brian Barrett
88cd561198 * bunch of fixes for Red Storm - missing header files and the like 
This commit was SVN r7325.
 2005-09-12 21:45:58 +00:00
Tim Woodall
304f6254e6 additional btl flags 
This commit was SVN r7324.
 2005-09-12 21:38:31 +00:00
Brian Barrett
15d48945c6 * fix communicator.h so that tree compiles again - needs to know what an ompi_proc_t is 
This commit was SVN r7323.
 2005-09-12 21:34:26 +00:00
George Bosilca
5caeb0295a Correct the includes and some indentation. 
This commit was SVN r7322.
 2005-09-12 20:36:04 +00:00
George Bosilca
948683215b And more fixes ... 
This commit was SVN r7321.
 2005-09-12 20:25:01 +00:00
Brian Barrett
79f7ea6856 * implement btl_put for Portals 
This commit was SVN r7320.
 2005-09-12 20:24:43 +00:00
George Bosilca
c9fb1f32f2 And more dependencies fixes. The big commit will follow shortly. 
This commit was SVN r7319.
 2005-09-12 20:22:59 +00:00
George Bosilca
fd1431bf0f Dependencies fixes. 
This commit was SVN r7318.
 2005-09-12 20:21:53 +00:00
George Bosilca
6eef57c5f2 More dependencies work. 
This commit was SVN r7317.
 2005-09-12 20:20:51 +00:00
George Bosilca
efac21256b Add some more dependencies. 
This commit was SVN r7316.
 2005-09-12 20:19:46 +00:00
George Bosilca
13f4191334 Remove the gm_config.h file as now the GM defines are in the ompi_config.h. 
Add some more header files for the next commit ... stay tuned ...

This commit was SVN r7315.
 2005-09-12 19:55:21 +00:00
George Bosilca
71b1cdb5f5 Cleanup the dependencies between the MPI layer and the rest of the ompi files. 
Modify the includes to start with the root directory (i.e orte/datatype/datatype.h).

This commit was SVN r7306.
 2005-09-12 09:17:44 +00:00
George Bosilca
1b031c153b Last commit to make the threaded case compiling without warnings. Next step try to make it working ... 
Correct the spring of the vpid problem (similar to the one in the SM PTL).

Add one more argument to the MCA_BTL_SM_FIFO_WRITE macro who will get passed down to the 
MCA_BTL_SM_SIGNAL_PEER macro to allow it to have the fifo_fd file descriptor.

This commit was SVN r7305.
 2005-09-11 20:55:22 +00:00
George Bosilca
f0e4cacd4d procs[proc]->proc_name.vpid is a size_t, so try to cast it before using sprintf. 
This commit was SVN r7304.
 2005-09-11 20:52:02 +00:00
George Bosilca
f8d9f6121c Typo correction ... 
This commit was SVN r7303.
 2005-09-11 20:49:27 +00:00
George Bosilca
8b1c28e068 No more warnings for the threaded case compilation. Some of the fields have to be declared 
as volatile to make sure the compiler don't try to optimize them out. 

This commit was SVN r7302.
 2005-09-11 20:48:37 +00:00
George Bosilca
e501768f10 Sleep is defined in the unistd.h file, do not forget to include it on the dependencies list. 
This commit was SVN r7301.
 2005-09-11 20:05:17 +00:00
George Bosilca
c24eb702bb Correctly compute the default sizes for the fragments. 
This commit was SVN r7299.
 2005-09-11 20:02:55 +00:00
George Bosilca
5b1303edef Correctly handle the convertor flags. 
This commit was SVN r7298.
 2005-09-11 18:51:47 +00:00
George Bosilca
6b2d154df3 Add more debug messages to the pack and unpack functions. 
This commit was SVN r7297.
 2005-09-11 18:51:06 +00:00
George Bosilca
95497b156e Cleanup the flags when we clone or copy a convertor. Add safety checks for the set_position function. 
This commit was SVN r7296.
 2005-09-11 18:49:46 +00:00
George Bosilca
df26395a83 Add some mask to cleanup the convertor flags. 
This commit was SVN r7295.
 2005-09-11 18:48:50 +00:00