Timeline


and .

04/09/09:

21:05 Changeset [4246] by thakur
fixed comment
18:00 Ticket #521 (Availability of fortran lib with lowercase symbols with MPICH2 64-bit on ...) created by "Jeyapandian Kottalam" <jkottalam@…>
[…]
17:11 Ticket #520 (mpicc -o test.c deletes test.c) created by "Rajeev Thakur" <thakur@…>
[…]
14:19 Changeset [4245] by jayesh
Check if the local shared mem handle is valid before closing it
14:09 Ticket #519 (weak symbol support on Mac OS X) created by goodell
OS X does support weak symbols, but in classic Apple fashion it uses a …
14:03 Ticket #518 (Hydra: stderr forwarding for failed processes) created by balaji
Hydra does not seem to be forwarding stderr for processes that segfault …
13:41 Ticket #517 (Compiling cpi.c with MPICH2 on SiCortex fails) closed by goodell
fixed: Should be fixed by r4244. -Dave
13:40 Changeset [4244] by goodell
Fix for SiCortex? atomics. Analogous to r43 in the OpenPA repository. No …
12:10 Ticket #517 (Compiling cpi.c with MPICH2 on SiCortex fails) created by jayesh
I get compile time errors when compiling cpi.c with MPICH2 on SiCortex?, …

04/08/09:

09:55 Changeset [4240] by rbutler
Fix minor mis-spelling in 2 print msgs.

04/07/09:

16:28 Changeset [4239] by jayesh
Handle comments and blank lines in the host file
14:24 Changeset [4238] by jayesh
1) Adding HYDU_CALLOC (2) Fixing several mem leaks (3)Also indent code - …

04/04/09:

21:40 Ticket #516 (Nemesis tests hang with PGI compiler) created by "Rajeev Thakur" <thakur@…>
[…]
20:24 Ticket #515 (Mpich2 on MacBook Pro) closed by balaji
worksforme

04/03/09:

20:49 Changeset [4237] by balaji
This takes a first step towards a distributed PMI server. The server is …
16:51 Changeset [4236] by jayesh
Removing creating a new session for persistent proxy. We will add it later …
16:35 Changeset [4235] by jayesh
1) Moved fork and exit functionality to a util function. 2) Removed …
15:53 Ticket #515 (Mpich2 on MacBook Pro) created by Douglas Dommermuth <douglas.dommermuth@…>
[…]
10:02 Changeset [4234] by jayesh
1) By default, the persistent proxies launched by the BSCD now fork a …

04/02/09:

16:05 Changeset [4233] by balaji
Committing in the first draft of PMI-2. This is incorrect in a lot of …
16:02 Ticket #514 (Fortran 90 module support on Windows MPICH2 x86_64) created by "Jeyapandian Kottalam" <jkottalam@…>
[…]
14:31 Changeset [4232] by jayesh
Don't abort if UI/launcher cannot connect to a proxy to shut it down. …
13:12 Ticket #512 (MPD and PGI compilers) closed by goodell
fixed: Rajeev added --disable-totalview to the old nightlies, so this shouldn't …
10:45 Ticket #513 (Is MPICH2 connection failure aware ?) created by "Jayesh Krishna" <jayesh@…>
[…]
09:48 Ticket #512 (MPD and PGI compilers) created by "Rajeev Thakur" <thakur@…>
[…]

04/01/09:

17:06 Ticket #511 (Configure check for C and Fortran compiler compatibility) created by Douglas Dommermuth <DOUGLAS.G.DOMMERMUTH@…>
[…]
15:10 Ticket #510 (Resolved: Re: python error) created by Anil Deane <deane@…>
[…]
10:07 Ticket #509 (Cannot build mpich2-1.0.8p1 (nemesis) with PGI 8.0-4 on Linux x86_64) created by gus@…
[…]
02:49 Changeset [4231] by balaji
This should fix the second socket we weren't closing.
02:37 Changeset [4230] by balaji
This should fix one of the two file descriptors that we are leaking. One …
00:38 Ticket #508 (query on MPI testing fail) created by 汪迪 <otheryou@…>
[…]

03/31/09:

21:12 Changeset [4229] by balaji
We now use threads to connect to different persistent proxies in parallel …
14:37 Ticket #507 (--enable-g option not recognized) created by Hormoz Zarnani <hzarnani@…>
[…]
12:58 Ticket #505 (MPIDI_msg_sz_t versus MPI_Aint in MPIR_Localcopy) closed by goodell
fixed: This should be fixed in r4228. Let us know if you have any other trouble, …
12:56 Changeset [4228] by goodell
Fix for MPIR_Localcopy with non-standard MPI_Aint sizes (ticket #505). …

03/30/09:

19:40 Changeset [4227] by balaji
Bug fixes to the persistent proxy code for the case when the proxy is …
17:15 Ticket #506 (bug in MPIU_Find_local_and_external) closed by goodell
fixed: Joe, this should be fixed by r4226. Sorry for missing the incredibly …
17:13 Changeset [4226] by goodell
MPIU_Find_local_and_external bug fix for node unaware devices (ticket …
15:55 Ticket #506 (bug in MPIU_Find_local_and_external) created by goodell
[…]
15:55 Ticket #505 (MPIDI_msg_sz_t versus MPI_Aint in MPIR_Localcopy) created by goodell
[…]
12:28 Ticket #504 (Need help installing mpich2) created by alahane1@…
[…]
10:37 Changeset [4225] by mercierg
switch to events-based iface for recv + bug fixes
02:00 Changeset [4224] by balaji
Added some place holder files for the communication sub-system. There's no …

03/29/09:

23:43 Changeset [4223] by balaji
Compiler and valgrind warning stomp.
22:57 Changeset [4222] by balaji
Merge the code paths for the runtime and persistent launch modes so that …

03/28/09:

16:38 Changeset [4221] by gropp
Clean out unused error strings (saving them in errnames-old.txt)
16:37 Changeset [4220] by gropp
Added a script to help find unused error strings
16:35 Changeset [4219] by gropp
Removed incorrect fixme and added comment in answer
16:34 Changeset [4218] by gropp
Added additional tests for fault handling
16:34 Changeset [4217] by gropp
Added support for additional mpiexec arguments and for environment …
16:33 Changeset [4216] by gropp
Ensure dtpack frees created datatypes

03/27/09:

18:36 Changeset [4215] by balaji
White space commit.
15:18 Changeset [4214] by jayesh
Hydra job launcher can now boot proxies on multiple hosts
13:49 Ticket #503 (Nemesis on niagara) created by "Rajeev Thakur" <thakur@…>
[…]
13:11 Changeset [4213] by jayesh
# Initial cut of distributed proxies support in hydra - launch/shutdown …
13:02 Changeset [4212] by balaji
Fixing yet another bug with the environment variable passing that got …

03/26/09:

18:04 Changeset [4209] by balaji
Corrections and clean-up to the release notes. Reviewed by thakur and …
15:37 Changeset [4208] by buntinas
Retagging 1.0.8p1
15:36 Changeset [4207] by buntinas
Adding more fixes into p1 patch. Removing old tag.
15:31 Changeset [4206] by buntinas
merged r4143 from 1_0 branch: fixed typo in ia64 atomics
15:30 Changeset [4205] by buntinas
merged r3982 from 1_0 branch: fixed gm compile error reported in ticket …
14:46 Changeset [4204] by balaji
Retagging 1.0.8p1.
14:46 Changeset [4203] by balaji
Updated version number to 1.0.8p1.
14:46 Changeset [4202] by balaji
Forgot to update the version number.
14:45 Changeset [4201] by balaji
Tagging the 1.0.8p1 release.
14:44 Changeset [4200] by balaji
Merging r4199 from the 1.0.x branch.
14:33 Changeset [4199] by balaji
Fix to ROMIO to work on the latest PVFS version. Thanks Dries!
10:16 Changeset [4198] by mercierg
New version of the newmad module
10:10 Changeset [4197] by mercierg
Added definition of COMM_OVERRIDES for newmad module
10:10 Changeset [4196] by mercierg
Removed C++ style comments
10:06 Changeset [4195] by mercierg
Added mx_probe.c file
09:42 Changeset [4194] by buntinas
test commit
09:39 Changeset [4193] by buntinas
test commit
01:07 Changeset [4192] by balaji
Some more reduction in the dependency on MPICH2.
00:41 Changeset [4191] by balaji
Minimize MPI-ism in Hydra.

03/25/09:

21:36 Changeset [4190] by balaji
This includes two bug fixes to the environment passing code: 1. This …
19:37 Ticket #502 (Simultaneous multi-compiler support) created by balaji
We need to verify that applications compiled with different fortran …
18:48 Changeset [4189] by balaji
Added "use warnings;" instead of calling "perl -w" directly, as that seems …
17:17 Ticket #483 (please help me) closed by balaji
worksforme: Resolving until we hear further.
16:57 Ticket #331 (corrupted block allocated in segment.c[222]) closed by balaji
worksforme: Resolving until we hear further.
16:55 Ticket #321 (f77/buildiface bugs around HAVE_MULTIPLE_PRAGMA_WEAK) closed by balaji
fixed
16:55 Ticket #315 (Minor wart during MPE install) closed by balaji
fixed
16:52 Ticket #299 (ERROR MESSAGE: ../../../include/mpitypedefs.h:17:25: sys/bitypes.h: No ...) closed by balaji
fixed: Resolving till we hear back from the user.
15:22 Ticket #501 (Multiple process manager default) created by balaji
Is there a reason why we don't build multiple process managers by default …
13:50 Changeset [4188] by goodell
Fix non-profiling ranlib printing in the VERBOSE=0 case. Reviewed by …
11:30 Ticket #493 (About the Jumpshot) closed by chan
wontfix: resolved this unless user replies back.
11:14 Ticket #500 (Fwd: Problem with MPE with IBM MPI) created by Anthony Chan <chan@…>
[…]
10:56 Ticket #492 ([patch] Random data are not programs, don't strip them) closed by balaji
fixed: Fixed in r4186. Resolving (again).
10:52 Changeset [4187] by balaji
Removed faq text (used to generate the FAQ pages on the older website). …
10:43 Changeset [4186] by balaji
Use "/usr/bin/env perl" instead of "/usr/bin/perl" for portability. …

03/24/09:

17:50 Changeset [4184] by goodell
Minor fixes for updated OSU tests. No reviewer.
16:44 Ticket #492 ([patch] Random data are not programs, don't strip them) reopened by balaji
Can you try changing it top-line in the script to "#! /usr/bin/env perl …
16:22 Ticket #499 (Add test for ANY_SOURCE + ANY_TAG) created by William Gropp <wgropp@…>
[…]
16:15 Changeset [4183] by buntinas
Added SIGTERM signal in addition to SIGKILL in hydra proxy
16:05 Ticket #77 (configure support for memory barriers) closed by buntinas
fixed: fixed in r4178.
16:04 Ticket #486 (Missing service not discovered until runtime) closed by buntinas
fixed: Fixed in r4178 and r4180.
16:03 Changeset [4180] by buntinas
fixed sun atomic primitives
14:43 Changeset [4179] by goodell
Update the OMB benchmarks to version 3.1.1. No reviewer.
13:12 Changeset [4178] by buntinas
moved memory barriers implementations into per-architecture aromic …
08:18 Ticket #492 ([patch] Random data are not programs, don't strip them) closed by balaji
fixed: Committed the patch to trunk in r4177. Resolving.
08:17 Changeset [4177] by balaji
Added an install rule for DATADIR which makes sure simplemake does not …
07:08 Ticket #498 (rlog tracing granularity) created by balaji
Currently we only have two levels of rlog tracing -- summarized and …
07:06 Ticket #497 (irlog2rlog) created by balaji
irlog2rlog is not being placed in the $prefix/bin directory when …
06:59 Ticket #496 (Inlining support for MPID functions) created by balaji
Currently the MPID functions are not inlined. Systems with low-frequency …
04:32 Changeset [4176] by balaji
For files that are used by the updatefiles script to auto-generate other …
04:31 Changeset [4175] by balaji
The default environment setting had incorrectly gotten changed to NONE in …
04:30 Changeset [4174] by balaji
Use realid instead of recalculating, so later changes to how realid is …

03/23/09:

18:54 Changeset [4173] by balaji
Allow some more features to be controlled through the environment.
10:52 Ticket #494 (Re: About the Jumpshot) closed by chan
duplicate: resolved as duplicate of req#493.
08:49 Ticket #495 (Communication between two groups of spawned processes) closed by goodell
invalid: Hi Edwin, It looks like you are calling MPI_Intercomm_create incorrectly …
06:22 Ticket #495 (Communication between two groups of spawned processes) created by Edwin Vollebregt <edwin.vollebregt@…>
[…]
04:29 Ticket #457 (Hydra: Process-core mapping ability) closed by balaji
fixed: All process-core mapping options planned so far have been added into the …
04:29 Changeset [4172] by balaji
Added some README notes on the different mapping schemes.
04:20 Changeset [4171] by balaji
This should allow the user to specify process-core mappings within the …
01:38 Ticket #494 (Re: About the Jumpshot) created by 林庚佑 <shocklink99@…>
[…]
01:36 Changeset [4170] by balaji
Added a first cut of user-defined process-core mappings. This is still …
01:32 Ticket #493 (About the Jumpshot) created by 林庚佑 <shocklink99@…>
[…]

03/22/09:

23:53 Changeset [4169] by balaji
Dealing with C's weird problems with reading text files with and without a …
22:57 Ticket #485 (broken type tests) closed by goodell
fixed: I believe that the last of these problems is resolved in r4168.
22:57 Changeset [4168] by goodell
Fix for ticket #485, hopefully the last one. MPIR_CXX_*COMPLEX_VALUE …
22:50 Changeset [4167] by balaji
Bug fix for the case where the number of proxies to be launched is lesser …
22:01 Changeset [4166] by balaji
Updated the confdb version number for romio and hydra to propagate the …
21:31 Changeset [4165] by balaji
svn:ignore fixes.
19:30 Changeset [4164] by balaji
Added an option to specify no binding at runtime and made that the default …
19:23 Changeset [4163] by balaji
Modified the process-core binding model to get the topology information at …
18:32 Ticket #492 ([patch] Random data are not programs, don't strip them) created by Aleksej Saushev <asau@…>
[…]
18:09 Changeset [4162] by balaji
Fix for a major problem with the PMI_ID calculation that creeped in r4156
15:25 Changeset [4161] by balaji
Some more fixes to the environment passing. This time for environment …
13:09 Changeset [4160] by balaji
Bug fix when a host file is passed.
04:37 Changeset [4159] by balaji
Using PLPA_NAME instead of direct function names as recommended.
04:31 Changeset [4158] by balaji
Added an initial version of process-core mapping and some documentation …
01:18 Changeset [4157] by balaji
Fixed a bug in the passing of local environment variables + general …
00:19 Ticket #464 (Hydra: Multi-executable launches on the same node) closed by balaji
fixed: Fixed in r4156. Resolving.
00:18 Changeset [4156] by balaji
Added support for allowing each proxy to handle non-contiguous PMI_IDs. …

03/21/09:

22:20 Ticket #491 (Device-level connection state machine) created by balaji
We should consider moving the connection state machine to the device …
21:45 Ticket #490 (PMI Abort Communicator) created by balaji
Currently the communicator for which an MPI_Abort is called is completely …
21:39 Ticket #489 (Fallback Singleton Init Warning) created by balaji
This is an old bug that was reported by Darius (I recently ran into this …
21:27 Ticket #488 (Environment variable documentation) created by balaji
The environment variables available in MPICH2 need to be documented …
18:57 Ticket #468 (Intel test suite with Hydra and Intel compiler) closed by balaji
fixed: None of the tests last night seem to have any processes hanging. Resolving …

03/20/09:

23:27 Ticket #487 (PMI-1 Initialization sequence) created by balaji
The PMI-1 initialization sequence is broken. The client sends the server …
23:05 Changeset [4155] by balaji
Suck in all the PMI related functionality into the process manager …
18:39 Changeset [4154] by balaji
Process manager code refactoring. This is not complete yet, but should …
17:54 Ticket #480 (1.1b1 on icrunch) closed by balaji
duplicate: This is a duplicate of many tickets, one of which is #77.
17:24 Changeset [4153] by goodell
Add LOGICAL_EXTRA groups to l{and,or,xor} reduction operations. This …
17:17 Changeset [4152] by goodell
Fix broken MPI_DOUBLE_COMPLEX type in predefined ops (ticket #485). No …
17:17 Changeset [4151] by goodell
Fix broken MPI_Type_get_name from ticket #485. This is a correction to …
15:36 Ticket #300 (cross-compiling is broken in MPICH2-1.1a2) closed by gropp
fixed: Fixed. I had a more detailed description, but trac destroyed it in yet …
15:30 Ticket #325 (maint/updatefiles fails to create configure on niagara) closed by gropp
fixed: Fixed. updatefiles now detects broken versions of test that abort the …
15:23 Changeset [4150] by gropp
Added a test for whether test accepts -nt or whether using -nt will cause …
15:20 Ticket #486 (Missing service not discovered until runtime) created by gropp
I built MPICH2 on Niagara and everything went fine until I tried to run …
14:56 Ticket #485 (broken type tests) created by goodell
[…]
14:35 Changeset [4149] by balaji
Refactored the launcher code to reduce duplication. Also moved out all …
13:14 Ticket #484 (Scaling problem with MPI_Comm_create()) created by Brian Smith <smithbr@…>
[…]
13:09 Changeset [4148] by balaji
Update the confdb link to use the r4146 changes.
12:50 Changeset [4147] by balaji
Get rid of the ROMIO confdb and link it from the mpich2 trunk's version to …
12:15 Ticket #483 (please help me) created by aspho nour <aspho_nour@…>
[…]
12:00 Changeset [4146] by gropp
Updated the README for confdb with a history and file manifest; removed …
11:31 Ticket #482 (hydra confdb problem) closed by balaji
fixed: Fixed in r4145. Resolving.
11:29 Ticket #482 (hydra confdb problem) created by gropp
After my last svn update ; maint/updatefiles, I now have the directory …
11:22 Changeset [4145] by balaji
Use svn:externals in a way that's more portable across svn versions.
09:53 Changeset [4144] by buntinas
fixed typo in ia64 atomics
09:51 Changeset [4143] by buntinas
fixed typo in ia64 atomics
09:43 Ticket #481 (mpich2-1.0.8 make : compiling error on SGI altix 350) closed by buntinas
fixed: Yves, It looks like I have a fix for this. Download the patch from here: …
07:36 Ticket #481 (mpich2-1.0.8 make : compiling error on SGI altix 350) created by yves.champagnac@…
[…]
01:01 Changeset [4141] by balaji
We no longer have the dbg directory.
01:00 Changeset [4140] by balaji
Several cleanups to the overall Hydra code. We now use a consistent naming …

03/19/09:

23:21 Ticket #480 (1.1b1 on icrunch) created by Anthony Chan <chan@…>
[…]
21:53 Changeset [4139] by balaji
Update the version of confdb that we want in Hydra.
21:47 Changeset [4138] by goodell
Refactor the MPI_Type_get_name implementation slightly. This should …
21:47 Changeset [4137] by goodell
Refactor predefined reduction operations. This version should greatly …
21:46 Changeset [4136] by balaji
Removing some unnecessary files and updating some more to autoconf-2.62.
19:26 Changeset [4134] by balaji
Added an svn:externals link to get the mpich2 confdb to the local …
18:38 Ticket #479 (Hydra warnings with Intel compiler) closed by balaji
fixed: Fixed in r4133. Resolving.
18:37 Changeset [4133] by balaji
The compiler doesn't seem to be very happy about comparing a constant …
15:56 Changeset [4132] by goodell
add CPPFLAGS to start of make output No reviewer.
15:52 Ticket #478 (error compiling mpich2-1.1b1 with ch3:nemesis:gm device) closed by goodell
fixed: Fixed in r4131. Thanks again for the bug report. -Dave
15:50 Changeset [4131] by goodell
Fixed gm compile error from ticket #478. See also r3982 and ticket #429. …
10:43 Ticket #479 (Hydra warnings with Intel compiler) created by "Rajeev Thakur" <thakur@…>
[…]
10:37 Ticket #478 (error compiling mpich2-1.1b1 with ch3:nemesis:gm device) created by "Audet, Martin" <Martin.Audet@…>
[…]
10:03 Ticket #468 (Intel test suite with Hydra and Intel compiler) reopened by thakur
Last night, a "node" process was left behind on smash that prevented the …
10:01 Ticket #463 (Re: MPICH2 installation problem) closed by thakur
wontfix: Resolving until we hear further.
10:00 Ticket #471 (FW: Cron <thakur@octagon> /home/MPI/nightly/basictests -cc=gcc -cxx=g++ ...) closed by thakur
wontfix
01:34 Changeset [4130] by balaji
Throw some warnings when binding fails.
01:29 Changeset [4129] by balaji
Used some more detailed features of PLPA to check how many processing …
00:26 Changeset [4128] by balaji
More configure bugs.

03/18/09:

23:03 Changeset [4127] by balaji
Fixed a small error in the configuration that was causing the process …
21:19 Changeset [4126] by balaji
Set svn:ignore properties to the faults directory.
19:05 Changeset [4125] by balaji
Changed --enable-g=dbg,log,mem,meminit to --enable-g=all.
18:47 Changeset [4124] by balaji
First draft of the process-binding code (ticket #457).
16:15 Changeset [4123] by gropp
Added an optional test for handling of faults in MPI programs - MPICH2 …
16:09 Milestone mpich2-1.1b1 completed
15:15 Changeset [4122] by robl
found something else interesting to visualize with mpe
14:46 Changeset [4121] by balaji
Initial support for slurm (ticket #443).
14:14 Ticket #442 (Hydra stdin support) closed by balaji
fixed: Fixed in r4120. Resolving.
14:14 Changeset [4120] by balaji
Minor bug in the stdin code, that was causing stdin tests with more than …
14:13 Changeset [4119] by balaji
Move the sys/socket.h include to the appropriate header file, so other …
12:25 Changeset [4113] by balaji
Tagging trunk for the 1.1b1 release.
12:24 Changeset [4112] by balaji
This is a branch directory. We need to tag it. Duh!
11:53 Ticket #476 (Get nightlies working on windows) created by jayesh
We need to get the nigthly tests working once again on windows (The …
11:41 Ticket #475 (Get rid of all warnings with Visual Studio compiler) created by jayesh
This has been postponed for long. This ticket is to remind me that we need …
11:20 Changeset [4111] by balaji
Tagging trunk for the 1.1b1 release.
10:48 Changeset [4110] by balaji
Minor fix for fork. This is too orthogonal for something that's passing …
10:45 Changeset [4109] by buntinas
removed c++ style comments
10:20 Changeset [4108] by jayesh
Adding notes about SMPD deficiencies in 1.1b1 release
10:08 Ticket #474 (spawnmanyargv fails on smpd) created by jayesh
spawnmanyargv fails with "too many arguments" for smpd -jayesh
09:21 Changeset [4107]
Removed Asserts that caused one-sided ops to crash

03/17/09:

18:16 Changeset [4106] by balaji
Added a direct link to the Hydra documentation.
18:08 Changeset [4105] by goodell
Update the CHANGES file in preparation for the 1.1b1 release. Reviewed by …
17:26 Changeset [4104] by jayesh
Explicitly specifying path to executables in the fortran project template …
17:01 Changeset [4103] by goodell
Fix for an incorrect valgrind warning in the memory allocation code. No …
16:37 Ticket #473 (Building MPICH2 under AIX) created by William Gropp <wgropp@…>
[…]
14:28 Changeset [4102] by jayesh
Updating the product code & ver numbers in windows installers for the …
13:46 Ticket #468 (Intel test suite with Hydra and Intel compiler) closed by balaji
fixed: Fixed in r4101. Resolving.
13:45 Changeset [4101] by balaji
The size of the buffer passed to read was more than what the buffer could …
12:16 Changeset [4100] by jayesh
Handle unhandled defns - don't use _ALL_SOURCE for struct ifconf, no …
11:57 Ticket #428 (MPI_Win_fence memory consumption) closed by jayesh
fixed
11:55 Ticket #462 (test1_dt failed in nightlies) closed by jayesh
fixed: Fixed as of r4099. The original code ignore iov offsets when loading recv …
11:47 Changeset [4099] by jayesh
Unpack to the *right* iov buffers. The original code ignored iov offsets. …
11:42 Changeset [4098] by jayesh
Prevent sending the same data twice - Since we are sending 2 MPI_INTS …
09:44 Ticket #472 (document how to support networks that support matching) created by buntinas
This is a placeholder to remind me to write wiki documentation on how to …
08:43 Ticket #471 (FW: Cron <thakur@octagon> /home/MPI/nightly/basictests -cc=gcc -cxx=g++ ...) created by "Rajeev Thakur" <thakur@…>
[…]

03/16/09:

22:06 Ticket #466 (meet error, when run mpdboot) closed by thakur
fixed: Ralph helped find the problem. Resolving.
18:30 Changeset [4097] by gropp
Added socket.h, required for socket on AIX
18:29 Changeset [4096] by gropp
Added _ALL_SOURCE as option for struct ifconf; needed for AIX
16:38 Ticket #470 (add OpenPA to trunk, replace MPIDU_Atomic_* with OPA_*) created by goodell
This is a feature reminder for us to convert from the MPIDU_Atomic code to …
16:35 Ticket #469 (add knem to trunk) created by goodell
This is a feature ticket to remind us to integrate knem into the MPICH2 …
16:01 Ticket #468 (Intel test suite with Hydra and Intel compiler) created by "Rajeev Thakur" <thakur@…>
[…]
15:39 Changeset [4093] by balaji
b1 was hardcoded in one of the places in the spec file.
15:22 Changeset [4092] by balaji
Adding the first draft of the Fedora spec file contributed by Deji …
14:13 Changeset [4091] by balaji
Use strerror where available for better error information.
13:26 Ticket #467 (MPICH2 does not work with Boost.MPI) created by Dries Kimpe <dkimpe@…>
Dries forwarded this from the boost MPI list as a thread. I've copied the …
13:20 Changeset [4090] by balaji
Give out a better error message when the file is not found.
13:15 Changeset [4089] by balaji
Warning stomp.
12:11 Changeset [4088] by balaji
Don't enclose the environment in double quotes, the local shell will try …
10:02 Changeset [4087] by balaji
Converted the error checks to the MPICH2 style check-and-pop macros.
02:06 Ticket #466 (meet error, when run mpdboot) created by 王强 <xiaoqiang714@…>
[…]
00:31 Changeset [4086] by balaji
Some updates to the CHANGES file in preparation for the 1.1b1 release.

03/15/09:

23:49 Changeset [4085] by balaji
Allow mpiexec to take options of the form --bootstrap=ssh.
22:35 Ticket #456 (Hydra: Runtime selectable boot-strap servers) closed by balaji
fixed
22:32 Changeset [4084] by balaji
This commit will allow us to pick which bootstrap servers we want to …
22:19 Ticket #465 (PSM netmod for Nemesis) created by balaji
This ticket is a reminder for us to cleanup the PSM netmod in nemesis …
19:36 Ticket #464 (Hydra: Multi-executable launches on the same node) created by balaji
Hydra uses a separate proxy whenever there is a separate executable name. …
19:32 Changeset [4083] by balaji
Relax constraints on when the global parameters must occur.
19:14 Changeset [4082] by balaji
Allow the bootstrap server to be specified using an environment variable.
17:54 Changeset [4081] by goodell
Fix for the keyval_double_free test failures on Solaris. We weren't …
15:45 Ticket #463 (Re: MPICH2 installation problem) created by William Gropp <gropp@…>
[…]
11:26 Ticket #378 (ROMIO cross-compile bug on BGP) closed by balaji
duplicate: The remaining cross-compile issue in this ticket is a duplicate of ticket …
11:24 Ticket #461 (mpich2 error in build) closed by balaji
invalid: Resolving until we hear back from the user.
10:47 Ticket #462 (test1_dt failed in nightlies) created by jayesh
test1_dt (rma) failed in the nightlies. We should consider this separately …
02:23 Changeset [4079] by balaji
Cleanup and documentation.

03/14/09:

23:31 Changeset [4078] by balaji
Get rid of some unnecessary header files. Eventually, we want to get …
23:24 Changeset [4077] by balaji
Warning stomp.
23:02 Changeset [4076] by balaji
Fixed one more bug in the environment passing code. Some bootstrap servers …
22:35 Changeset [4075] by balaji
First cut of the runtime selection of bootstrap server (ticket #456).
22:17 Ticket #461 (mpich2 error in build) created by Jack Bryan <dtustudy68@…>
[…]
19:53 Changeset [4074] by balaji
Got rid of irrelevant continue statement.
19:52 Changeset [4073] by balaji
Added support for mpiexec --version.
19:04 Ticket #460 (last night's failures) closed by goodell
fixed: These were all fixed by r4066. I just ran them on atlantis with svn …
18:24 Changeset [4072] by balaji
Allocate/free resources in a more organized manner, so we don't leak …
17:26 Changeset [4071] by balaji
Changed the proxy to be called hydproxy to make it less generic.
16:50 Changeset [4070] by gropp
Fixed bug in simplemake when name of executable program does not match any …
16:03 Changeset [4069] by balaji
Additional help when bad arguments are passed to mpiexec.
16:01 Changeset [4068] by balaji
Change -v to --verbose to avoid confusion with version information.
16:00 Changeset [4067] by balaji
We were missing quote marks to contain the environment setting before …
12:00 Changeset [4066] by goodell
Followup to r4060. Missed one case and a warning. No reviewer.
11:11 Ticket #460 (last night's failures) created by rajeev_s_thakur@…
[…]
07:22 Changeset [4065] by balaji
Fixes to the stdout management code. When there was a lot of stdout data, …

03/13/09:

20:08 Ticket #459 (MPICH2 on Vista & windows 7) created by balaji
I was going over the older tickets in mpich2tkreq and moving relevant and …
20:00 Ticket #458 (Flow control in MPICH2) created by balaji
This is a reminder that we need to add communication flow-control in …
19:43 Ticket #457 (Hydra: Process-core mapping ability) created by balaji
We need to be able to allow processes to be bound to a processor or core …
19:36 Ticket #456 (Hydra: Runtime selectable boot-strap servers) created by balaji
We need to add runtime selection ability for Hydra's boot-strap servers. …
18:50 Changeset [4064] by balaji
install-alt should build the proxy too, together with mpiexec.
18:26 Changeset [4063] by balaji
Warning stomp.
18:08 Changeset [4062] by jayesh
1) Modifying the datatype code to find the number of contig blocks in an …
18:00 Ticket #356 (Hydra Intel test suite hang) closed by balaji
fixed: This one seems to have been fixed with the recent process cleanup commits. …
17:59 Changeset [4061] by jayesh
1) Adding mpir-mpioinit.c to windows makefile (2) Updating simple makefile …
17:50 Ticket #455 (Intel test c/grp_ctxt_comm/functional/MPI_Keyval3 failed last night) closed by goodell
fixed: Fixed in r4060. Resolving.
17:49 Changeset [4060] by goodell
Attribute bugfix (fixes ticket #455). We were not zeroing the attribute …
13:01 Ticket #454 (IOR segfault in MPI_Init) closed by robl
invalid: sigh. how embarrassing. a 'make clean' did the trick here. as I told …
13:00 Changeset [4059] by robl
reword
12:55 Changeset [4058] by robl
note about lustre
10:33 Changeset [4057] by gropp
Added support for environment variable RANLIBNEEDED which, if set to NO, …
10:32 Changeset [4056] by gropp
Enhanced the interface builders to handle multiple comments in a function …
10:30 Changeset [4055] by robl
The improved Lustre driver, contributed by Sun with help from Weikuan Yu.
09:44 Ticket #455 (Intel test c/grp_ctxt_comm/functional/MPI_Keyval3 failed last night) created by "Rajeev Thakur" <thakur@…>
[…]
06:28 Changeset [4054] by balaji
Don't completely ignore the free; free it after the process has been …
06:19 Ticket #453 (Warnings with hydra on octagon with thread-multiple) closed by balaji
fixed
06:17 Changeset [4053] by balaji
Missed a subtle statement on the putenv man page that states that the …
03:53 Changeset [4052] by balaji
We were not handling the case where mpiexec is called from a relative …
03:36 Changeset [4051] by balaji
Missed a change in the simplemake naming conventions. Argh.
00:17 Ticket #425 (bugs in test/mpi/configure.in) closed by chan
fixed: Fixed by applying the patch provided by David Gingold so that native …
00:15 Changeset [4050] by chan
Fixed f77 and f90 build when the base f77/f90 compilers are naive …

03/12/09:

22:24 Ticket #454 (IOR segfault in MPI_Init) created by Robert Latham <robl@…>
[…]
20:52 Changeset [4049] by balaji
This was a big mistake in the proxy implementation that was completely …
18:19 Ticket #399 (Subcommunicators, init, and brief-global) closed by goodell
fixed: These context IDs are now calculated from the parent context ID, so there …
18:04 Ticket #449 (ROMIO: free keyvals inside cleanup routines registered to run at ...) closed by robl
fixed: fixed in r4048
18:02 Ticket #222 (Initialization code in MPI-IO) closed by robl
fixed: OK, this and all related bugs seem to be fixed by a flurry of commits …
18:01 Changeset [4048] by robl
yay. the last of the unallocated handles. *and* it passes the romio tests. …
17:34 Ticket #282 (should --enable-fast=all disable error messages?) closed by goodell
fixed: Fixed as of r4047. I ran a the netpipe in the mpich2 test suite on …
17:33 Changeset [4047] by goodell
Fix for ticket #282, --enable-fast no longer disables error msgs. …
17:32 Changeset [4046] by goodell
Fix for a bad part of r3981. No reviewer.
17:20 Changeset [4045] by robl
found a situation where we would not build these tests
17:19 Changeset [4044] by robl
clean up a lot of old, unused code that got copied in from ROMIO, but only …
17:13 Changeset [4043] by balaji
Some of the messages that were showing up as errors were really warnings. …
16:58 Changeset [4042] by jayesh
Adding mpir-mpioinit.c added in r4010 to windows proj files
16:54 Changeset [4041] by balaji
Added a note on enable-strict=posix
16:48 Ticket #451 (Hydra failures on octagon) closed by balaji
invalid: This seems to be a problem with --enable-strict that Dave had pointed out …
16:38 Changeset [4040] by jayesh
Handling unhandled defns in windows configure script. 1) No need for …
16:05 Ticket #452 (Help with mpich2) closed by thakur
worksforme
15:50 Ticket #279 (Re: MPICH2-1.0.8 on windows with Compaq f90) reopened by jayesh
Satish reported that the syntax issue still exists for Compaq F90. …
15:44 Changeset [4039] by jayesh
1) Moving wintcp_module to wintcp - to follow naming convention of nemesis …
15:34 Ticket #453 (Warnings with hydra on octagon with thread-multiple) created by "Rajeev Thakur" <thakur@…>
[…]
15:34 Changeset [4038] by balaji
Committing some configuration setup changes to support runtime selection …
15:33 Changeset [4037] by jayesh
This change breaks MPICH2 windows compiln - renaming files to follow …
15:10 Ticket #279 (Re: MPICH2-1.0.8 on windows with Compaq f90) closed by jayesh
fixed: The change for ticket #276 should fix this too (hopefully). Satish, Can …
15:08 Ticket #276 (MPICH2-1.0.8 on windows with g77) closed by jayesh
fixed: Using "INTEGER *8" instead of "INTEGER (KIND=8)" for offset kinds in …
15:06 Changeset [4036] by jayesh
Using (INTEGER *8) fortran type for offsets instead of (INTEGER (KIND=8)) …
14:43 Ticket #402 (NPB IS hangs) closed by buntinas
fixed: Fixed in r4031.
14:28 Changeset [4035] by balaji
Missed merging the 4031 changeset. This should be the last in this round …
14:26 Changeset [4034] by balaji
Merging 4031:4033 from trunk (people are committing stuff faster than I …
14:22 Changeset [4033] by balaji
Merging 3965:4030 from trunk.
14:18 Changeset [4032] by goodell
Minor warnings stomp. No reviewer.
14:17 Changeset [4031] by buntinas
fixes #402 (NPB IS hangs) and other allgatherv hangs. Also narrowed down …
13:44 Ticket #452 (Help with mpich2) created by Ignacio Rodríguez <ignaciorj@…>
[…]
13:12 Ticket #434 (Intel fortran test MPI_Send_init_overtake fails only with Intel compiler ...) closed by goodell
wontfix: I can reproduce this, but I can't figure out what's going on. gdb doesn't …
12:58 Changeset [4030] by robl
function prototype for keyval copy and delete callbacks takes an 'int' …
12:43 Changeset [4029] by goodell
Fix a double-free bug with attribute keyvals. This is partly a fix for …
12:42 Changeset [4028] by goodell
Add a new test for double-freeing attriute keyvals. No reviewer.
12:42 Changeset [4027] by goodell
Add assertions to keyval reference counting macros. These make testing …
11:39 Ticket #451 (Hydra failures on octagon) created by Pavan Balaji <balaji@…>
[…]
10:50 Changeset [4026] by robl
free another keyval before exiting
09:49 Ticket #356 (Hydra Intel test suite hang) reopened by thakur
Not sure if this has been fixed. I just killed a "node" process left over …
08:07 Changeset [4025] by robl
explicitly free keyvals so handle-leak checks won't carp
04:03 Changeset [4024] by balaji
Committing in the initial version of the fork bootstrap server.
03:50 Changeset [4023] by balaji
Added some debug information. A lot more is needed.
02:38 Changeset [4022] by balaji
Pass the working directory to the proxy instead of explicitly cd'ing to …
02:00 Changeset [4021] by balaji
Cleaned up the code to find the location of the proxy.
01:53 Changeset [4020] by balaji
Several fixes to the environment passing routines.
00:29 Ticket #447 (Hydra process cleanup) closed by balaji
fixed: Committed the fix in r4019. Resolving.
00:28 Changeset [4019] by balaji
Added a connection setup between the job launcher and the proxies in case …

03/11/09:

21:58 Ticket #450 (A problem , when installed mpich2) closed by balaji
worksforme: Looks like you don't have a C++ compiler (e.g., g++), a fortran 77 …
21:20 Ticket #450 (A problem , when installed mpich2) created by shenghui wei <wsh.grxx@…>
[…]
16:16 Ticket #449 (ROMIO: free keyvals inside cleanup routines registered to run at ...) reopened by goodell
Fair enough, my misunderstanding. Reopening.
16:06 Ticket #449 (ROMIO: free keyvals inside cleanup routines registered to run at ...) closed by goodell
duplicate: Hi Lisandro, Thanks for the bug report. We are aware of this particular …
15:51 Changeset [4018] by balaji
Missed a socket close call when there is no stdin data. This was causing …
15:42 Ticket #449 (ROMIO: free keyvals inside cleanup routines registered to run at ...) created by Lisandro Dalcin <dalcinl@…>
[…]
14:43 Ticket #448 (Fwd: MPIR_Grequest_waitall) created by koop@…
[…]
14:20 Ticket #447 (Hydra process cleanup) created by balaji
When an application needs to be cleaned up, Hydra currently does this …
13:55 Changeset [4017] by buntinas
fixed related to %d vs MPIDI_MSG_SZ_FMT in debug messages on 64-bit …
13:38 Changeset [4016] by balaji
We don't support the HYDRA_USE_LOCALHOST option anymore. If nothing is …
13:32 Changeset [4015] by balaji
If the executable path starts with a /, strtok misses it.
13:15 Changeset [4014] by robl
as pointed out in #222, we should place attributes on COMM_SELF, not …
12:51 Ticket #438 (Ranlib needed) closed by goodell
fixed: Reverted in r4013.
12:50 Changeset [4013] by goodell
Revert "Cut out unnecessary ranlib step from linking process to speed up …
12:45 Changeset [4012] by balaji
Added a few known major changes for the 1.1b1 release to the CHANGES file.
12:44 Changeset [4011] by balaji
Enable MX netmod in the release tarballs.
11:36 Ticket #438 (Ranlib needed) reopened by goodell
I still don't believe that this is necessary (I think you have some wonky …
10:54 Changeset [4010] by robl
make ADIO initialization code a common routine. Also, place ROMIO …
04:33 Changeset [4009] by balaji
Fix incorrect freeing of variables.
03:47 Changeset [4008] by balaji
Fixed incorrect error message.
03:46 Ticket #356 (Hydra Intel test suite hang) closed by balaji
fixed: Committed a fix for this in r4007. Resolving.
03:45 Changeset [4007] by balaji
Cleanup processes when the application aborts abruptly. This should fix …
03:06 Changeset [4006] by balaji
Move the sh shell invocation to the bootstrap server, since this is …
02:57 Changeset [4005] by balaji
Fixed a bad bug that occurred when a macro was using the same variable as …
01:36 Changeset [4004] by balaji
With this fix, the proxy does not have to be in the user's path.

03/10/09:

18:07 Ticket #438 (Ranlib needed) closed by goodell
worksforme: Resolving this until we hear back from Bill.
17:51 Changeset [4003] by balaji
Got rid of an invalid case that was missed in r4002.
17:48 Changeset [4002] by balaji
Added back the code to wrap around host files when the number of nodes is …
17:22 Ticket #147 (Hydra process manager) closed by balaji
fixed: A draft of the Hydra process manager has been committed in to trunk. There …
17:20 Changeset [4001] by goodell
Compute sub-communicator context IDs from the parent's context ID. This …
17:15 Ticket #446 (Windows support for Hydra) created by balaji
This is a place holder for windows support for Hydra.
17:14 Ticket #445 (Hydra proxy enhancements) created by balaji
The current proxy implementation in Hydra is fairly simple. This needs to …
17:09 Ticket #444 (Dynamic process support in Hydra) created by balaji
This is a place holder for dynamic process support in Hydra.
17:06 Ticket #443 (Hydra: Support for additional tools) created by balaji
The following tools are currently supported by Hydra. Bootstrap …
16:54 Changeset [4000] by balaji
Move -f to a global option that affects all local executables. The code …
16:49 Ticket #442 (Hydra stdin support) created by balaji
Hydra does not currently support stdin when more than one process is …
16:34 Changeset [3999] by balaji
Missed some changes in the function arguments.
16:18 Changeset [3998] by balaji
Squashing several compile warnings in Hydra. Also made the demux engine …
14:42 Ticket #285 (Add channel/netmod context field for MPID_Request objects) closed by balaji
fixed
14:40 Ticket #403 (mpiexec kills the remote login shell) closed by balaji
invalid: Resolving until we hear further.
14:23 Ticket #184 (emulated atomics don't work on ia64 with nemesis) closed by balaji
wontfix: This will be a part of the OPA. Resolving.
12:48 Ticket #441 (mpiu_shm_wrappers warnings) created by balaji
I get a lot of warning messages (mostly in mpiu_shm_wrappers) when …
03:50 Changeset [3997] by balaji
Move the base utilities and types into a hydra_base.h header and common …
03:24 Changeset [3996] by balaji
Several bug fixes for the stdin management code.
00:26 Changeset [3995] by balaji
Some cleanup to the environment handling scripts, but more things need to …
Note: See TracTimeline for information about the timeline view.