Context Navigation

13:47 Changeset [5273] by goodell: Internal checks for context ID leaks when handle debugging is enabled. …
13:47 Changeset [5272] by goodell: Fix a multi-threaded context ID leak. Under multi-threaded contention the …
09:17 Changeset [5271] by goodell: Add --with-pm=none as a synonym for --with-pm=no. Some of the wiki docs …
09:17 Changeset [5270] by goodell: Another fix for MTest comm creation broken by r5229. This is a similar …

08/26/09:

17:49 Changeset [5269] by balaji: Use the process-mapping PMI attribute in PMI-1 as well, like MPD does.
17:48 Changeset [5268] by balaji: Remove the explicit format for process mapping; this is unused, untested …
09:29 Ticket #835 (Hydra: Persistent proxy launch race) created by balaji: Here's an error reported by Scott Atchley from Myricom: […]

08/25/09:

20:50 Ticket #834 (Hydra: Persistent proxy aborts) created by balaji: With mpiexec is killed on an abnormal termination, the persistent proxies …
12:17 Ticket #833 (Installation under ubuntu 8.04) created by Leonardo Bargigli <leonardo.bargigli@…>: […]

08/24/09:

16:49 Changeset [5262] by goodell: Avoid integer overflow in allgather size calculations. Thanks to Kim …
14:16 Changeset [5261] by goodell: Clean up the probe-intercomm test. It made some potentially unsafe …
14:16 Changeset [5260] by goodell: Check for MPI_COMM_NULL when using MTestGetIntercomm in several tests. …
14:16 Changeset [5259] by goodell: Set the is_low_group field when creating an intercomm via MPI_Comm_create. …
14:16 Changeset [5258] by goodell: Fix hanging intercommunicator collective tests broken by r5229. The logic …

08/22/09:

11:09 Changeset [5256] by balaji: Use the -distrib option to maint/updatefiles when creating releases. …

08/20/09:

15:45 Changeset [5254] by goodell: Don't clobber files named FOO when FOO.vin isn't present. Running …
15:45 Changeset [5253] by goodell: WIP Implement MPI_Reduce_local and MPI_Op_commutative (MPI-2.2 ticket …
15:45 Changeset [5252] by goodell: Add MPI_IN_PLACE support to MPI_Exscan (MPI-2.2 ticket #94). Per MPI-2.2, …
15:45 Changeset [5251] by goodell: MPI_Exscan cleanup in prep for MPI Forum ticket #94 implementation. No …
15:45 Changeset [5250] by goodell: Attribute delete callbacks are now ordered (MPI-2.2 ticket #71). …
15:45 Changeset [5249] by goodell: Fix MPI_Comm_create for intercomms and add MPI-2.2 ticket #66 tests. This …
15:44 Changeset [5248] by goodell: Permit MPI_REQUEST_NULL in MPI_Request_get_status (MPI-2.2 ticket #137). …
15:19 Ticket #832 (mpich2 1.1.1p1 does not install on win2k systems) created by jayesh: The user was able to install 1.0.7 successfully on his win2k machine. …
09:16 Changeset [5247] by mercierg: Slight change in cancel code
02:47 Changeset [5246] by mercierg: typos
02:43 Changeset [5245] by mercierg: Adding some information about the current state of the MX Module
00:03 Ticket #831 (Fatal error in MPI_Wait) created by Preeti <preeti@…>: […] untitled-part.html

08/19/09:

21:37 Ticket #829 (subscribe mpich-discuss) closed by goodell: wontfix: We've had other reports of trouble on older ia64 machines (like ticket …
20:37 Ticket #829 (subscribe mpich-discuss) created by "163aaronxue" <163aaronxue@…>: […] untitled-part.html
18:05 Changeset [5244] by goodell: Add '*' to function ptr args to MPI::Grequest::Start (MPI-2.2 ticket #6). …
18:05 Changeset [5243] by goodell: Implement MPI-2.2 ticket #7, _errhandler_fn --> _errhandler_function. …
18:01 Changeset [5242] by goodell: Tweaks for maint/updatefiles. We now understand --do= as well as -do=. …
14:07 Ticket #828 (Provide 32-bit and 64-bit SDKs with windows installation) created by jayesh: This is a place holder to remind me to provide both 32-bit and 64-bit SDKs …
12:38 Ticket #797 (Urgent - MPICH2 installer broken) reopened by smithpd@…: […] untitled-part-2.html
12:33 Changeset [5241] by balaji: We were accidentally removing PMI2 in the releases.
11:55 Ticket #797 (Urgent - MPICH2 installer broken) closed by jayesh: fixed: Hi, The patch release, 1.1.1p1, is a complete release of MPICH2 (no …
11:52 Changeset [5240] by jayesh: Updating the GUIDs for windows installers for the next release in 1.1.1 …
10:29 Ticket #801 (Unable to use svn, even with new version) closed by goodell: worksforme: Not that I know of. It was probably some transient networking issue that …
10:03 Ticket #800 (Scancel test hangs when MPI_Issend not cancelled?) closed by gropp: invalid
09:38 Ticket #802 (MPD error) created by "163aaronxue" <163aaronxue@…>: […] untitled-part.html
09:38 Ticket #801 (Unable to use svn, even with new version) created by William Gropp <wgropp@…>: […]
05:07 Changeset [5237] by mercierg: Fixed the DirectSsend? routine that didn't set the request ptr
05:06 Changeset [5236] by mercierg: Added the same fix that Dave applied to mpid_isend for requests allocation
04:54 Ticket #800 (Scancel test hangs when MPI_Issend not cancelled?) created by mercierg: It seems to me that the scancel test in the test suite is not correct when …

08/18/09:

17:24 Changeset [5235] by balaji: Forgot to remove SIGTSTP from the list of handled signals.
17:09 Changeset [5234] by balaji: Use SIGUSR1 instead of SIGTSTP for checkpointing since a child process …
17:05 Ticket #798 (MPI_Reduce_local test in Fortran (for MPI-2.2 ticket #24)) created by goodell: Anthony has kindly volunteered to write Fortran versions of the C test …
14:33 Ticket #797 (Urgent - MPICH2 installer broken) created by Peter Smith <smithpd@…>: […] untitled-part.html [[Image(...)]]
11:02 Changeset [5233] by mercierg: Fix corrected
10:58 Changeset [5232] by mercierg: Fixed error reporting
09:25 Changeset [5231] by mercierg: Momentary fix for alltoallw1/tresized/tresized2 tests
05:59 Ticket #796 (Troubleshooting) created by "#LEONG HON WAI#" <LEON0101@…>: […] untitled-part.html c.txt

08/17/09:

16:06 Changeset [5230] by goodell: Make MPIR_GroupCheckVCRSubset a bit safer with some argument checks. No …
16:06 Changeset [5229] by goodell: Fix potential deadlock in mtest communicator creation. The 'done' …
16:06 Changeset [5228] by goodell: Some of the intercomm tests weren't checking for MPI_COMM_NULL. No …
16:06 Changeset [5227] by goodell: Fix intercomm localcomm context ID calculation. They were very broken …
16:06 Changeset [5226] by goodell: Fix inconsistent alloc/free of context IDs for intercomms. The various …
15:21 Ticket #616 (Migrate tests to new servers) closed by thakur: fixed
13:49 Changeset [5225] by buntinas: ignoring any leading spaces in initial breakpoint response. This is an …

08/14/09:

22:23 Ticket #783 (allred test fails) reopened by thakur: The make install step looks for .so files in the install directory and …
17:09 Ticket #783 (allred test fails) closed by gropp: worksforme: Will someone send the trac authors an introductory book on databases - …
16:53 Ticket #794 (allred test assume MPI 2.2) closed by goodell: fixed: Fixed by r5220 and r5221. Resolving.
16:52 Changeset [5221] by goodell: Make MPI-2.2 tests conditional on the MPI_(SUB)VERSION (fixes #794). The …
16:52 Changeset [5220] by goodell: Add a new MTEST_ macro to simplify MPI version number checks. No …
15:32 Ticket #794 (allred test assume MPI 2.2) created by gropp: The MPICH2 test suite is intended as a general MPI test suite. The IBM …
15:10 Changeset [5219] by balaji: Update confdb and mpl versions.
15:06 Changeset [5218] by balaji: Check if plpa and blcr work correctly before setting them as the default …
15:04 Ticket #793 (Fwd: [mpich2-dev] fandcattrf90 test on BGP) created by William Gropp <wgropp@…>: […] untitled-part.html
15:01 Changeset [5217] by balaji: Change the PAC_CHECK_HEADER_LIB to return an error instead of aborting. …
13:26 Changeset [5215] by balaji: Pass the PMI_PORT to the restart process so it can be reset. Also, allow …
13:07 Changeset [5214] by balaji: Handle the case where there is no binding or checkpoint library installed. …
04:44 Changeset [5212] by balaji: When a process is restarted from the checkpoint, the checkpointing system …
04:40 Changeset [5211] by balaji: Add a help message for the restart path.
04:38 Changeset [5210] by balaji: Initial draft of the checkpoint restart code.
03:44 Changeset [5209] by balaji: Bug fix for the persistent proxy case.
03:37 Changeset [5208] by balaji: Initial draft of the checkpoint initiation code; this is not fully tested …
02:55 Changeset [5207] by balaji: The binding library specified on the command-line should always have a …
02:41 Changeset [5206] by balaji: Don't expose the semantics of the binding and checkpointing tools to the …
01:59 Changeset [5205] by balaji: 1. Modify the UI to breakup the handling of different options into …

08/13/09:

13:11 Changeset [5204] by balaji: Added macros for pushing variables into a stack to help with nesting …
01:51 Changeset [5203] by balaji: When the bootstrap server creates a nodes list, make sure that we don't …

08/12/09:

16:40 Ticket #792 (A error about omp.h) closed by chan: invalid
15:51 Ticket #792 (A error about omp.h) reopened by yy2250@…: […]
14:54 Ticket #792 (A error about omp.h) closed by chan: invalid: The native compiler used by mpicxx must support OpenMP. If this is the …
14:13 Ticket #792 (A error about omp.h) created by yy2250@…: […]
12:52 Ticket #791 (Fortran 90 errors with autoconf 2.64) created by balaji: ./maint/updatefiles fails with autoconf 2.64. Here's the error: […]
11:12 Ticket #790 (Hydra pm don't print exit signals) closed by balaji: fixed: Thanks! This is fixed in r5202 (it also prints out the signal number now). …
11:11 Changeset [5202] by balaji: Use WTERMSIG to get the correct signal number when the process exits with …
10:47 Ticket #790 (Hydra pm don't print exit signals) created by https://www.google.com/accounts/o8/id?id=AItOawlmBmkqBufNwgbjwoMgyyV90a0c_jmqP2k: Package mpich2-1.1.1p1 src/pm/hydra/ui/mpiexec/mpiexec.c must be patched …
10:22 Changeset [5201] by buntinas: Stomped some strict alias warnings, and a couple others

08/11/09:

14:30 Changeset [5199] by buntinas: added blcr code to hydra
14:28 Changeset [5197] by buntinas: added quoting to macros
13:49 Changeset [5195] by goodell: Fix cpp undef lines in oputil.h that were added in r5142. Thanks to Pavan …
13:49 Changeset [5194] by goodell: Removing spurious newlines in the handle debug messages. No reviewer.
08:53 Changeset [5193] by gropp: Improve use of const in PM util routines
08:52 Changeset [5192] by gropp: Move Fortran size-specific types within the eif-endif for Fortran support
08:52 Changeset [5191] by gropp: Change test to output information on failure as per the style of the other …
08:50 Changeset [5190] by gropp: Prepare configure to support aint instead of void * for holding …

08/10/09:

18:28 Ticket #781 (Build of either "mpich2-1.1.1" or "mpich2-1.1.1p1" fails for "initcxx.cxx") reopened by ejotoo@…: […]
17:43 Ticket #781 (Build of either "mpich2-1.1.1" or "mpich2-1.1.1p1" fails for "initcxx.cxx") closed by chan: worksforme: […] I don't use bash often, so not sure if .bash_profile is the file …
15:26 Ticket #691 (parallel make support in MPE) closed by chan: fixed: fixed by patches r5077, r5082 and r5171.
14:50 Changeset [5186] by buntinas: update mpl
14:49 Changeset [5185] by buntinas: update confdb
14:49 Changeset [5184] by buntinas: update confdb
14:45 Changeset [5183] by buntinas: merged /mpich2/branches/dev/ckpt2:r5182 -- fixed up checking for blcr …
14:37 Changeset [5181] by buntinas: merged /mpich2/branches/dev/ckpt2/confdb:r5180 -- added convenience macros …
14:08 Changeset [5179] by gropp: The macro to test whether a compiler option could be used failed in some …
12:42 Ticket #788 (aclocal_mpi.m4 missing macros) closed by balaji: duplicate: I've restored these two macros in r5177. Since we are handling this issue …
12:40 Changeset [5177] by balaji: Revert PAC_PROG_F77_CMDARGS and PAC_PROG_F77_RUN_PROC_FROM_C that were …
10:33 Changeset [5175] by goodell: Fix bad condition stringification in MPIU_Assert_fmt_msg. No reviewer.
02:30 Ticket #789 (Problems building MPE2 with IBM MPI) closed by chan: fixed: resurrected the old file and tested on bp-login and typical linux box. …
02:14 Changeset [5174] by chan: r5067 incorrectly replaced MPE_Thread* by MPIU_Thread* that causes …

08/09/09:

19:23 Ticket #789 (Problems building MPE2 with IBM MPI) created by gropp: Thanks for catching the bug. It looks like r5067 incorrectly replaced …
13:14 Changeset [5173] by balaji: Restore aclocal_mpi.m4 since that seems to be used in other projects.
12:22 Ticket #788 (aclocal_mpi.m4 missing macros) created by gropp: The macros PAC_ARG_MPI_TYPES PAC_LIB_MPI are missing from aclocal_mpi.m4 …

08/08/09:

15:34 Ticket #123 (MPIDU_Yield and MPID_Thread_yield) reopened by balaji: Reverted r5166 in r5172. This was causing device-specific definitions of …
15:33 Changeset [5172] by balaji: Essentially revert r5166 as it was causing device-specific definitions of …
12:29 Ticket #165 (Config and binary file conflict) closed by balaji: wontfix: Based on some discussion within the group and binary package creators, …
12:25 Ticket #689 (Duplicate flags) closed by balaji: fixed: This was fixed by Dave for the 1.1.1p1 release. Resolving.
11:08 Ticket #783 (allred test fails) created by gropp: I now see this: […] The failing datatype (strangely, not identified …
01:15 Changeset [5171] by chan: added more parallel make support through the use of AC_PROG_CC_C_O.
00:54 Ticket #781 (Build of either "mpich2-1.1.1" or "mpich2-1.1.1p1" fails for "initcxx.cxx") created by EJO <ejotoo@…>: Hi, Both "mpich2-1.1.1" and "mpich2-1.1.1p1" fail in attempt to compile …
00:30 Changeset [5170] by balaji: Move the nightly tests back to trunk.
00:28 Changeset [5169] by balaji: Update confdb in romio and mpl in Hydra.
00:28 Changeset [5168] by balaji: Update confdb in MPL.
00:25 Ticket #395 (etags configure checks) closed by balaji: fixed: Fixed in r5167. Resolving.
00:23 Changeset [5167] by balaji: Remove tags checks. This is completely unused and broken. This also …
00:22 Ticket #123 (MPIDU_Yield and MPID_Thread_yield) closed by balaji: fixed: Fixed in r5166. Resolving.
00:20 Changeset [5166] by balaji: Move MPIDU_Yield and MPID_Thread_yield to MPIU_ functionality and unify …
00:15 Changeset [5165] by balaji: Cleanup macros in confdb that are unused or irrelevant at this time since …

08/07/09:

16:02 Ticket #779 (Lustre compilation error on FC7 Linux (but not FC4)) created by "Audet, Martin" <Martin.Audet@…>: […]
06:03 Changeset [5161] by mercierg: Typo
05:58 Changeset [5160] by mercierg: Complementary fix when not using a specific context for ADI msg
04:27 Changeset [5159] by mercierg: This fixes the known bug of the OSU MPI test suite. It passes all tests in …

Note: See TracTimeline for information about the timeline view.

Download in other formats:

RSS Feed