Timeline
11/07/09: Today
- 07:29 Changeset [5733] by
- Fix hanging autoconf display message.
- 07:29 Changeset [5732] by
- Fixed help message that states that enable-strict is only for gcc, which …
11/06/09: Yesterday
- 20:10 Ticket #928 (Changes to LIBS are lost between configure in ch3 and configure in nemesis) created by
- The value of LIBS at the beginning of ch3/configure is set by the top …
- 19:59 Changeset [5731] by
- added ftb to ch3 and nemesis
- 18:19 Ticket #926 (enable-fast) closed by
- wontfix: Probably not. Resolving it for now.
- 13:47 Ticket #927 (Spawn() fails on remote node with nemesis on windows) created by
- Actually, it does work locally but fails remotely, with channel nemesis. …
- 10:18 Changeset [5727] by
- Yet another workaround for bb problems -- wait for longer after allocating …
- 10:15 Ticket #926 (enable-fast) created by
- Bcast seems to be hanging with enable-fast in the new nightly tests: …
- 09:19 Changeset [5726] by
- Update version number in preparation for the 1.2.1rc1 release.
- 09:18 Changeset [5725] by
- Updating the guids of the installers for the upcoming rc release
- 09:10 Changeset [5724] by
- Since disabling thread package is broken, make the thread package …
- 09:07 Changeset [5723] by
- Adding async.c - added in r5694 - into VS projs
- 08:41 Changeset [5722] by
- Temporary fix. Setting this pointer seems to corrupt hwloc's topology …
- 08:09 Changeset [5721] by
- renamed MPI_COMBINER_ENUM to MPIR_Combiner_enum
11/05/09:
- 21:35 Ticket #902 (mpd PM post-execution error: close failed in file object destructor) closed by
- fixed: Fixed and/or worked-around in r5720. There might be something smarter to …
- 21:34 Changeset [5720] by
- Fix for mpd on FC12 (ticket #902). mpdman closes all non-standard file …
- 16:04 Ticket #925 (Update website) closed by
- fixed
- 15:58 Changeset [5717] by
- Missed state declarations in the async code causing rlog builds to fail.
- 15:48 Ticket #925 (Update website) created by
- This page needs to be updated. …
- 15:10 Ticket #924 (environment vars) closed by
- fixed: Fixed in r5714. Resolving.
- 15:05 Changeset [5714] by
- Add information about some of the environment variables in README.vin. …
- 14:26 Changeset [5713] by
- Memory allocation bug fix in the Hydra process binding code. Thanks to …
- 14:24 Changeset [5712] by
- White space cleanup.
- 14:21 Changeset [5711] by
- hwloc was assuming autoconf-2.64. Fix the usage to make it autoconf-2.63 …
- 14:13 Changeset [5710] by
- hwloc changed its autogen mechanism for the final release. It's better to …
- 13:36 Changeset [5709] by
- New binding file for hwloc: matches Hydra's new data structure
- 12:56 Ticket #910 (Hydra: Documentation for Internal Interfaces) closed by
- fixed: Fixed in r5708. Resolving.
- 12:55 Changeset [5708] by
- Doxygen powered documentation for Hydra's internal interfaces. Fixes …
- 11:58 Changeset [5707] by
- Updated hwloc to the 0.9.2 release.
- 11:56 Changeset [5706] by
- mpiu_thread.h was never explicitly included with the async code. We were …
- 09:10 Ticket #924 (environment vars) created by
- We need to document environment variables like MPICH_NO_LOCAL, …
- 00:26 Changeset [5701] by
- Added an autogen file to be used when building Hydra independently.
- 00:19 Changeset [5700] by
- Get rid of ltmain.sh. This was initially required when we relied on …
- 00:19 Changeset [5699] by
- Suppress function unused warning.
- 00:19 Changeset [5698] by
- Move the tools API prototypes to their appropriate header files (closer to …
- 00:19 Changeset [5697] by
- Add a FIXME to correct the critical section code as per Bill's …
11/04/09:
- 23:37 Ticket #923 (mscvcr80.dll was not found) created by
- I download and install your excellent program; but when I run "mpiexec" on …
- 17:51 Ticket #917 (Asynchronous Progress in MPICH2) closed by
- fixed: The above described implementation has been added into the MPICH2 trunk in …
- 17:45 Changeset [5695] by
- Configure protect the asynchronous progress stuff. This will be released …
- 17:45 Changeset [5694] by
- Simple version of the asynchronous progress code, where we have a thread …
- 15:23 Ticket #915 (SMPD does not handle env vars ending in backslashes correctly) closed by
- fixed: This is fixed in r5689 and should be available in the next release. …
- 15:22 Changeset [5689] by
- Making sure that we don't have the ESCAPE character at the end of a …
- 14:48 Changeset [5688] by
- added count=0 case for both C and forttran version of reduce_local test.
- 12:41 Changeset [5686] by
- Disabling suppression of pre-defined refcounts on windows
- 12:34 Changeset [5685] by
- Handling unhandled defns - added in r5666 - in windows configure script
- 12:20 Changeset [5684] by
- Changing *VERSION to MPICH2_*VERSION in windows configure - this change is …
- 11:08 Changeset [5683] by
- Disabling VS make logs for C test projects - Speeds up the build for test …
- 11:05 Ticket #922 (allctypesf.f and allctypesf90.f90 fail on windows) closed by
- fixed: Fixed in r5682. -Jayesh
- 11:05 Changeset [5682] by
- 1) Adding manually defined types and types defined from other types into …
- 11:01 Ticket #922 (allctypesf.f and allctypesf90.f90 fail on windows) created by
- These tests fail with error message that several MPI datatypes …
- 10:26 Ticket #920 (hydra's Makefile does not read VERBOSE) closed by
- fixed: Fixed in r5680. Now simplemake accepts both V=1 and VERBOSE=1 (if either …
- 10:24 Changeset [5681] by
- Export the MPICH2 version and other relevant information for Hydra to pick …
- 10:24 Changeset [5680] by
- Allow VERBOSE=1 or V=1 format for simplemake. Fixes ticket #920. Reviewed …
- 10:02 Ticket #618 (nesting and subarray) closed by
- fixed: OK, think this is fixed in r5679 -- moved nest-decr into fn_exit. It's a …
- 09:59 Changeset [5679] by
- make sure all code paths, even errors, decrement the nesting level
11/03/09:
- 22:16 Ticket #918 (Build warnings with --enable-strict=noopt) closed by
- fixed: Never mind that. This works correctly. My git setup seems to be giving …
- 20:31 Ticket #918 (Build warnings with --enable-strict=noopt) reopened by
- It doesn't look like this problem is fixed. I still get the same warnings: …
- 20:24 Ticket #921 (MPICH2 performance in long-term climate simulations) created by
- Hi, I conduct regional climate simulations using the Weather Research and …
- 17:13 Changeset [5678] by
- errcheck for null newtype
- 16:47 Changeset [5677] by
- removed unused variables.
- 15:12 Ticket #798 (MPI_Reduce_local test in Fortran (for MPI-2.2 ticket #24)) closed by
- fixed: resolving this by #5676.
- 15:11 Changeset [5676] by
- fixed reduce_local.c and added fortran version, reducelocalf.f, and added …
- 13:33 Ticket #920 (hydra's Makefile does not read VERBOSE) created by
- It appears that hydra's Makefile does not read VERBOSE=1
- 13:28 Ticket #919 (Allgatherv performance) created by
- One of the users reported a performance problem for Allgatherv for …
- 13:26 Changeset [5675] by
- Fixes for MPI_Reduce_local and the reduce_local test. Anthony noticed …
- 11:15 Changeset [5674] by
- update to ltmain.sh
- 11:13 Ticket #918 (Build warnings with --enable-strict=noopt) closed by
- fixed: Fixed in r5673. Returning structures is legal in ANSI C, so it appears …
- 11:10 Changeset [5673] by
- removed -Waggregate-return warning. Fixes #918.
- 10:46 Ticket #918 (Build warnings with --enable-strict=noopt) created by
- There are several warnings with compiled with --enable-strict=noopt with …
- 09:16 Ticket #917 (Asynchronous Progress in MPICH2) created by
- There have been a bunch of requests for asynchronous progress capability …
- 08:19 Changeset [5669] by
- Delete the older code for allgatherv (large messages) which has now been …
11/02/09:
- 23:21 Changeset [5668] by
- Update svn:ignore properties to include libtool generated macro files.
- 18:10 Changeset [5667] by
- Allow users to pick the PMI version they want based on environment …
- 17:31 Changeset [5666] by
- Work around Darwin strict behavior and sys/stat.h in ROMIO. When MPICH2 …
- 17:31 Changeset [5665] by
- Fix GREQ_CLASS leaks in the generalized request extensions. This commit …
- 17:25 Changeset [5664] by
- Name-space the PMI2 (and PMI-1.1) process-mapping key value. Reviewed by …
- 16:20 Ticket #916 (Disable error return tests in the Intel test suite) created by
- When MPICH2 is configured with --enable-fast or --disable-error-checking, …
- 16:10 Ticket #915 (SMPD does not handle env vars ending in backslashes correctly) created by
- set CNTRL_MPI_DIR=d:\mpi\ mpiexec -n 1 -genvlist CNTRL_MPI_DIR cmd /c set …
- 14:09 Changeset [5663] by
- The -tp=piv option is PGI specific and doesn't even seem to work on the …
- 08:52 Ticket #913 (Failures when weak symbols are disabled) closed by
- fixed: Fixed by r5659, resolving.
- 08:51 Changeset [5659] by
- Avoid duplicate symbols in comm_create if weak symbols aren't available. …
11/01/09:
- 16:18 Changeset [5656] by
- Added diablo-jdk in the list of paths to search for a Java installation; …
- 16:11 Changeset [5655] by
- Add a dependency on txt2man for the release tarball creation script (this …
- 15:58 Changeset [5654] by
- Fix the man page generation in mpd. Thanks to Lucas Nussbaum for reporting …
- 05:09 Ticket #911 (Hydra: Extend process binding glue layer to accomodate hwloc's features) closed by
- fixed: This should be fixed in r5653. Resolving.
- 05:08 Changeset [5653] by
- Update the HYDT_bind_info structure to make it more flexible and future …
- 04:21 Ticket #914 (Hydra: Process binding to multiple OS indices) created by
- We need to allow process binding to multiple OS indices in Hydra. This is …
10/31/09:
- 23:08 Ticket #905 (mpiexec -machinefile :broken) closed by
- fixed: There doesn't seem to be a cleaner way to do this because of how mpd is …
- 20:43 Changeset [5652] by
- Get rid of unused code.
- 16:07 Changeset [5651] by
- Work around another nasty bug on our breadboard imaging system. All …
- 15:52 Ticket #913 (Failures when weak symbols are disabled) created by
- The Intel test suite is failing with the --disable-weak-symbols option. …
- 15:44 Changeset [5650] by
- Bug fix for the failing timeout test in the nightly tests.
- 13:12 Changeset [5649] by
- Call the process cleanup directly instead of raising a signal and …
- 11:16 Changeset [5648] by
- Missed escaping some more quotes.
- 09:43 Changeset [5647] by
- added error macros for internal errors
10/30/09:
- 21:16 Changeset [5643] by
- Escape nested quotes.
- 21:11 Changeset [5642] by
- Disable OPA atomic primitives for suncc.
- 19:01 Ticket #903 (Hydra: MPICH_INTERFACE_HOSTNAME support) closed by
- fixed: Fixed in r5641. Resolving.
- 19:01 Changeset [5641] by
- Fix for the MPICH_INTERFACE_HOSTNAME issue (ticket #903). Now Hydra will …
- 17:56 Changeset [5639] by
- Fix mpd process_mapping in another case (ticket #905). This is yet …
- 16:05 Changeset [5637] by
- Disabling --enable-g for the SMPD nightly tests, as per Jayesh's request.
- 15:55 Ticket #912 (fortran test MPI_Errhandler_fatal fails with SMPD on the nightlies) created by
- […]
- 15:48 Changeset [5636] by
- added svn:ignores
- 14:56 Ticket #911 (Hydra: Extend process binding glue layer to accomodate hwloc's features) created by
- Hydra's process binding glue layer is not sophisticated enough to allow …
- 14:14 Ticket #776 (Problem installing MPICH2 on a Windows XP machine) closed by
- fixed: This should be fixed (no more XP SP3 pre-req) as of r5634. The patch will …
- 14:13 Ticket #832 (mpich2 1.1.1p1 does not install on win2k systems) closed by
- fixed: This should be fixed as of r5634. The patch will be available in the next …
- 13:22 Changeset [5635] by
- Reenable C++ tests for the Sun compiler. This got fixed earlier than …
- 12:40 Changeset [5634] by
- Handling the case where we don't have info about logical procs on windows …
- 11:24 Changeset [5633] by
- Define thread specific abstractions on windows as in unix. This change …
- 04:09 Changeset [5631] by
- Bind on the right core number.
10/29/09:
- 23:03 Changeset [5630] by
- Use consistent naming convention for the process-core binding parameters …
- 23:03 Changeset [5629] by
- Bug fix for the user-specified bindings code. Thanks to Guillaume to …
- 21:12 Ticket #910 (Hydra: Documentation for Internal Interfaces) created by
- Hydra doesn't have proper documentation for the internal interfaces. The …
- 20:55 Changeset [5628] by
- Fix the unhandled opa defn - sizeof opa ptr - on windows correctly. The …
- 16:57 Changeset [5627] by
- ROMIO wibbles. Reviewed by robl@.
- 16:57 Changeset [5626] by
- Fix errhandler leak in C++ introduced in r850. The fix so that …
- 16:57 Changeset [5625] by
- Reset the errhandler in MPI_File_close and ADIO_End. Prior to this commit …
- 16:56 Changeset [5624] by
- Fix MPI_File errhandler leak. Prior to this commit we only released …
- 16:56 Changeset [5623] by
- Never refcount builtin errhandlers. Reviewed by robl@.
- 16:56 Changeset [5622] by
- Fix errhandler leaks in fileerrret, fileerrretx, and userioerr tests. …
- 16:48 Changeset [5621] by
- Handling undefined opa defns - sizeof opa ptr - on windows
- 16:38 Changeset [5620] by
- Adding C versions of MPI datatypes in fortran for windows - added in unix …
- 15:00 Changeset [5618] by
- bug fixed by Pavan directly in my directory
- 14:40 Changeset [5617] by
- Allow --with-libtool option and MPICH2_LIBTOOL_DIR environment for …
- 10:48 Changeset [5612] by
- You can't bind when the topology is not initialized!
- 07:30 Changeset [5611] by
- Make sure that we run the selected autoreconf, not whatever is first in …
- 05:03 Changeset [5610] by
- Fix for issues when No Nodes/Socket level is present
- 04:04 Changeset [5609] by
- Fix for the case where no NUMA node is present in the hierarchy
10/28/09:
- 21:17 Changeset [5608] by
- Incoming data was being overwritten in this test, which is not MPI …
- 18:24 Changeset [5607] by
- Fix MPI_Info object allocation/deallocation. Also, refactor/simplify some …
- 18:24 Changeset [5606] by
- Add an ERRHAND message class for debug logging. Also, add missing ROMIO …
- 18:24 Changeset [5605] by
- Fix MPI_Comm/MPI_Win errhandler leak. Prior to this commit we only …
- 18:24 Changeset [5604] by
- Add errhandler resetting to the commcall test. This is a regression test …
- 17:31 Ticket #909 (MPICH2 shared library builds) created by
- MPICH2's create shared library macros seem to pick the compiler option …
- 17:31 Ticket #797 (Urgent - MPICH2 installer broken) closed by
- fixed
- 17:29 Ticket #891 (Can't initialize with MPI_THREAD_MULTIPLE when not run via mpiexec) closed by
- fixed
- 17:29 Ticket #889 (Mpich2 not running on 64 bit multple node) closed by
- fixed
- 16:56 Changeset [5603] by
- Added more fatal error macros.
- 15:31 Changeset [5602] by
- The case where a kvs get should return an error was not correctly …
- 15:24 Changeset [5601] by
- D'oh! My brain is not working. r5600 was an incorrect "fix".
- 15:10 Changeset [5600] by
- Fix a bug that crept in, in r5591.
- 13:32 Changeset [5598] by
- Disable C++ for the sun compilers since it's broken on breadboard.
- 12:59 Changeset [5597] by
- There was an error in r5554 that didn't actually fix ticket #906. This …
- 11:15 Changeset [5596] by
- Fixed wrong boundaries for loops
- 11:12 Changeset [5595] by
- Fixed a misleading commentary
- 11:10 Changeset [5594] by
- New version of the Hwloc binding file: lot of fixes!
- 03:49 Changeset [5593] by
- Same fix as for the MX Module (see ticket #9O7)
10/27/09:
- 18:10 Changeset [5591] by
- Get rid of the nesting code in PMI-2. This is unnecessary, and just an …
- 18:10 Changeset [5590] by
- Merge equivalent parts of the PMI-1 and PMI-2 code bases.
- 18:10 Changeset [5589] by
- mpiexec -help is a valid command; show the usage, but don't report an …
- 18:10 Changeset [5588] by
- Code style check cleanup.
- 17:59 Changeset [5587] by
- We don't need to run autoreconf on hwloc from maint/updatefiles. It is now …
- 17:57 Changeset [5586] by
- Update config.guess to the latest version.
- 17:57 Changeset [5585] by
- Added depcomp and missing to work with autoreconf.
- 17:53 Changeset [5584] by
- Fix the build problems with using hwloc caused by the fact that hwloc was …
- 16:34 Changeset [5583] by
- Added ltmain.sh (version 2.2.6) to work with libtool.
- 13:07 Ticket #764 (configure failed) closed by
- fixed
- 09:32 Changeset [5576] by
- First version of HWLOC bindings file -- Compiles but needs testing.
- 04:30 Changeset [5575] by
- Forgot to add this file
- 04:22 Ticket #907 (mpich2 1.2 with channel ch3:nemesis:mx is broken) closed by
- fixed: Hi Martin, The problem should be fixed now. The current repository version …
- 04:18 Changeset [5574] by
- Fix for Newmad for one-sided ops.
- 04:18 Changeset [5573] by
- The current MPID_nem_netmod_funcs_t definition doesn't include the …
10/26/09:
- 14:06 Changeset [5572] by
- Use AC_MSG_NOTICE in PAC_CONFIG_SUBDIR. This way the message will also …
- 13:53 Changeset [5571] by
- Enhance the error reporting if the errcode file cannot be opened
- 13:47 Changeset [5569] by
- Extra error checking for MPIU_THREAD_REFCOUNT. No reviewer.
- 13:47 Changeset [5568] by
- Pick better defaults for thread configure values. Fixes shm and ssm …
- 12:52 Changeset [5560] by
- use typedefed name for pthread_mutex_t
- 10:02 Ticket #906 (Nemesis build failure on ppc) closed by
- fixed: Added the comment and committed this in r5554. Resolving.
- 10:01 Changeset [5554] by
- The padding in the nemesis cell structures was assuming that the OPA_ptr_t …
- 09:11 Ticket #908 (Unicode command line arguments) created by
- It appears that MPIEXEC converts UNICODE arguments to 8bit. Consider the …
10/25/09:
- 23:20 Changeset [5553] by
- Added a note about the PMI-2 updates to the CHANGES file.
- 22:45 Ticket #183 (Support for Sun Studio compiler) closed by
- fixed: This should be fixed by r5552 barring other errors that show up on the …
- 22:44 Changeset [5552] by
- Update the svn:externals link for OPA to r96 to bring in the pthread_mutex …
10/24/09:
- 17:04 Changeset [5548] by
- Updates to the release script to remove unnecessary code. Also added an …
- 12:32 Ticket #731 (32-bit Intel C++ compile failures in the mpich2 test suite) closed by
- invalid: With the Intel compiler updated to 11 on breadboard, this error doesn't …
- 12:26 Ticket #867 (Installing multiple versions of MPICH2 on a windows system) closed by
- duplicate: This seems like a duplicate of ticket #901. Resolving.
- 12:21 Ticket #904 (Configure file included in svn?) closed by
- fixed: The configure and other auto-generated files from hwloc are removed. We …
- 12:08 Ticket #907 (mpich2 1.2 with channel ch3:nemesis:mx is broken) created by
- Hi MPICH2_Developers, It seems that there are again problems with the …
10/23/09:
- 10:24 Ticket #906 (Nemesis build failure on ppc) created by
- This bug showed up on the Debian builds -- Nemesis build fails on PPC …
- 01:51 Changeset [5544] by
- The release script doesn't need to have a default path; it should pick …
10/22/09:
- 19:31 Changeset [5543] by
- fixed bug reported by Fab Tillier
- 19:03 Changeset [5542] by
- One more bug fix -- /sandbox might not be a separate partition on machines …
- 18:46 Ticket #905 (mpiexec -machinefile :broken) created by
- Hi all I have made an mich2 cluster, I have one quadcore and three …
- 17:31 Changeset [5541] by
- Include mpichconfconst.h in mpishared.h as in r5492. A classic example of …
- 16:10 Changeset [5538] by
- fixed bug where int was passed in place of str
- 12:46 Changeset [5533] by
- Fix a bug in the raid-probe script to work on machines that don't have a …
- 12:46 Changeset [5532] by
- Add a bit more debug messages for failure cases.
- 11:35 Changeset [5531] by
- Missed two files in hwloc from the 0.9.1rc2 release because of my git-svn …
- 10:32 Changeset [5530] by
- Update hwloc to the 0.9.1rc2 release. This fixes a bunch of issues with …
- 09:55 Changeset [5529] by
- squashed warnings about type-punning
- 09:42 Changeset [5528] by
- r5514 had fixed a bug with the incorrect endif placement in r5508, but …
10/21/09:
- 18:45 Changeset [5527] by
- Get rid of GNU'ism in macro definitions.
- 17:51 Changeset [5526] by
- Remove one more unused macro. We should try to remove MPICH-specific …
- 16:43 Changeset [5525] by
- Get rid of unused macros and ones which are already provided by autoconf.
- 16:33 Changeset [5524] by
- Define MPIU_Handle_get_kind_str unconditionally. Another good catch by …
- 16:33 Changeset [5523] by
- Fix bad format string specifier in SendQ log msg from r5506. A good catch …
- 16:33 Changeset [5522] by
- A few more bullets for the CHANGES file. No reviewer.
- 15:47 Changeset [5521] by
- Accidentally committed two unnecessary checks in r5520. Reverting those …
- 15:46 Changeset [5520] by
- Add useful configure checks automatically so packages using MPL don't need …
- 15:34 Changeset [5519] by
- Get rid of PAC_PROG_MKDIR_P and use AC_PROG_MKDIR_P instead. It seems to …
- 14:25 Changeset [5518] by
- Only check buffer aliasing at the root. Thanks to Kenneth Inghram for …
- 11:36 Changeset [5517] by
- Added changes log for 1.2.1.
- 09:59 Changeset [5516] by
- We don't need two ifdef cases for MPIU_VG_PRINTF_BACKTRACE, since both of …
10/20/09:
- 20:31 Changeset [5515] by
- Separately run autoreconf in the hwloc directory.
- 20:22 Changeset [5514] by
- Fix a bug that got introduced in r5508 causing some MPIU_VG macros to not …
- 19:30 Changeset [5513] by
- Added a new autogen.sh script to run autoreconf correctly in all …
- 19:30 Changeset [5512] by
- Use a sandbox configure for hwloc instead of adding it to CONFIG_SUBDIRS. …
- 18:46 Changeset [5511] by
- Added a PAC macro to allow for sandbox sub-configures. This allows us to …
- 15:35 Changeset [5510] by
- Avoid leaking another context ID. This fixes the leak in the …
- 15:35 Changeset [5509] by
- Better valgrind integration: object labeling and mempools. Valgrind …
- 15:35 Changeset [5508] by
- Add more valgrind client request macros. These mostly have to do with …
- 15:35 Changeset [5507] by
- Quiet valgrind false positives in MPIU_trdump. No reviewer.
- 15:35 Changeset [5506] by
- Nemesis SendQ fixes for multithreading. This change adds debug logging to …
- 15:35 Changeset [5505] by
- Twiddle some ref counting to keep valgrind happy. These were not strictly …
- 15:35 Changeset [5504] by
- Add HANDLE_SET_MPI_KIND macro for parity. No reviewer.
- 15:35 Changeset [5503] by
- Use proper allocation routines for predefined pair-types. For some reason …
- 15:35 Changeset [5502] by
- Zero out free'd user-defined error handlers in MPI_Finalize. Otherwise …
- 15:35 Changeset [5501] by
- Improve handle reference counting log messages. No reviewer.
- 15:35 Changeset [5500] by
- Don't refcount predefined objects when using lock-free refcounting. This …
- 15:35 Changeset [5499] by
- Make lock-free reference counting work again. There were several layers …
- 15:35 Changeset [5498] by
- Twiddle the include guard in mpiatomic.h. No reviewer.
- 15:35 Changeset [5497] by
- Remove references to non-existent OPA_busy_wait() macros. At some point …
- 15:35 Changeset [5496] by
- Use MPIU_QUOTE instead of MPIDI_QUOTE in mpiatomic.h. No reviewer.
- 15:35 Changeset [5495] by
- Delete FUNC_{ENTER,EXIT} macros from mpiatomic.h. I'm deleting them …
- 15:35 Changeset [5494] by
- Move the CONTEXTID critical section macros to mpiimplthread.h. This is …
- 15:35 Changeset [5493] by
- Delete the BRIEF_GLOBAL thread granularity level. It provides little to …
- 15:35 Changeset [5492] by
- Refactor configure-used constants to a new header included before …
- 15:34 Changeset [5491] by
- Update OPA to r94 to pick up the new static initializers. No reviewer.
- 14:29 Changeset [5490] by
- Use the svn version of hwloc-0.9.1rc1 instead of the release tarball. …
- 11:31 Changeset [5489] by
- added srcid field to pmi2 get
- 03:20 Changeset [5488] by
- Move the include of mpif.h so that the Fortran 90 versions that use the …
10/19/09:
- 22:04 Ticket #887 (MPID_nem_init assumes malloc(0) returns a valid pointer) closed by
- fixed: I've (with Jayesh's help in spotting a bug) fixed the macros to only …
- 21:59 Ticket #899 (Some important test links broken) closed by
- fixed: This may be fixed - we'll see the next time the Web server loses the NFS …
- 21:57 Ticket #904 (Configure file included in svn?) created by
- I get the following when I do an update: […] There is a configure.ac …
- 17:33 Changeset [5487] by
- renamed pmi v2 functions to PMI2_; removed mpi error codes from pmi2
- 16:31 Ticket #903 (Hydra: MPICH_INTERFACE_HOSTNAME support) created by
- Hydra currently doesn't have support for MPICH_INTERFACE_HOSTNAME that …
- 14:24 Changeset [5486] by
- Rename misleading variable and function names.
- 14:24 Changeset [5485] by
- The hostfile parsing code was not common between the UI and the RMK which …
- 12:58 Ticket #902 (mpd PM post-execution error: close failed in file object destructor) created by
- I've been testing MPICH2-1.2 for the upcoming Fedora 12 release and have …
- 03:05 Changeset [5484] by
- Added check for nbytes_ > 0 to the memory macros to allow the use of 0 …
- 01:29 Ticket #375 (Need to define C datatypes in Fortran's mpif.h) closed by
- fixed: I've improved the test, fixed some bugs in the existing Datatype support, …
10/18/09:
- 14:21 Changeset [5483] by
- Simplify the directory structure for MPL.
- 11:50 Changeset [5482] by
- Fix a copy-paste error resulting in compilation failure when plpa is not …
- 11:15 Changeset [5481] by
- Move macro definitions out of aclocal.m4 to aclocal_util.m4 (aclocal.m4 …
- 08:18 Changeset [5480] by
- Added an initial draft of the hwloc process binding library as an …
10/16/09:
- 13:52 Changeset [5479] by
- Added (or corrected) support for C versions of MPI Datatypes in Fortran
- 13:50 Changeset [5478] by
- Correct the datatype name tests for both C (make sure that all required …
- 11:36 Ticket #896 (smpd is not to find in vista task manager) closed by
- fixed
- 11:36 Ticket #894 (Why does nemesis build set SO_REUSEADDR on Windows?) closed by
- fixed
- 11:35 Changeset [5477] by
- Providing the MPICH2 channel info to SMPD manager when spawning from …
- 11:31 Changeset [5476] by
- Printing out the context name and event retval instead of a context …
- 11:00 Changeset [5475] by
- Code indentation - no src code change
- 09:21 Ticket #901 (MPICH2 multi-version install on windows) created by
- Currently we don't support installing multiple versions of MPICH2 on a …
- 09:14 Ticket #897 (smpd cannot be installed on windows with a user specified port) closed by
- fixed: Hi, I am closing this ticket and we can continue our discussion on the …
10/15/09:
- 23:44 Changeset [5474] by
- Use smaller variable names where possible.
- 23:40 Changeset [5473] by
- Get rid of unnecessary prefixes in the debug output.
- 23:24 Changeset [5472] by
- Use a consistent naming convention everywhere as a part of the overall …
- 19:10 Ticket #900 (mpecc.in and mpefc.in macros not expanded after installation) closed by
- fixed: Replying to Deji Akingunola <dakingun@…>: > Replying to …
- 18:50 Changeset [5471] by
- remove r5470. Added better support to avoid installing all *.in that …
- 17:42 Changeset [5470] by
- avoid installing mpecc.in and mpefc.in during inpath build.
- 13:25 Changeset [5469] by
- Verbose output for the nightly test logs.
- 13:24 Changeset [5468] by
- Bug fix with the new format of fulltests.
- 12:10 Changeset [5467] by
- White space cleanup.
- 12:10 Changeset [5466] by
- We now separate out the basic utility functions with tools. This makes it …
- 12:10 Changeset [5465] by
- Allow only contiguous segments within each proxy. Multiple contiguous …
- 11:36 Changeset [5464] by
- Added style/copyright/credit
- 10:53 Ticket #900 (mpecc.in and mpefc.in macros not expanded after installation) created by
- If MPE2 is enabled during build, both mpecc.in and mpecc.in under …
- 10:29 Changeset [5463] by
- Added a test program for use of MPI_Comm_disconnect, provided by the …
- 09:19 Ticket #899 (Some important test links broken) created by
- On page http://wiki.mcs.anl.gov/mpich2/index.php/Developer_Documentation , …
10/14/09:
- 21:43 Ticket #898 (MPICH2 Compilation error with --enable-debuginfo) created by
- When I compile it on my Linux laptop with : […] I get : […]
- 16:46 Changeset [5462] by
- Enabling user selection of the PM port during installation on windows. …
- 07:10 Changeset [5461] by
- White space changes.
- 07:03 Changeset [5460] by
- Fix more memory leaks.
- 06:39 Ticket #893 (Hydra: Failures with repeating hostname) closed by
- fixed: This should be fixed in a series of commits up to r5455. Resolving.
- 06:36 Changeset [5459] by
- Fix some memory leaks.
- 06:36 Changeset [5458] by
- Unify structure initialization.
- 06:35 Changeset [5457] by
- Rearrange functionality into more appropriate files.
- 06:35 Changeset [5456] by
- Remove the pgid variable for now. It's currently unused since we don't …
- 06:35 Changeset [5455] by
- Fix a bug in multi-node launches that got introduced in r5446. Cleaned up …
- 06:35 Changeset [5454] by
- Organize the proxy parameters to allow for better structure sharing …
- 06:35 Changeset [5453] by
- Change variable name to maintain consistency with the other names.
10/13/09:
- 09:41 Ticket #897 (smpd cannot be installed on windows with a user specified port) created by
- SMPD does not read the "-port" option during the install (smpd -install …
10/12/09:
- 17:16 Changeset [5452] by
- noopt should default to the same as "all" without -O2. Didn't catch this …
- 17:10 Changeset [5451] by
- Added a noopt flag to --enable-strict to allow for all flags, except -O2 …
- 17:04 Changeset [5450] by
- Added non-verbose make support.
- 16:57 Ticket #896 (smpd is not to find in vista task manager) created by
- Dear mpich2 team, I installed your 32bit Binary Version of mpich2 on my …
- 16:48 Changeset [5449] by
- Warning squashes from a strict compile.
- 13:43 Changeset [5447] by
- Check the number of processes launched to see if a proxy is active, …
10/11/09:
- 18:33 Changeset [5446] by
- This is the second part of the cleanup to decouple the proxy/host …
- 15:32 Changeset [5445] by
- Modified the binding code to deal with the querying of all information …
- 00:44 Changeset [5444] by
- This is the first part of the cleanup to decouple the proxy/host …
10/10/09:
- 23:19 Changeset [5443] by
- Some more cleanup to the environment utilities.
- 20:12 Changeset [5442] by
- Redid parts of the environment management code to make it more manageable, …
10/09/09:
- 16:02 Changeset [5441] by
- Added support for non-verbose make.
- 15:00 Changeset [5440] by
- Removing REUSEADDR option in windows tcp network module. In winsock …
- 09:38 Ticket #895 (Failure to send/recv largish messages on Windows using nemesis sockets) created by
- See attached program - when run across two windows boxes, using the …
- 05:29 Ticket #894 (Why does nemesis build set SO_REUSEADDR on Windows?) created by
- Testing mpich2-1.2 on Windows using the nemesis channel, I find that if I …
10/08/09:
- 13:47 Changeset [5439] by
- The bballoc script has a locking bug as per Rick Bradshaw that's causing …
- 13:46 Changeset [5438] by
- The new images all have a /soft link already created. We don't need to do …
- 13:45 Changeset [5437] by
- Updates to the test list to use the new format of fulltests.pl with …
- 12:36 Changeset [5436] by
- Merge r5435 from the trunk -- Updated the README to mention the 8 EP …
- 12:35 Changeset [5435] by
- Updated the README to mention the 8 EP restriction in MX and removed it …
- 09:06 Changeset [5434] by
- Merge r5432 and r5433 from trunk -- fixes to the MX compilation error …
- 05:01 Changeset [5433] by
- The previous commit was a fix for ticket #892. Some experimental code was …
- 04:44 Ticket #892 (Compilation error while compiling mpich2-1.2 with ch3:nemesis:mx device) closed by
- fixed
- 04:42 Changeset [5432] by
- Forgot to declare a variable
Note: See TracTimeline
for information about the timeline view.
