Timeline
05/11/09:
- 21:54 Changeset [4516] by
- Update openpa to r68. This should improve support for BG/P. No reviewer.
- 20:38 Changeset [4515] by
- disabling poll-avoidance optimization in tcp and wintcp when called from a …
- 18:13 Changeset [4514] by
- Add a --with-automake command to the release.pl script. Reviewed by …
- 17:54 Changeset [4513] by
- No more tee'ing so failure in a command is correctly reported on 0
- 17:42 Ticket #595 (xlc man page dumped repeatedly in config.log) created by
- This was mentioned in ticket #362 but the main point of that ticket was …
- 17:20 Ticket #594 (release.pl script exit status) created by
- The run_cmd subroutine of the maint/release.pl script uses tee to append …
- 17:09 Changeset [4512] by
- added svn:ignore properties
- 16:29 Ticket #593 (Compile error in MPICH2 1.1b1 with external PMPI library) created by
- […]
- 15:29 Changeset [4511] by
- Fix global symbol errors reported by checkforglobs. (mostly) fixes ticket …
- 14:42 Ticket #292 (Some thread tests hang with Nemesis and SMPD) closed by
- fixed: See c4502, c4504, c4505, c4506, c4507, c4510 on trunk for details. SMPD …
- 14:38 Ticket #29 (nemesis ext_procs optimization) closed by
- fixed: Ported the changes to windows netmod in r4509 -jayesh
- 14:36 Changeset [4510] by
- Adding the include paths for smpd sock utils to windows makefile. This …
- 14:30 Changeset [4509] by
- Manually porting c4429 & c4431 for improving shared mem perf on Nemesis to …
- 14:28 Changeset [4508] by
- Removed the control system which was mainly just passing through commands …
- 14:20 Ticket #592 (Shared library build for MPICH2 1.0.8p1 failing on Solaris) created by
- Shared library builds are failing on Solaris. Here's the relevant part of …
- 14:16 Changeset [4507] by
- Fixing dependencies in windows VS proj files after r4502. Also adding a …
- 13:44 Changeset [4506] by
- 1) Removing Makefile.in - accidently checked-in during merge (2) Use PMPI* …
- 13:37 Changeset [4505] by
- dbg_printf.c still depends on device+channel defns since it includes …
- 13:27 Changeset [4504] by
- Handling unhandled defns in smpd configure in windows configure script
- 12:55 Changeset [4503] by
- warning stomp
- 12:52 Changeset [4502] by
- # Merging pm-sock-util branch to trunk - SMPD no longer uses device …
- 11:25 Changeset [4500] by
- added 'test' to list of output for clmake to ignore
- 11:13 Changeset [4499] by
- fix warning on intel compilers about checking that an unsigned is >= 0
- 11:13 Changeset [4498] by
- non-error behavior now much less chatty
- 10:40 Ticket #470 (add OpenPA to trunk, replace MPIDU_Atomic_* with OPA_*) closed by
- fixed: Windows support was integrated in r4493 (Thanks Jayesh!). Resolving.
- 10:20 Changeset [4497] by
- Minor fix to conform to line length standard
- 09:44 Ticket #514 (Fortran 90 module support on Windows MPICH2 x86_64) closed by
- fixed: This should be fixed as of r4410 -Jayesh
- 09:41 Ticket #366 (can not find function MPI_Type_create_f90_real) closed by
- fixed: This should be fixed as of r4406 - trunk. -Jayesh
- 09:00 Changeset [4495] by
- Remove error message introduced in r4331. Empty node arguments is not an …
- 04:04 Ticket #591 (problem about running mpich2-1.0.7 on linux) created by
- […]
05/10/09:
- 19:04 Ticket #585 (Hydra: Help output) closed by
- fixed: Fixed in r4494. Resolving.
- 19:03 Changeset [4494] by
- Several improvements to the help and error messages for user-passed …
- 18:07 Changeset [4493] by
- Modifying windows configure & build scripts to recognize OpenPA. Fixes …
- 14:11 Ticket #589 (Hydra: Bootstrap assumptions) created by
- Hydra's upper layers currently assume that the bootstrap server is too …
- 11:12 Changeset [4492] by
- Handle timeout expiration correctly. If the user specified a timeout and …
- 01:54 Ticket #586 (Hydra: print-all-exitcodes) closed by
- fixed: Fixed in r4491. Resolving.
- 01:53 Changeset [4491] by
- Pass the exit codes for all processes to the UI. The UI can print all of …
- 01:42 Changeset [4490] by
- Fixed an error in the blocking socket read call; when the socket itself is …
- 00:34 Changeset [4489] by
- Accidentally deleted some code which sets necessary sockets to …
- 00:31 Changeset [4488] by
- The sock routines were not explicitly checking to see if all the data that …
05/09/09:
- 22:57 Changeset [4487] by
- Update version number in preparation for the internal release.
- 22:56 Changeset [4486] by
- Use similar naming convention between the server and the proxies.
- 22:53 Changeset [4485] by
- Better naming for some variables to avoid confusion.
- 22:42 Ticket #575 (mpiexec rank mapping output feature in hydra) closed by
- fixed
- 22:41 Changeset [4484] by
- Added support for --print-rank-map. Ticket #575.
- 21:04 Changeset [4483] by
- Fixed the exit status passing. Now the exit status is blindly passed the …
- 20:33 Changeset [4482] by
- Added better support for verbose output. Now, we pass the debug status to …
- 20:16 Changeset [4481] by
- Installed both autoconf and automake into a common directory; MPICH2 needs …
- 20:14 Ticket #588 (add --with-automake option to maint/updatefiles) created by
- With openpa we now have a dependency on automake as well. Currently the …
- 11:18 Changeset [4480] by
- updated names of hints
- 10:59 Ticket #501 (Multiple process manager default) closed by
- fixed: Fixed in r4479. Resolving.
- 10:59 Changeset [4479] by
- Allow multiple process managers to be compiled in by default. The default …
- 10:36 Ticket #587 (Nemesis test failures) created by
- […]
- 01:35 Changeset [4478] by
- Stomp some valgrind warnings.
- 01:13 Changeset [4477] by
- Warning stomp and fixes to some typos that creeped in.
- 01:03 Changeset [4476] by
- Together with the user-specified and the default path, we now also look in …
- 00:29 Changeset [4475] by
- Accidentally committed in a patch that doesn't belong to the IBM change.
- 00:28 Changeset [4474] by
- IBM's patch for improvements to the dcmfd device. Thanks!
- 00:23 Ticket #544 (Hydra: pmi_proxy path detection) closed by
- fixed
- 00:22 Changeset [4473] by
- Hackish fix for ticket #544. If mpiexec is used from the user's path, …
05/08/09:
- 23:04 Ticket #584 (Hydra: Dynamic port binding) closed by
- fixed: This has been fixed in a series of commits till r4472. Resolving.
- 23:03 Changeset [4472] by
- Support for dynamic port binding. This should fix ticket #584.
- 21:32 Changeset [4471] by
- Got rid of some code duplication.
- 21:15 Changeset [4470] by
- Use a consistent method for passing arguments for both runtime and …
- 20:25 Changeset [4469] by
- Some more code refactoring to allow for RUNTIME proxies to receive …
- 20:03 Changeset [4468] by
- Use the launch mode to know what the proxy type is, instead of creating a …
- 19:46 Changeset [4467] by
- Simplify the persistent connections code to protect only the gethostbyname …
- 18:58 Changeset [4466] by
- Fix SMPD after the opa integration. This change reverses the processing …
- 18:28 Changeset [4465] by
- Updating to openpa@r64. No reviewer.
- 17:46 Changeset [4464] by
- fixed %t parameter in filename for debugging log file. With this option, …
- 17:43 Changeset [4463] by
- log address of mutex in cs enter/exit dbg messages
- 17:41 Changeset [4462] by
- fix for #539 (brief global for nemesis)
- 17:19 Changeset [4461] by
- fixed global symbol name
- 17:16 Changeset [4460] by
- Fix for #582. Update openpa to r63. This brings in change r63 in which …
- 16:48 Changeset [4459] by
- Set env vars to tell autoreconf where to find automake and friends. …
- 16:44 Changeset [4458] by
- updates from Sun for Lustre: - rename a couple hints to be more meaningful …
- 16:29 Changeset [4457] by
- Fix nemesis initialization when using OPA. Also fix a couple of lurking …
- 16:29 Changeset [4456] by
- Add libopa.a as a dependency for cpi (and other examples). This way cpi …
- 16:29 Changeset [4455] by
- Make openpa autoreconf failures fatal for maint/updatefiles. Reviewed by …
- 16:29 Changeset [4454] by
- Wei-keng finds a bug in the mapping of OSTs to I/O clients. Patch by …
- 16:00 Ticket #586 (Hydra: print-all-exitcodes) created by
- We need a way to allow Hydra to print all exit codes from all processes, …
- 15:53 Ticket #577 (64 bit systems) closed by
- wontfix
- 15:50 Ticket #585 (Hydra: Help output) created by
- The help output when the user passes incorrect arguments is pretty naive …
- 15:48 Ticket #584 (Hydra: Dynamic port binding) created by
- We need to dynamically find ports and bind to them in Hydra, instead of …
- 14:52 Changeset [4453] by
- # Adding mpiparam.h which contains the defns exported by util/param # …
- 14:32 Changeset [4452] by
- fix warning
- 11:54 Changeset [4451] by
- from ibm: this index already declared earlier
- 11:52 Changeset [4450] by
- from IBM: we can use the striping unit in common code now, so process that …
- 11:51 Changeset [4449] by
- from IBM: clean up an extern declaration that doesn't need to be there
- 11:41 Changeset [4448] by
- Fix for #582 which was introduced in r4441. maint/updatefiles was …
- 11:08 Ticket #583 (Older MPICH2 porting issues) created by
- […]
- 10:57 Ticket #528 (Some design docs point to obsolete www-unix links) closed by
- fixed: Fixed.
- 10:42 Ticket #582 (Problems with openpa) created by
- I can't build MPICH2 anymore. The openpa package doesn't integrate into …
- 10:25 Changeset [4447] by
- from IBM: sync up with common-code changes. Addresses a test failure for …
- 09:59 Changeset [4446] by
- quite a 'declared but not used' warning
- 09:57 Changeset [4445] by
- renamed this function, but didn't update declaration
- 09:47 Changeset [4444] by
- initialize potentially "used uninitialized" variable
- 09:47 Changeset [4443] by
- update declaration of heapsort funcs to match renamed version
- 09:45 Changeset [4442] by
- remove unused variable
- 00:14 Changeset [4441] by
- Remove MPIDU_Atomic_ primitives code and use OPA_ instead. Reviewed by …
05/07/09:
- 22:29 Changeset [4440] by
- added prototypes to fix warnings
- 22:25 Changeset [4439] by
- fixed hangs in odd-even cliques
- 19:17 Ticket #580 (Parameterize the collective algorithms) created by
- This is a placeholder to complete the implementation of the parameter …
- 18:14 Changeset [4438] by
- fix typo
- 17:33 Changeset [4437] by
- Reverting removal of mpiimpl.h from dbg_printf.c - done in r4432. Without …
- 17:00 Changeset [4436] by
- define alltoall_throttle
- 16:36 Changeset [4435] by
- throttle no. of irecvs/isends posted as suggested by Tony Lad
- 16:36 Changeset [4434] by
- clean up yesterday's changes
- 16:31 Ticket #540 (PMI-2 additions) closed by
- fixed: Done in r4366.
- 15:56 Changeset [4433] by
- Adding ability to explicitly exclude dirs, files from windows build
- 15:48 Changeset [4432] by
- Move MPIU dbg stuff from mpiimpl.h to mpidbg.h . Now utils don't have to …
- 15:28 Changeset [4431] by
- changed inactive poll value for tcp
- 14:53 Changeset [4430] by
- more fixes.
- 14:44 Ticket #29 (nemesis ext_procs optimization) reopened by
- Since the changes are in the tcp network module, I need to port the …
- 14:42 Ticket #284 (Retrofit job attributes to PMI v1.1) closed by
- duplicate: Duplicate of #538.
- 14:39 Ticket #29 (nemesis ext_procs optimization) closed by
- fixed: Fixed in r4429. The original plan was to poll the network only when …
- 14:29 Changeset [4429] by
- fixes #29: In tcp, we're calling poll only once in a long while until we …
- 12:26 Changeset [4427] by
- Reverting change to explicitly define MPI_AINT32 as 32-bit VS intrinsic …
- 09:18 Changeset [4426] by
- guess i was too cute here. call the MPIU_ functions with ifdefs
- 07:37 Ticket #579 (Error launching MPI apps with mpiexec delegate option on windows - Ticket ...) created by
- […]
- 06:57 Ticket #577 (64 bit systems) created by
- […]
- 00:45 Changeset [4425] by
- Code refactoring to function-ize parts that will be reused in other …
05/06/09:
- 23:40 Changeset [4424] by
- Changed proxy_args to exec_args to allow the proxies to use a different …
- 23:21 Changeset [4423] by
- Several changes to separate out information needed by the bootstrap …
- 23:01 Changeset [4422] by
- checking in changes suggested by Tony Lad. May need to change further
- 17:00 Changeset [4421] by
- Add KVS get caching to mpd. This substantially reduces MPI_Init times, …
- 16:50 Changeset [4420] by
- from IBM: another level of debugging
- 16:47 Changeset [4419] by
- from IBM: more (optional) memory management logging
- 16:46 Changeset [4418] by
- 1) Explicitly define MPI_AINT32 with 32-bit int (2) Correct typo - …
- 16:36 Changeset [4417] by
- Fix VS compiler warnings - Handling unreferenced arg & removing return …
- 16:18 Changeset [4416] by
- "we've got a bum commit". fixed. and fixed again.
- 15:57 Changeset [4415] by
- removed unneeded in_or_out parameter from network_poll functions. Removed …
- 15:46 Changeset [4414] by
- from IBM: improve fcntl locking errno handling
- 15:36 Changeset [4413] by
- BGL driver doesn't ever invoke ADIOI_GEN_SetInfo and itself is just a copy …
- 15:20 Ticket #441 (mpiu_shm_wrappers warnings) closed by
- fixed: Fixed by r4412.
- 15:11 Changeset [4412] by
- Fix some printf warnings in gcc 4.3 (ticket #441). Resolves ticket #441, …
- 14:34 Ticket #507 (--enable-g option not recognized) closed by
- fixed
- 14:25 Changeset [4411] by
- Fixed nemesis to correctly set vc state (instead of setting all vcs to …
- 14:18 Changeset [4410] by
- Adding rules for building fortran 90 modules to windows makefile/scripts & …
- 13:58 Changeset [4409] by
- setting fd->comm to COMM_SELF for this create-on-one case is both …
- 13:55 Changeset [4408] by
- "scalable stat" to broadcast file system blocksize information to all MPI …
- 13:07 Changeset [4407] by
- Added AC_CONFIG_AUX_DIR in rlog's configure.in to get rid of install-sh.
- 12:20 Ticket #576 (cleanup rlog configure.in) created by
- […]
- 11:19 Changeset [4406] by
- Adding F90 Extended support - only the type create funcs - sources to …
- 11:17 Changeset [4405] by
- Adding the windows header file for NT intrinsics - to define the …
- 10:52 Ticket #575 (mpiexec rank mapping output feature in hydra) created by
- Scott Atchley recently ran into some trouble with the way mpd was mapping …
- 10:38 Changeset [4404] by
- Clearly separate out upstream connections so we can easily add downstream …
- 10:35 Changeset [4403] by
- Updated header files to get the right definitions for FUNC_ENTER/EXIT …
- 10:17 Changeset [4402] by
- Correcting the format of integer model & integer model map in windows …
- 02:06 Ticket #573 (Fwd: [mpich-discuss] How to build libmpe.so ?) created by
- […]
- 01:49 Ticket #497 (irlog2rlog) closed by
- fixed: fixed the rlog configure.in and Makefile.sm as well as top-level …
- 01:49 Changeset [4401] by
- fixed the rlog configure.in and Makefile.sm as well as top-level …
- 01:43 Ticket #550 (Fwd: [mvapich-discuss] examples_collchk issue (fwd)) closed by
- fixed: fixed in the trunk by r4400. resolved this unless we hear more from user.
- 01:39 Changeset [4400] by
- fixed for req #550.
05/05/09:
- 23:56 Changeset [4399] by
- The vector mapping code was uncommitted earlier.
- 15:03 Changeset [4398] by
- Adding F90 type create funcs - part of Extended fortran support - to …
- 14:50 Changeset [4397] by
- Fixed bug in Makefile if X_INC is null.
- 12:35 Ticket #441 (mpiu_shm_wrappers warnings) reopened by
- RobL took a stab at fixing this in r4391, but Bill noticed that this is …
- 11:36 Ticket #441 (mpiu_shm_wrappers warnings) closed by
- worksforme: I am resolving this for now since we cannot reproduce these warnings. (We …
- 09:09 Ticket #571 (mpi installation) closed by
- invalid: I'm not sure where you downloaded mpich2-1.0 from (it's very old), but you …
- 04:09 Ticket #571 (mpi installation) created by
- […]
- 01:08 Changeset [4396] by
- Adding an initial version of the process mapping strings. This might …
- 00:01 Changeset [4395] by
- Changed launcher to UI.
05/04/09:
- 22:57 Ticket #569 (Hydra with --enable-g) closed by
- fixed: Fixed in r4394. Resolving.
- 22:57 Changeset [4394] by
- Got rid of some ancient debug code that was sitting around causing the …
- 16:22 Ticket #569 (Hydra with --enable-g) created by
- Hydra compile fails with --enable-g.
- 16:05 Changeset [4393] by
- take care of the i/o related non-conforming symbols as raised in #566
- 12:39 Changeset [4392] by
- Disabling PMI2 on windows. SMPD currently does not support PMI2.
- 11:46 Ticket #568 (Cygwin gcc libs for MPICH2 on windows x86_64) created by
- Currently we don't provide Cygwin gcc libs with MPICH2 64-bit on windows. …
- 11:39 Changeset [4391] by
- gcc-4.3 complains with "format not a string literal and no format …
- 11:24 Ticket #567 (Nemesis tests last night) created by
- […]
- 08:25 Changeset [4390] by
- Fix ifdef -> if
05/03/09:
- 22:45 Ticket #548 (ISU016 - mpich2-1.1b1 fails to build on SUN Solaris 5.10 with GNU 3.4.2) closed by
- worksforme
- 22:42 Ticket #556 (I have question) closed by
- wontfix
- 21:24 Changeset [4389] by
- Fix for per-object - at least it now compiles and runs most of the tests
- 19:20 Changeset [4388] by
- In the case where the signal raised by mpiexec does not terminate it, we …
- 16:31 Ticket #518 (Hydra: stderr forwarding for failed processes) closed by
- fixed
- 16:29 Changeset [4387] by
- 1. Modified the proxy to return the exit status as is, instead of trying …
- 07:40 Changeset [4386] by
- More warning stomp.
05/02/09:
- 12:39 Changeset [4385] by
- Missing semi-colon. Shocking that this even compiled so far.
- 10:50 Ticket #566 (Global symbols require renaming) created by
- See http://www.mcs.anl.gov/research/projects/mpich2/todo/globsyms.htm . …
- 02:43 Ticket #511 (Configure check for C and Fortran compiler compatibility) closed by
- fixed: changed the warn action of the F77 and C object file compatibility test to …
- 02:41 Changeset [4384] by
- Replaced AC_MSG_WARN by AC_MSG_ERROR in F77 and C object file …
- 00:38 Ticket #152 (PAC_F90_CHECK_COMPILER_OPTION + pgf90) closed by
- fixed: Added --enable-check-compiler-flags to allow for disabling of the checks …
- 00:22 Changeset [4383] by
- Added --enable-check-compiler-flags option to allow user to disable the …
05/01/09:
- 22:27 Changeset [4382] by
- White space cleanup.
- 22:26 Changeset [4381] by
- Fixes warnings.
- 19:27 Changeset [4380] by
- First draft of nested launching of mpiexec (r4379 was from a wrong …
- 19:26 Changeset [4379] by
- Initial support for nested launching of mpiexec.
- 15:31 Changeset [4378] by
- renamed MPIDI_PG_Get_vc macro to more appropriate …
- 12:53 Changeset [4377] by
- removed dead and invalid PMI routines from pmi.h and from the simple and …
- 12:50 Changeset [4376] by
- Added a test for weak_import
- 12:46 Changeset [4375] by
- Added note about the unexpected reqreuiement to use -cpp with the IBM mpCC …
- 12:45 Changeset [4374] by
- Fixed use of restrict to match the someone silly restrictions in C
- 12:44 Changeset [4373] by
- Fixed error in use of comm parameter
- 12:35 Changeset [4372] by
- Changed == to eq in string comparison
- 11:01 Ticket #557 (Last night's tests) closed by
- fixed: Yeah, my merge of the pmi2 branch broke for sock and nemesis when using …
- 10:58 Ticket #531 (support for multiple executables) closed by
- fixed: I'm closing this ticket. If you still have problems with this, feel free …
- 10:53 Changeset [4371] by
- Fixed compile error when compiling nemesis with pmi1
- 10:43 Ticket #555 (ch3u_init_sock no longer compiles) closed by
- fixed: Fixed in r4370.
- 10:42 Changeset [4370] by
- fixed copy/paste error. fixes #555
- 10:07 Ticket #557 (Last night's tests) created by
- […]
- 06:44 Ticket #556 (I have question) created by
- […]
04/30/09:
- 18:42 Ticket #555 (ch3u_init_sock no longer compiles) created by
- […]
- 18:24 Ticket #554 (MPE: fine-grained logging support) created by
- […]
- 18:16 Ticket #553 (Jumpshot: Replace JScrollbar) created by
- […]
- 17:42 Ticket #552 (Jumpshot right-click zoom support on drawable) created by
- […]
- 17:17 Ticket #551 (MacOSX & jumpshot) created by
- […]
- 17:12 Ticket #550 (Fwd: [mvapich-discuss] examples_collchk issue (fwd)) created by
- […]
- 16:43 Ticket #510 (Resolved: Re: python error) closed by
- worksforme: So it looks like the --with-python problem is that using --with-python is …
- 16:18 Changeset [4369] by
- revert mistaken commit
- 16:16 Changeset [4368] by
- fix mpich2prereq to look in pmi2 (renamed from simple2)
- 15:51 Changeset [4367] by
- added cast for cases where iov_base is not a void*
- 15:46 Changeset [4366] by
- adding PMI2 functionality. PMI2 is still experimental.
- 15:19 Ticket #549 (Update the windows developer's guide) created by
- The MPICH2 windows developer's guide contains a lot of now-irrelevant …
- 14:40 Ticket #548 (ISU016 - mpich2-1.1b1 fails to build on SUN Solaris 5.10 with GNU 3.4.2) created by
- […]
- 12:01 Ticket #484 (Scaling problem with MPI_Comm_create()) closed by
- fixed: I have implemented a fix that checks whether the processes in the group …
- 11:52 Ticket #499 (Add test for ANY_SOURCE + ANY_TAG) closed by
- fixed: I've added a simple test.
04/29/09:
- 16:01 Ticket #503 (Nemesis on niagara) closed by
- worksforme: I just tried this on trunk and it worked. Let me know if you still see …
- 15:17 Changeset [4358] by
- added missing #define of dprintf debugging statement
- 12:33 Changeset [4357] by
- Added the rsh bootstrap server. Also set ignore properties on a bunch of …
- 11:31 Changeset [4356] by
- Added an option to allow the user to pick a non-default path for the …
04/28/09:
- 18:38 Changeset [4353] by
- Fixed support for dependency creation for programs (simplemake did …
- 18:33 Changeset [4352] by
- Added an optimization for MPI_COMM_CREATE for the case where the input …
- 18:29 Changeset [4351] by
- Added comments about why the f77 and f90 buildiface must use DOUBLE …
- 18:27 Changeset [4350] by
- Added some documentation for the MPIU_DBG_MSG_FMT macro
- 18:26 Changeset [4349] by
- Fix the error count check and initialization and add a pointer about why …
- 18:22 Changeset [4348] by
- Added a test that creating a communicator from a group that includes …
- 18:18 Changeset [4347] by
- Added a test that sending data to receives with MPI_ANY_TAG and …
- 18:15 Changeset [4346] by
- Added a test for invoking delete functions on attributes added to comm …
- 18:13 Changeset [4345] by
- Added a test on lower bounds that MPICH2 recently failed on some 64-bit …
- 18:11 Changeset [4344] by
- Added more tests of communicator creation from groups
- 18:11 Changeset [4343] by
- Added options to display rate that comm_split runs and to change loop …
- 17:37 Changeset [4342] by
- Added some more details about using simplemake
- 17:31 Changeset [4341] by
- Update updatefiles to check that we have autoconf 2.62 since some parts of …
- 17:19 Ticket #547 (ISU015 - mpich2-1.1b1/test/mpi/f90/attr/fandcattr* needs update) closed by
- fixed: This fix is wrong; the Cray implementation must be buggy. See …
- 16:50 Ticket #547 (ISU015 - mpich2-1.1b1/test/mpi/f90/attr/fandcattr* needs update) reopened by
- 15:22 Changeset [4340] by
- we haven't really used 'longlongisze' since the pre-mpich2 days
- 13:10 Ticket #547 (ISU015 - mpich2-1.1b1/test/mpi/f90/attr/fandcattr* needs update) closed by
- fixed: Thanks for the report. I've fixed the bug wrt the use of MPI_Aint *val. …
- 13:01 Ticket #401 (BARRIER instructions not generated in CH3 SHM device on X86_64 linux.) closed by
- fixed: Fixed in r4338. Thanks for the bug report and fix!
- 12:59 Changeset [4338] by
- Fix memory barriers for the shm channel on x86_64. Fixes ticket #401. …
- 10:36 Changeset [4337] by
- One more change to the knem tarball mime-type. Hopefully the last one.
- 10:33 Changeset [4336] by
- set mimetype property to x-gzip on tar.gz file
- 10:29 Changeset [4335] by
- set mimetype property on tar.gz file
- 10:28 Changeset [4334] by
- Add missing multisend test to Makefile.sm. The test suite basically still …
04/27/09:
- 15:29 Ticket #474 (spawnmanyargv fails on smpd) closed by
- fixed: This should be fixed as of r4331 on trunk. -jayesh
- 15:28 Changeset [4331] by
- Increasing the length of the executable supported. Without the fix …
- 11:53 Changeset [4330] by
- Added missing include for strcmp
- 11:08 Ticket #547 (ISU015 - mpich2-1.1b1/test/mpi/f90/attr/fandcattr* needs update) created by
- […]
04/26/09:
- 12:19 Changeset [4328] by
- Merged the MPI-UPC branch changes into the trunk. This also includes some …
- 11:59 Changeset [4327] by
- Merging r4321 from the PMI2 branch (the correct KVS Fence implementation).
04/25/09:
- 17:27 Ticket #546 (Problem: MPI setting) closed by
- worksforme: It looks like those guys copied an extremely old version of the MPICH2 …
- 17:07 Ticket #546 (Problem: MPI setting) created by
- […]
- 14:25 Ticket #487 (PMI-1 Initialization sequence) closed by
- wontfix: All process managers which I know use the PMI_PORT functionality of PMI …
04/24/09:
- 20:00 Ticket #305 (Location of the console - hardcoded /tmp) closed by
- fixed: Ralph has fixed this in r4326 of the svn source using MPD_TMPDIR instead …
- 18:32 Changeset [4326] by
- Add support for MPD_TMPDIR (for "advanced" users only :-).
- 18:05 Ticket #541 (bsend of a packed buffer) closed by
- fixed: Thanks. Applied in r4325.
- 18:04 Changeset [4325] by
- Brian's patch to fix bug in bsend of packed buffer, ticket #541
- 16:06 Changeset [4324] by
- 1) Move complex macros to inlines (2) Type cast variables correctly before …
- 16:04 Changeset [4323] by
- Local shared mem handles are always init'ed for MMAP,SYSV,NT shms. …
- 15:45 Changeset [4322] by
- 1) Define the abstract atomic type with a long - Since windows 32-bit …
- 15:43 Ticket #508 (query on MPI testing fail) closed by
- worksforme: I wish I knew what was going on with that bsend1cxx test, but I am unable …
04/23/09:
- 22:44 Ticket #543 (MPI_Alltoallw Datatype checks) closed by
- fixed: fixed in r4319
- 22:44 Changeset [4319] by
- check datatypes only if count>0. fixes #543
- 22:09 Changeset [4318] by
- Got rid of unnecessary debug statements.
- 21:20 Changeset [4316] by
- Merging the PMI-2 server parts from the pmi2 branch.
- 19:39 Ticket #544 (Hydra: pmi_proxy path detection) created by
- Currently, Hydra only handles two cases for detecting the path where the …
- 19:21 Ticket #543 (MPI_Alltoallw Datatype checks) created by
- In mpich2-1.0.8p1/src/mpi/coll/alltoallw.c, lines 361 - 372, error …
- 18:57 Ticket #541 (bsend of a packed buffer) created by
- […]
- 17:30 Ticket #469 (add knem to trunk) closed by
- fixed: Should be resolved by r4310 and r4311. -Dave
- 17:29 Ticket #204 (mpid_abort warning) closed by
- fixed: I believe that Darius fixed this in r4249. -Dave
- 17:27 Ticket #146 (Shared memory capable collectives) closed by
- fixed: This is all implemented enough to be able to close this ticket. Any …
- 17:21 Changeset [4311] by
- Create contrib directory and add knem-0.5.0 there initially. Also add …
- 17:21 Changeset [4310] by
- Initial merge of the knem and vmsplice LMT backends to trunk. This …
- 17:21 Changeset [4309] by
- Cleanup the COOKIE exchange mechanism in the LMT harness. Reviewed by …
- 16:18 Ticket #487 (PMI-1 Initialization sequence) reopened by
- I think there is an actual implementation bug here. If you look at the …
- 15:47 Ticket #487 (PMI-1 Initialization sequence) closed by
- fixed: This is all true. The PMIv1 that was implemented was different than the …
- 14:53 Ticket #540 (PMI-2 additions) created by
- PMI-2 client and server code need to be added.
- 14:46 Ticket #539 (Fine-grain threads) created by
- Add fine-grained threading support to MPICH2.
- 14:44 Ticket #538 (Nemesis MPI_Init scalability issue) created by
- […]
- 14:34 Ticket #534 (Unable to open a file for writing on a Lustre filesystem when configured ...) closed by
- fixed
- 14:24 Ticket #513 (Is MPICH2 connection failure aware ?) closed by
- duplicate: We plan to make MPICH2 more failure aware in the near future. That will …
- 14:13 Ticket #504 (Need help installing mpich2) closed by
- worksforme
- 13:32 Ticket #537 (SIGBUS in typelb test from MPICH1 on 64-bit Solaris) closed by
- fixed
- 11:05 Changeset [4307] by
- Fixed serious error - MAX_MSGS was set to one, making the test useless. …
- 09:53 Ticket #221 (Attribute handling on 64-bit platforms) closed by
- fixed: I've now tested this with Jeff's test (modified to run on Solaris) on the …
- 09:41 Changeset [4306] by
- Fix bug due to inadvertent type casting of an int array to MPI_Aint array. …
- 08:09 Ticket #537 (SIGBUS in typelb test from MPICH1 on 64-bit Solaris) created by
- […]
04/22/09:
- 15:35 Changeset [4303] by
- update lustre function table
- 15:04 Ticket #534 (Unable to open a file for writing on a Lustre filesystem when configured ...) created by
- […]
- 14:17 Changeset [4301] by
- 1) The Bootstrap control device only contacts partitions with …
- 12:18 Ticket #530 (Re: run mpich on two architectures) closed by
- duplicate
- 12:15 Ticket #525 (FW: Fortran attributes) closed by
- fixed: Fixed by Bill in r4299
- 12:14 Ticket #523 (mpd error message) closed by
- worksforme
- 12:13 Ticket #520 (mpicc -o test.c deletes test.c) closed by
- wontfix
- 09:52 Changeset [4299] by
- Change from MPI_Aint to MPIR_Pint to allow MPI_Aint to be a different size …
04/21/09:
- 13:14 Ticket #532 (MPI_ADDRESS_KIND = 0 on linux 64) closed by
- wontfix
- 13:11 Ticket #533 (bitypes.h: No such file or directory) closed by
- fixed
- 10:32 Ticket #533 (bitypes.h: No such file or directory) created by
- […]
04/20/09:
04/19/09:
- 17:50 Changeset [4262] by
- Allow a different parser to be used for PMI-1 and PMI-2. There's still a …
- 16:02 Ticket #531 (support for multiple executables) created by
- […]
- 14:33 Ticket #530 (Re: run mpich on two architectures) created by
- […]
- 11:29 Changeset [4258] by
- renamed pmi2 api functions form PMI2_ to PMI_
04/18/09:
- 17:54 Ticket #529 (mpdgdbdrv.py python deprecation warnings) created by
- […]
- 11:16 Ticket #528 (Some design docs point to obsolete www-unix links) created by
- […]
04/17/09:
- 13:57 Changeset [4257] by
- reverted r4249 (making mpid_abort return void) and instead removed the …
- 11:48 Ticket #526 (Using mpich2 with python and pypar on windows) created by
- […]
- 09:47 Changeset [4256] by
- Clean out old attribute code (#if 0ed for several releases)
- 09:45 Changeset [4255] by
- minor text fix
- 09:45 Changeset [4254] by
- Use the correct extension for the debugger dll
04/16/09:
- 23:08 Changeset [4253] by
- initial coding for pmi2 client interface
- 16:44 Ticket #525 (FW: Fortran attributes) created by
- […]
- 12:46 Ticket #524 (Fwd: mpich2 error ifort makelog) closed by
- fixed: Since you did configure and make as 2 separate users, you and su, the …
- 12:06 Ticket #524 (Fwd: mpich2 error ifort makelog) created by
- […]
- 10:57 Changeset [4250] by
- Removing trailing spaces in the name of perl pgm. This was causing these …
- 09:58 Ticket #523 (mpd error message) created by
- […]
04/15/09:
- 17:33 Changeset [4249] by
- Made MPID_Abort return void to get rid of compiler warnings about noreturn …
Note: See TracTimeline
for information about the timeline view.
