April 2008 - Mpich2-core - mailman3.cels.anl.gov

Guillaume visit
by Darius Buntinas 21 Apr '08

21 Apr '08

Guillaume is thinking of visiting June 16-27. Is this a good time for everyone? -d

3 2

Restart conversation with IBM
by Pavan Balaji 20 Apr '08

20 Apr '08

Given that 1.0.7 is out and the EuroPVM and SC deadlines are up, should we restart the conversation with IBM with respect to merging BGP related to changes? I saw that there was some discussion regarding ROMIO, but we should probably restart discussion on all parts of the stack. -- Pavan -- Pavan Balaji http://www.mcs.anl.gov/~balaji

5 7

test suites for sicortex
by Darius Buntinas 17 Apr '08

17 Apr '08

SiCortex is asking for tarballs of the test suites we use. They need them for the acceptance tests. I send them a link from the old website: http://www-unix.mcs.anl.gov/mpi/mpi-test/tsuite.html But the tarballs are not all up to date. Where can I find the most up to date tarballs of the test suites? MPICH2: I know where to find that one :-) Intel: We have it on svn, but do we make periodic tarballs? C++: I found this on the old cvs, but, again, are there up to date tarballs? MPICH: Is this in the MPICH(1) tarball? Thanks, -d

2 1

Resolving svn conflicts
by William Gropp 16 Apr '08

16 Apr '08

I'm changing the handling of THROWS_EXCEPTIONS to use MPIR_ instead of MPI_ (which is consistent with other places where we have the same issue) and one of my updates ran afoul of a separate fix. I've addressed all of the conflict markers in the files, but svn status still shows that files as type "C". What else do I need to do to resolve conflicts? Bill William Gropp Paul and Cynthia Saylor Professor of Computer Science University of Illinois Urbana-Champaign

3 3

Re: [mpich2-announce] Announcing the availability of MPICH2-1.0.7
by Pavan Balaji 07 Apr '08

07 Apr '08

Thanks Seppo. Cc'ing mpich2-core for everyone's information. -- Pavan On 04/07/2008 01:52 PM, Seppo Sahrakorpi wrote: > Hi Pavan, > > Thanks. I have smoke tested your earlier pre-releases and things looked > OK. We will include this final version to our more thorough testing > system and hopefully we can declare support soon. I will keep you updated. > > Thanks, > Seppo > > On 04/04/2008 09:43 PM, Pavan Balaji wrote: >> Hi all, >> >> A new release of MPICH2 (1.0.7) is available from >> http://www.mcs.anl.gov/research/projects/mpich2/downloads/tarballs/mpich2-1…. >> Please try it out and let us know if you encounter any problems. The >> major changes since 1.0.6p1 are highlighted below. >> >> (Windows binaries are available from >> http://www.mcs.anl.gov/research/projects/mpich2/downloads/tarballs/mpich2-1… >> and >> http://www.mcs.anl.gov/research/projects/mpich2/downloads/tarballs/mpich2-1…) >> >> >> Regards, >> The MPICH2 Team >> >> --------------------------------------------------------------------- >> Main changes in 1.0.7 >> ===================== >> >> # OVERALL: Initial ROMIO device for BlueGene/P (the ADI device is >> also added but is not configurable at this time). >> >> # OVERALL: Major clean up for the propagation of user-defined and >> other MPICH2 flags throughout the code. >> >> # OVERALL: Support for STI Cell Broadband Engine. >> >> # OVERALL: Added datatype free hooks to be used by devices >> independently. >> >> # OVERALL: Added device-specific timer support. >> >> # OVERALL: make uninstall works cleanly now. >> >> # ROMIO: Support to take hints from a config file >> >> # ROMIO: more tests and bug fixes for nonblocking I/O >> >> # PM/PMI: Added support to use PMI Clique functionality for process >> managers that support it. >> >> # PM/PMI: Added SLURM support to configure to make it transparent to >> users. >> >> # PM/PMI: SMPD Singleton Init support. >> >> # WINDOWS: Fortran 90 support added. >> >> # SCTP: Added MPICH_SCTP_NAGLE_ON support. >> >> # MPE: Updated MPE logging API so that it is thread-safe (through >> global mutex). >> >> # MPE: Added infrastructure to piggyback argument data to MPI states. >> >> # DOCS: Documentation creation now works correctly for VPATH builds. >> >> # Many other bug fixes, memory leak fixes and code cleanup. A full >> list of changes is available using: >> svn log -r100:HEAD >> https://svn.mcs.anl.gov/repos/mpi/mpich2/branches/release/MPICH2_1_0_7 >> ------------------------------------------------------------------- >> > -- Pavan Balaji http://www.mcs.anl.gov/~balaji

1 0

1.0.7 tarball
by Pavan Balaji 07 Apr '08

07 Apr '08

We decided to push out the 1.0.7 release today itself. The Nemesis bug has been added to the known issues for this release. I've created the tarball and placed it in my MCS home directory: /homes/balaji/mpich2-1.0.7.tar.gz Anthony -- can you check if this looks OK? Darius -- can you run some Nemesis tests to make sure it works fine? Jayesh -- can you create the windows binaries with this tarball? I'll check the sock version to make sure everything is fine. -- Pavan -- Pavan Balaji http://www.mcs.anl.gov/~balaji

4 11

NAS BTIO Full Failure
by Jonathan L. Perkins 03 Apr '08

03 Apr '08

MPICH2 Group: During mvapich2 testing we ran into a problem with the NAS benchmark. We've been able to duplicate this issue with mpich2-1.0.7rc2. I built mpich2 with the default settings (./configure && ./make). We executed the test using 4 processes across two nodes. We set the class to B. The test fails due to a collective abort early in the benchmark and generates 4 core files. Each core file presented the same backtrace as below... [perkinjo@ro0-cr NPB3.2-MPI]$ gdb -batch bin/bt.B.4.mpi_io_full core.9445 -x gdb.cmd (no debugging symbols found) Using host libthread_db library "/lib64/tls/libthread_db.so.1". warning: core file may not match specified executable file. Core was generated by `./bin/bt.B.4.mpi_io_full'. Program terminated with signal 11, Segmentation fault. (no debugging symbols found) (no debugging symbols found) (no debugging symbols found) (no debugging symbols found) (no debugging symbols found) (no debugging symbols found) (no debugging symbols found) #0 0x0000000000437dbb in ADIOI_Calc_my_req () #0 0x0000000000437dbb in ADIOI_Calc_my_req () #1 0x00000000004358ef in ADIOI_GEN_WriteStridedColl () #2 0x00000000004634c1 in MPIOI_File_write_all () #3 0x000000000042f474 in PMPI_File_write_at_all () #4 0x0000000000417c5c in pmpi_file_write_at_all__ () #5 0x00000000004171d9 in output_timestep__ () #6 0x0000000000402e47 in MAIN__ () #7 0x0000000000481a42 in main () Have you guys seen this problem before? -- Jonathan Perkins http://www.cse.ohio-state.edu/~perkinjo

2 1

1.0.7 release delayed
by Pavan Balaji 03 Apr '08

03 Apr '08

Hi all, As I had mentioned to some of you yesterday, some of the changes hadn't gone into the release branch, and hence were never tested. We'll be delaying the release some more till the entire test run goes through fine. This might take several days, so I'll send out an update once we are closer. -- Pavan -- Pavan Balaji http://www.mcs.anl.gov/~balaji

1 0

Re: [Systems Req #49467] atlantis (fwd)
by Anthony Chan 02 Apr '08

02 Apr '08

Atlantis is back up just in case if we are still doing nightly test for the release on the machine... A.Chan ---------- Forwarded message ---------- Date: Wed, 2 Apr 2008 12:17:52 -0500 (CDT) From: Ken Raffenetti <raffenet(a)mcs.anl.gov> Reply-To: systems(a)mcs.anl.gov To: chan(a)mcs.anl.gov Cc: systems(a)mcs.anl.gov Subject: Re: [Systems Req #49467] atlantis Anthony, I borrowed some power from a neighboring rack to bring atlantis back up. Ken > > > Is atlantis still down. Can't ssh to it: > > /homes/chan/ts_C> ssh atlantis.mcs.anl.gov > ssh: connect to host atlantis.mcs.anl.gov port 22: No route to host > > Thanks, > A.Chan > >

1 0

Preparation for the 1.0.7 release
by Pavan Balaji 02 Apr '08

02 Apr '08

Hi all, I've updated the documentation and committed changes that need to go on the website for the 1.0.7 release into the SVN. Several files that differ between the trunk and the 1.0.7 branch (listed below). Are these supposed to be left out of 1.0.7? Given the pain in maintaining the release branch and the trunk, I would like to bring up my original discussion on when the branch should be created again -- I still prefer creating a branch at the time of a full release, and not at the time of an RC or earlier. Though Dave volunteered to maintain the release branch, it has gotten quite difficult for one person to take care of this. People who make the commits forget to merge it into the release branch and someone has to keep checking the commits and reminding people about them -- this is a pain. It's also a big headache to keep finding what all is missing in the release branch and making sure everything is consistent. I understand that other software organizations might be branching early, but given the limited developer resources we have, we need to be more conservative with respect to how long we want to maintain multiple branches. Consider how little benefit we had for the 1.0.7 release with respect to being able to commit to trunk, as opposed to the amount of extra work we had to put in with respect to maintaining two different branches. -- Pavan ----------------------------------------------------- MPE: src/mpe2/src/slog2sdk/aclocal_java.m4 ROMIO: src/mpi/romio/test/file_info.c src/mpi/romio/adio/include/adioi.h src/mpi/romio/adio/common/ad_fstype.c src/mpi/romio/adio/common/ad_read_coll.c src/mpi/romio/adio/common/ad_aggregate.c src/mpi/romio/adio/common/ad_open.c src/mpi/romio/adio/common/ad_write_coll.c src/mpi/romio/adio/common/ad_hints.c src/mpi/romio/adio/ad_bgl/ad_bgl_rdcoll.c src/mpi/romio/adio/ad_bgl/ad_bgl_wrcoll.c OVERALL: src/mpid/ch3/setup_device NEMESIS: src/mpid/ch3/channels/nemesis/nemesis/net_mod/tcp_module/tcp_module_poll.c src/mpid/ch3/channels/nemesis/nemesis/src/mpid_nem_init.c ----------------------------------------------------- -- Pavan Balaji http://www.mcs.anl.gov/~balaji

6 7