- devel - mailman3.cels.anl.gov

Re: [mpich-devel] Assertion failed in file src/mpid/ch3/channels/nemesis/src/ch3_progress.c at line 643: pkt->type >= 0 && pkt->type < MPIDI_CH3_PKT_END_ALL internal ABORT - process 1
by Jeff Hammond 29 Dec '15

29 Dec '15

Build info: MPICH Version: 3.2b3 MPICH Release date: Thu Jun 4 07:59:11 CDT 2015 MPICH Device: ch3:nemesis MPICH configure: --prefix=/usr --sysconfdir=/etc --libdir=/usr/lib --mandir=/usr/share/man --infodir=/usr/share/info --docdir=/usr/share/doc/mpich --htmldir=/usr/share/doc/mpich --dvidir=/usr/share/doc/mpich --pdfdir=/usr/share/doc/mpich --psdir=/usr/share/doc/mpich MPICH CC: gcc -O2 MPICH CXX: g++ -O2 MPICH F77: gfortran -O2 MPICH FC: gfortran -O2 gcc -I/usr/include -L/usr/lib -Wl,-rpath -Wl,/usr/lib -Wl,--enable-new-dtags -lmpi On Tue, Dec 29, 2015 at 12:44 PM, Jeff Hammond <jeff.science(a)gmail.com> wrote: > I do not, but I guess Marcelo Soares Souza does ( > http://lists.mpich.org/pipermail/discuss/2015-June/004045.html) > > Travis will log the output of mpichversion the next time it runs... > > Jeff > > On Tue, Dec 29, 2015 at 12:17 PM, Balaji, Pavan <balaji(a)anl.gov> wrote: > >> >> Do you have more details on how mpich was built? >> >> -- Pavan >> >> > On Dec 29, 2015, at 2:05 PM, Jeff Hammond <jeff.science(a)gmail.com> >> wrote: >> > >> > I am using mpich_3.2b3-1ubuntu_amd64.deb within Travis CI for ARMCI-MPI >> testing. I never saw this problem in Jenkins, so I can't tell if the >> problem is specific to the deb build, to Travis, or something else. >> > >> > Travis page has all the details so I will not repeat them here. >> > >> > Thanks, >> > >> > Jeff >> > >> > From https://travis-ci.org/jeffhammond/armci-mpi/jobs/99316268: >> > >> > Processing triggers for libc-bin ... >> > ldconfig deferred processing now taking place >> > +wget -q >> http://www.cebacad.net/files/mpich/ubuntu/mpich-3.2b3/mpich_3.2b3-1ubuntu_a… >> > +sudo dpkg -i ./mpich_3.2b3-1ubuntu_amd64.deb >> > Selecting previously unselected package mpich. >> > (Reading database ... 124680 files and directories currently installed.) >> > Unpacking mpich (from .../mpich_3.2b3-1ubuntu_amd64.deb) ... >> > Setting up mpich (3.2b3-1ubuntu) ... >> > Processing triggers for man-db ... >> > >> > >> > FAIL: benchmarks/contiguous-bench (exit: 1) >> > =========================================== >> > ARMCI-MPI initialized with 2 processes, MPI v3.1 >> > PROGRESS_THREAD = DISABLED >> > ALLOC_SHM used = TRUE >> > WINDOW type used = CREATE >> > STRIDED_METHOD = DIRECT >> > IOV_METHOD = DIRECT >> > IOV_CHECKS = FALSE >> > SHR_BUF_METHOD = COPY >> > NONCOLLECTIVE_GROUPS = FALSE >> > CACHE_RANK_TRANSLATION = TRUE >> > DEBUG_ALLOC = FALSE >> > Starting one-sided contiguous performance test with 2 processes >> > Trg. Rank Xfer Size Get (usec) Put (usec) Acc (usec) Get >> (MiB/s) Put (MiB/s) Acc (MiB/s) >> > 1 8 0.869 0.707 0.935 >> 8.783 10.795 8.158 >> > 1 16 0.926 0.768 0.914 >> 16.482 19.859 16.698 >> > 1 32 0.945 0.856 0.974 >> 32.284 35.671 31.336 >> > 1 64 0.938 0.781 0.956 >> 65.097 78.199 63.852 >> > 1 128 0.946 0.833 1.055 >> 128.992 146.623 115.730 >> > 1 256 0.969 0.819 1.065 >> 251.986 297.954 229.235 >> > 1 512 0.953 0.867 1.024 >> 512.188 562.918 477.005 >> > 1 1024 0.995 0.873 1.139 >> 981.813 1119.078 857.161 >> > 1 2048 1.010 0.850 1.044 >> 1932.970 2297.619 1870.578 >> > 1 4096 1.071 1.172 1.494 >> 3647.221 3334.105 2614.495 >> > 1 8192 1.634 1.721 2.068 >> 4780.515 4539.290 3777.801 >> > 1 16384 2.989 3.325 4.856 >> 5227.361 4698.842 3217.416 >> > 1 32768 5.321 7.480 9.465 >> 5872.833 4177.594 3301.464 >> > 1 65536 13.295 14.356 20.259 >> 4701.146 4353.478 3085.039 >> > 1 131072 23.574 31.835 45.756 >> 5302.534 3926.447 2731.862 >> > 1 262144 44.921 70.181 102.028 >> 5565.275 3562.231 2450.313 >> > Assertion failed in file >> src/mpid/ch3/channels/nemesis/src/ch3_progress.c at line 643: pkt->type >= >> 0 && pkt->type < MPIDI_CH3_PKT_END_ALL >> > internal ABORT - process 1 >> > >> > >> > -- >> > Jeff Hammond >> > jeff.science(a)gmail.com >> > http://jeffhammond.github.io/ >> > _______________________________________________ >> > To manage subscription options or unsubscribe: >> > https://lists.mpich.org/mailman/listinfo/devel >> >> _______________________________________________ >> To manage subscription options or unsubscribe: >> https://lists.mpich.org/mailman/listinfo/devel >> > > > > -- > Jeff Hammond > jeff.science(a)gmail.com > http://jeffhammond.github.io/ > -- Jeff Hammond jeff.science(a)gmail.com http://jeffhammond.github.io/

1 0

Re: [mpich-devel] Assertion failed in file src/mpid/ch3/channels/nemesis/src/ch3_progress.c at line 643: pkt->type >= 0 && pkt->type < MPIDI_CH3_PKT_END_ALL internal ABORT - process 1
by Jeff Hammond 29 Dec '15

29 Dec '15

I do not, but I guess Marcelo Soares Souza does ( http://lists.mpich.org/pipermail/discuss/2015-June/004045.html) Travis will log the output of mpichversion the next time it runs... Jeff On Tue, Dec 29, 2015 at 12:17 PM, Balaji, Pavan <balaji(a)anl.gov> wrote: > > Do you have more details on how mpich was built? > > -- Pavan > > > On Dec 29, 2015, at 2:05 PM, Jeff Hammond <jeff.science(a)gmail.com> > wrote: > > > > I am using mpich_3.2b3-1ubuntu_amd64.deb within Travis CI for ARMCI-MPI > testing. I never saw this problem in Jenkins, so I can't tell if the > problem is specific to the deb build, to Travis, or something else. > > > > Travis page has all the details so I will not repeat them here. > > > > Thanks, > > > > Jeff > > > > From https://travis-ci.org/jeffhammond/armci-mpi/jobs/99316268: > > > > Processing triggers for libc-bin ... > > ldconfig deferred processing now taking place > > +wget -q > http://www.cebacad.net/files/mpich/ubuntu/mpich-3.2b3/mpich_3.2b3-1ubuntu_a… > > +sudo dpkg -i ./mpich_3.2b3-1ubuntu_amd64.deb > > Selecting previously unselected package mpich. > > (Reading database ... 124680 files and directories currently installed.) > > Unpacking mpich (from .../mpich_3.2b3-1ubuntu_amd64.deb) ... > > Setting up mpich (3.2b3-1ubuntu) ... > > Processing triggers for man-db ... > > > > > > FAIL: benchmarks/contiguous-bench (exit: 1) > > =========================================== > > ARMCI-MPI initialized with 2 processes, MPI v3.1 > > PROGRESS_THREAD = DISABLED > > ALLOC_SHM used = TRUE > > WINDOW type used = CREATE > > STRIDED_METHOD = DIRECT > > IOV_METHOD = DIRECT > > IOV_CHECKS = FALSE > > SHR_BUF_METHOD = COPY > > NONCOLLECTIVE_GROUPS = FALSE > > CACHE_RANK_TRANSLATION = TRUE > > DEBUG_ALLOC = FALSE > > Starting one-sided contiguous performance test with 2 processes > > Trg. Rank Xfer Size Get (usec) Put (usec) Acc (usec) Get > (MiB/s) Put (MiB/s) Acc (MiB/s) > > 1 8 0.869 0.707 0.935 > 8.783 10.795 8.158 > > 1 16 0.926 0.768 0.914 > 16.482 19.859 16.698 > > 1 32 0.945 0.856 0.974 > 32.284 35.671 31.336 > > 1 64 0.938 0.781 0.956 > 65.097 78.199 63.852 > > 1 128 0.946 0.833 1.055 > 128.992 146.623 115.730 > > 1 256 0.969 0.819 1.065 > 251.986 297.954 229.235 > > 1 512 0.953 0.867 1.024 > 512.188 562.918 477.005 > > 1 1024 0.995 0.873 1.139 > 981.813 1119.078 857.161 > > 1 2048 1.010 0.850 1.044 > 1932.970 2297.619 1870.578 > > 1 4096 1.071 1.172 1.494 > 3647.221 3334.105 2614.495 > > 1 8192 1.634 1.721 2.068 > 4780.515 4539.290 3777.801 > > 1 16384 2.989 3.325 4.856 > 5227.361 4698.842 3217.416 > > 1 32768 5.321 7.480 9.465 > 5872.833 4177.594 3301.464 > > 1 65536 13.295 14.356 20.259 > 4701.146 4353.478 3085.039 > > 1 131072 23.574 31.835 45.756 > 5302.534 3926.447 2731.862 > > 1 262144 44.921 70.181 102.028 > 5565.275 3562.231 2450.313 > > Assertion failed in file > src/mpid/ch3/channels/nemesis/src/ch3_progress.c at line 643: pkt->type >= > 0 && pkt->type < MPIDI_CH3_PKT_END_ALL > > internal ABORT - process 1 > > > > > > -- > > Jeff Hammond > > jeff.science(a)gmail.com > > http://jeffhammond.github.io/ > > _______________________________________________ > > To manage subscription options or unsubscribe: > > https://lists.mpich.org/mailman/listinfo/devel > > _______________________________________________ > To manage subscription options or unsubscribe: > https://lists.mpich.org/mailman/listinfo/devel > -- Jeff Hammond jeff.science(a)gmail.com http://jeffhammond.github.io/

1 0

Re: [mpich-devel] Assertion failed in file src/mpid/ch3/channels/nemesis/src/ch3_progress.c at line 643: pkt->type >= 0 && pkt->type < MPIDI_CH3_PKT_END_ALL internal ABORT - process 1
by Balaji, Pavan 29 Dec '15

29 Dec '15

Do you have more details on how mpich was built? -- Pavan > On Dec 29, 2015, at 2:05 PM, Jeff Hammond <jeff.science(a)gmail.com> wrote: > > I am using mpich_3.2b3-1ubuntu_amd64.deb within Travis CI for ARMCI-MPI testing. I never saw this problem in Jenkins, so I can't tell if the problem is specific to the deb build, to Travis, or something else. > > Travis page has all the details so I will not repeat them here. > > Thanks, > > Jeff > > From https://travis-ci.org/jeffhammond/armci-mpi/jobs/99316268: > > Processing triggers for libc-bin ... > ldconfig deferred processing now taking place > +wget -q http://www.cebacad.net/files/mpich/ubuntu/mpich-3.2b3/mpich_3.2b3-1ubuntu_a… > +sudo dpkg -i ./mpich_3.2b3-1ubuntu_amd64.deb > Selecting previously unselected package mpich. > (Reading database ... 124680 files and directories currently installed.) > Unpacking mpich (from .../mpich_3.2b3-1ubuntu_amd64.deb) ... > Setting up mpich (3.2b3-1ubuntu) ... > Processing triggers for man-db ... > > > FAIL: benchmarks/contiguous-bench (exit: 1) > =========================================== > ARMCI-MPI initialized with 2 processes, MPI v3.1 > PROGRESS_THREAD = DISABLED > ALLOC_SHM used = TRUE > WINDOW type used = CREATE > STRIDED_METHOD = DIRECT > IOV_METHOD = DIRECT > IOV_CHECKS = FALSE > SHR_BUF_METHOD = COPY > NONCOLLECTIVE_GROUPS = FALSE > CACHE_RANK_TRANSLATION = TRUE > DEBUG_ALLOC = FALSE > Starting one-sided contiguous performance test with 2 processes > Trg. Rank Xfer Size Get (usec) Put (usec) Acc (usec) Get (MiB/s) Put (MiB/s) Acc (MiB/s) > 1 8 0.869 0.707 0.935 8.783 10.795 8.158 > 1 16 0.926 0.768 0.914 16.482 19.859 16.698 > 1 32 0.945 0.856 0.974 32.284 35.671 31.336 > 1 64 0.938 0.781 0.956 65.097 78.199 63.852 > 1 128 0.946 0.833 1.055 128.992 146.623 115.730 > 1 256 0.969 0.819 1.065 251.986 297.954 229.235 > 1 512 0.953 0.867 1.024 512.188 562.918 477.005 > 1 1024 0.995 0.873 1.139 981.813 1119.078 857.161 > 1 2048 1.010 0.850 1.044 1932.970 2297.619 1870.578 > 1 4096 1.071 1.172 1.494 3647.221 3334.105 2614.495 > 1 8192 1.634 1.721 2.068 4780.515 4539.290 3777.801 > 1 16384 2.989 3.325 4.856 5227.361 4698.842 3217.416 > 1 32768 5.321 7.480 9.465 5872.833 4177.594 3301.464 > 1 65536 13.295 14.356 20.259 4701.146 4353.478 3085.039 > 1 131072 23.574 31.835 45.756 5302.534 3926.447 2731.862 > 1 262144 44.921 70.181 102.028 5565.275 3562.231 2450.313 > Assertion failed in file src/mpid/ch3/channels/nemesis/src/ch3_progress.c at line 643: pkt->type >= 0 && pkt->type < MPIDI_CH3_PKT_END_ALL > internal ABORT - process 1 > > > -- > Jeff Hammond > jeff.science(a)gmail.com > http://jeffhammond.github.io/ > _______________________________________________ > To manage subscription options or unsubscribe: > https://lists.mpich.org/mailman/listinfo/devel

1 0

Assertion failed in file src/mpid/ch3/channels/nemesis/src/ch3_progress.c at line 643: pkt->type >= 0 && pkt->type < MPIDI_CH3_PKT_END_ALL internal ABORT - process 1
by Jeff Hammond 29 Dec '15

29 Dec '15

I am using mpich_3.2b3-1ubuntu_amd64.deb within Travis CI for ARMCI-MPI testing. I never saw this problem in Jenkins, so I can't tell if the problem is specific to the deb build, to Travis, or something else. Travis page has all the details so I will not repeat them here. Thanks, Jeff >From https://travis-ci.org/jeffhammond/armci-mpi/jobs/99316268: Processing triggers for libc-bin ... ldconfig deferred processing now taking place +wget -q http://www.cebacad.net/files/mpich/ubuntu/mpich-3.2b3/mpich_3.2b3-1ubuntu_a… +sudo dpkg -i ./mpich_3.2b3-1ubuntu_amd64.deb Selecting previously unselected package mpich. (Reading database ... 124680 files and directories currently installed.) Unpacking mpich (from .../mpich_3.2b3-1ubuntu_amd64.deb) ... Setting up mpich (3.2b3-1ubuntu) ... Processing triggers for man-db ... FAIL: benchmarks/contiguous-bench (exit: 1) =========================================== ARMCI-MPI initialized with 2 processes, MPI v3.1 PROGRESS_THREAD = DISABLED ALLOC_SHM used = TRUE WINDOW type used = CREATE STRIDED_METHOD = DIRECT IOV_METHOD = DIRECT IOV_CHECKS = FALSE SHR_BUF_METHOD = COPY NONCOLLECTIVE_GROUPS = FALSE CACHE_RANK_TRANSLATION = TRUE DEBUG_ALLOC = FALSE Starting one-sided contiguous performance test with 2 processes Trg. Rank Xfer Size Get (usec) Put (usec) Acc (usec) Get (MiB/s) Put (MiB/s) Acc (MiB/s) 1 8 0.869 0.707 0.935 8.783 10.795 8.158 1 16 0.926 0.768 0.914 16.482 19.859 16.698 1 32 0.945 0.856 0.974 32.284 35.671 31.336 1 64 0.938 0.781 0.956 65.097 78.199 63.852 1 128 0.946 0.833 1.055 128.992 146.623 115.730 1 256 0.969 0.819 1.065 251.986 297.954 229.235 1 512 0.953 0.867 1.024 512.188 562.918 477.005 1 1024 0.995 0.873 1.139 981.813 1119.078 857.161 1 2048 1.010 0.850 1.044 1932.970 2297.619 1870.578 1 4096 1.071 1.172 1.494 3647.221 3334.105 2614.495 1 8192 1.634 1.721 2.068 4780.515 4539.290 3777.801 1 16384 2.989 3.325 4.856 5227.361 4698.842 3217.416 1 32768 5.321 7.480 9.465 5872.833 4177.594 3301.464 1 65536 13.295 14.356 20.259 4701.146 4353.478 3085.039 1 131072 23.574 31.835 45.756 5302.534 3926.447 2731.862 1 262144 44.921 70.181 102.028 5565.275 3562.231 2450.313 Assertion failed in file src/mpid/ch3/channels/nemesis/src/ch3_progress.c at line 643: pkt->type >= 0 && pkt->type < MPIDI_CH3_PKT_END_ALL internal ABORT - process 1 -- Jeff Hammond jeff.science(a)gmail.com http://jeffhammond.github.io/

1 0

Re: [mpich-devel] MPI_Comm_split_type extensions
by Jeff Hammond 19 Nov '15

19 Nov '15

On Thu, Nov 19, 2015 at 1:14 PM, Rob Latham <robl(a)mcs.anl.gov> wrote: > > > On 11/19/2015 03:01 PM, Jeff Hammond wrote: > >> That is so cool! Do you have psychic powers or did you actually go from >> email to finished code in 65 minutes? >> > > Pavan maybe has the psychic powers? he asked me work on this a couple > months ago. > > That makes sense. Pavan did a Vulcan mind meld on me before I left Argonne :-) > but the two implementations have made interesting design choice: OpenMPI > went with the "add a lot of flags" approach. Pavan suggested we take an > approach that introduces only one new flag (MPIX_COMM_SPLIT_NEIGHBORHOOD) > but then uses an info argument to indicate *how* that neighborhood split > should occur. > > I strongly favor the info approach, since it is much easier to toggle features on and off as appropriate without thrashing the header. If MPICH were to support all the HWLOC options like OpenMPI, having it inside of an info key would be future-proof w.r.t. HWLOC feature changes, unlike an explicit list of types. It also makes it trivial to match the exact syntax of HWLOC commands since info keys are composed of strings. > As far as an implementation is concerned, both approaches are equivalent. > > Indeed. > As an end-user, Jeff, do you have an opinion (hah! oh, I kill me... I > thought I could type that with a straight face, I really did...) on which > approach you prefer? > > I have no opinions, just facts that others have yet to accept as truth ;-) Jeff > ==rob > > Jeff >> >> On Thu, Nov 19, 2015 at 12:39 PM, Rob Latham <robl(a)mcs.anl.gov >> <mailto:[email protected]>> wrote: >> >> >> >> On 11/19/2015 01:34 PM, Jeff Hammond wrote: >> >> https://github.com/open-mpi/ompi/pull/320 implements features >> related to >> what was proposed in >> https://svn.mpi-forum.org/trac/mpi-forum-web/ticket/297 and >> https://svn.mpi-forum.org/trac/mpi-forum-web/ticket/372. >> >> It would be nice if MPICH supported such things as well. Since >> MPICH >> also uses HWLOC, I imagine it is not too difficult. >> >> I will file a ticket about this in the absence of objections. >> >> >> >> BLAMMO! >> >> >> http://git.mpich.org/mpich-dev.git/shortlog/refs/heads/comsplit_commonfs >> >> ==rob >> >> >> Thanks, >> >> Jeff >> >> -- >> Jeff Hammond >> jeff.science(a)gmail.com <mailto:[email protected]> >> <mailto:[email protected] <mailto:[email protected]>> >> >> http://jeffhammond.github.io/ >> >> >> _______________________________________________ >> To manage subscription options or unsubscribe: >> https://lists.mpich.org/mailman/listinfo/devel >> >> >> -- >> Rob Latham >> Mathematics and Computer Science Division >> Argonne National Lab, IL USA >> _______________________________________________ >> To manage subscription options or unsubscribe: >> https://lists.mpich.org/mailman/listinfo/devel >> >> >> >> >> -- >> Jeff Hammond >> jeff.science(a)gmail.com <mailto:[email protected]> >> http://jeffhammond.github.io/ >> >> >> _______________________________________________ >> To manage subscription options or unsubscribe: >> https://lists.mpich.org/mailman/listinfo/devel >> >> > -- > Rob Latham > Mathematics and Computer Science Division > Argonne National Lab, IL USA > _______________________________________________ > To manage subscription options or unsubscribe: > https://lists.mpich.org/mailman/listinfo/devel > -- Jeff Hammond jeff.science(a)gmail.com http://jeffhammond.github.io/

1 0

Re: [mpich-devel] MPI_Comm_split_type extensions
by Rob Latham 19 Nov '15

19 Nov '15

On 11/19/2015 03:01 PM, Jeff Hammond wrote: > That is so cool! Do you have psychic powers or did you actually go from > email to finished code in 65 minutes? Pavan maybe has the psychic powers? he asked me work on this a couple months ago. but the two implementations have made interesting design choice: OpenMPI went with the "add a lot of flags" approach. Pavan suggested we take an approach that introduces only one new flag (MPIX_COMM_SPLIT_NEIGHBORHOOD) but then uses an info argument to indicate *how* that neighborhood split should occur. As far as an implementation is concerned, both approaches are equivalent. As an end-user, Jeff, do you have an opinion (hah! oh, I kill me... I thought I could type that with a straight face, I really did...) on which approach you prefer? ==rob > Jeff > > On Thu, Nov 19, 2015 at 12:39 PM, Rob Latham <robl(a)mcs.anl.gov > <mailto:[email protected]>> wrote: > > > > On 11/19/2015 01:34 PM, Jeff Hammond wrote: > > https://github.com/open-mpi/ompi/pull/320 implements features > related to > what was proposed in > https://svn.mpi-forum.org/trac/mpi-forum-web/ticket/297 and > https://svn.mpi-forum.org/trac/mpi-forum-web/ticket/372. > > It would be nice if MPICH supported such things as well. Since > MPICH > also uses HWLOC, I imagine it is not too difficult. > > I will file a ticket about this in the absence of objections. > > > > BLAMMO! > > http://git.mpich.org/mpich-dev.git/shortlog/refs/heads/comsplit_commonfs > > ==rob > > > Thanks, > > Jeff > > -- > Jeff Hammond > jeff.science(a)gmail.com <mailto:[email protected]> > <mailto:[email protected] <mailto:[email protected]>> > http://jeffhammond.github.io/ > > > _______________________________________________ > To manage subscription options or unsubscribe: > https://lists.mpich.org/mailman/listinfo/devel > > > -- > Rob Latham > Mathematics and Computer Science Division > Argonne National Lab, IL USA > _______________________________________________ > To manage subscription options or unsubscribe: > https://lists.mpich.org/mailman/listinfo/devel > > > > > -- > Jeff Hammond > jeff.science(a)gmail.com <mailto:[email protected]> > http://jeffhammond.github.io/ > > > _______________________________________________ > To manage subscription options or unsubscribe: > https://lists.mpich.org/mailman/listinfo/devel > -- Rob Latham Mathematics and Computer Science Division Argonne National Lab, IL USA

1 0

Re: [mpich-devel] MPI_Comm_split_type extensions
by Jeff Hammond 19 Nov '15

19 Nov '15

That is so cool! Do you have psychic powers or did you actually go from email to finished code in 65 minutes? Jeff On Thu, Nov 19, 2015 at 12:39 PM, Rob Latham <robl(a)mcs.anl.gov> wrote: > > > On 11/19/2015 01:34 PM, Jeff Hammond wrote: > >> https://github.com/open-mpi/ompi/pull/320 implements features related to >> what was proposed in >> https://svn.mpi-forum.org/trac/mpi-forum-web/ticket/297 and >> https://svn.mpi-forum.org/trac/mpi-forum-web/ticket/372. >> >> It would be nice if MPICH supported such things as well. Since MPICH >> also uses HWLOC, I imagine it is not too difficult. >> >> I will file a ticket about this in the absence of objections. >> > > > BLAMMO! > > http://git.mpich.org/mpich-dev.git/shortlog/refs/heads/comsplit_commonfs > > ==rob > > >> Thanks, >> >> Jeff >> >> -- >> Jeff Hammond >> jeff.science(a)gmail.com <mailto:[email protected]> >> http://jeffhammond.github.io/ >> >> >> _______________________________________________ >> To manage subscription options or unsubscribe: >> https://lists.mpich.org/mailman/listinfo/devel >> >> > -- > Rob Latham > Mathematics and Computer Science Division > Argonne National Lab, IL USA > _______________________________________________ > To manage subscription options or unsubscribe: > https://lists.mpich.org/mailman/listinfo/devel > -- Jeff Hammond jeff.science(a)gmail.com http://jeffhammond.github.io/

1 0

Re: [mpich-devel] MPI_Comm_split_type extensions
by Rob Latham 19 Nov '15

19 Nov '15

On 11/19/2015 01:34 PM, Jeff Hammond wrote: > https://github.com/open-mpi/ompi/pull/320 implements features related to > what was proposed in > https://svn.mpi-forum.org/trac/mpi-forum-web/ticket/297 and > https://svn.mpi-forum.org/trac/mpi-forum-web/ticket/372. > > It would be nice if MPICH supported such things as well. Since MPICH > also uses HWLOC, I imagine it is not too difficult. > > I will file a ticket about this in the absence of objections. BLAMMO! http://git.mpich.org/mpich-dev.git/shortlog/refs/heads/comsplit_commonfs ==rob > > Thanks, > > Jeff > > -- > Jeff Hammond > jeff.science(a)gmail.com <mailto:[email protected]> > http://jeffhammond.github.io/ > > > _______________________________________________ > To manage subscription options or unsubscribe: > https://lists.mpich.org/mailman/listinfo/devel > -- Rob Latham Mathematics and Computer Science Division Argonne National Lab, IL USA

1 0

MPI_Comm_split_type extensions
by Jeff Hammond 19 Nov '15

19 Nov '15

https://github.com/open-mpi/ompi/pull/320 implements features related to what was proposed in https://svn.mpi-forum.org/trac/mpi-forum-web/ticket/297 and https://svn.mpi-forum.org/trac/mpi-forum-web/ticket/372. It would be nice if MPICH supported such things as well. Since MPICH also uses HWLOC, I imagine it is not too difficult. I will file a ticket about this in the absence of objections. Thanks, Jeff -- Jeff Hammond jeff.science(a)gmail.com http://jeffhammond.github.io/

1 0

Re: [mpich-devel] HYDRA_TOPO_DEBUG=1 needs a line break
by Kenneth Raffenetti 18 Nov '15

18 Nov '15

Just pushed a commit to master that should support many processing elements. Thanks for the report. Ken On 11/18/2015 10:24 AM, Kenneth Raffenetti wrote: > The newline is added after all the hwthreads are added to the string - > line 469. You might be overrunning the max string length, as that is not > well handled. Can you try doubling DBG_STR_LENGTH at the top of the file? > > Ken > > On 11/18/2015 09:09 AM, Jeff Hammond wrote: >> I followed the example on >> https://wiki.mpich.org/mpich/index.php/Using_the_Hydra_Process_Manager#Exam… >> >> but got a huge mess rather than the output shown on the Wiki. >> >> Is the lack of line break here intentional? I think it would be more >> useful if it did. >> >> jhammond@nid00218:~/> HYDRA_TOPO_DEBUG=1 mpiexec -rmk slurm -n 8 >> -bind-to core true >> >> process 0 binding: 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 >> 0 0 0 0 0 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0process 2 >> binding: 0 0 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 >> 0 0 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0process 3 binding: 0 0 0 1 >> 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 0 0 0 0 >> 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0process 5 binding: 0 0 0 0 0 1 0 0 0 0 0 0 >> 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 0 0 0 0 0 0 0 0 0 0 >> 0 0 0 0 0 0 0process 6 binding: 0 0 0 0 0 0 1 0 0 0 0 0 0 0 0 0 0 0 0 0 >> 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 >> 0process 1 binding: 0 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 >> 0 0 0 0 0 0 0 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0process 4 >> binding: 0 0 0 0 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 >> 0 0 0 0 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0process 7 binding: 0 0 0 0 >> 0 0 0 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 >> 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0jhammond@nid00218:~/> >> >> The offending line appears to be 463 in >> src/pm/hydra/tools/topo/hwloc/topo_hwloc.c, but I don't know if the "\n" >> goes there or in the definition of HYDU_snprintf. >> >> Thanks, >> >> Jeff >> >> -- >> Jeff Hammond >> jeff.science(a)gmail.com <mailto:[email protected]> >> http://jeffhammond.github.io/ >> >> >> _______________________________________________ >> To manage subscription options or unsubscribe: >> https://lists.mpich.org/mailman/listinfo/devel >> > _______________________________________________ > To manage subscription options or unsubscribe: > https://lists.mpich.org/mailman/listinfo/devel

1 0