Re: [OMPI users] Problem installing Dalton with OpenMPI overPelicanHPC
On May 15, 2009, at 1:23 AM, Silviu Groza wrote: I still not solved these errors. I need help in order to install Dalton quantum with OpenMPI. Thank you. ---> Linking sequential dalton.x ... mpif77.openmpi -march=x86-64 -O3 -ffast-math -fexpensive- optimizations -funroll-loops -fno-range-check -fsecond-underscore \ I notice the "-fsecond-underscore" option here; do you know who is insertting this option? If I had to guess, I'd say that that is forcing the Fortran linker to change its native name mangling scheme, and it therefore does not match the Fortran name mangling scheme that Open MPI was created with...? -o /root/Fig/dalton-2.0/bin/dalton.x abacus/dalton.o cc/ crayio.o abacus/linux_mem_allo.o \ abacus/herpar.o eri/eri2par.o amfi/amfi.o amfi/symtra.o gp/ mpi_dummy.o -Labacus -labacus -Lrsp -lrsp -Lsirius -lsirius -labacus -Leri -leri -Ldensfit -ldensfit -Lcc -lcc -Ldft -ldft -Lgp -lgp - Lpdpack -lpdpack -L/usr/lib -llapack -lblas dft/libdft.a(general.o): In function `mpi_sync_data': What happens if you copy/paste this entire "mpif77.openmpi" command line and add "--showme" to the end of it? If you chase down the libmpi.so that is used in that command line and run nm on it, do you see ompi_mpi_comm_world (and friends) listed? I'm afraid that I'm not familiar with Dalton and PelicanHPC -- are all of the support libraries listed above part of Darlton or Pelican HPC? Were they all compiled with Open MPI? More specifically: is it possible that they were compiled with a different MPI implementation? -- Jeff Squyres Cisco Systems
Re: [OMPI users] OpenMPI 1.3.2 with PathScale 3.2
Hah; this is probably at least tangentially related to http://www.open-mpi.org/faq/?category=building#pathscale-broken-with-mpi-c++-api I'll be kind and say that Pathscale has been "unwilling to help on these kinds of issues" with me in the past as well. :-) It's not entirely clear from the text, but I guess that sounds like Pathscale is unsupported on GCC 3.x systems...? Is that what you parse his answer to mean? On May 18, 2009, at 5:09 PM, Joshua Bernstein wrote: Well, I spoke Gautam Chakrabarti at Pathscale. It seems the long and short of it is that using OpenMP with C++ with a GNU3.3 (RHEL4) frontend creates some limitations inside of pathCC. On a RHEL4 system, the compilier activates the proper frontend for GCC 3.3, this is what creates the crash. As suggested I forced the compilier to use the newer frontend with the -gnu4 option and the build completes without an issue. Sad though that they aren't trying to be backwards compatible, or even testing on RHEL4 systems. I imagine there is still large group of people using RHEL4. Perhaps this is an OMPI FAQ entry? The full response from Pathscale appears below: ---SNIP--- It appears you are using the compiler on a relatively old linux distribution which has a default GCC compiler based on version 3.3. Our compiler has a front-end that is activated on such systems, and a different newer improved front-end which is activated on the newer GCC4-based systems. Our compiler is tested on GCC-based systems with versions up to 4.2. I see that you are using OpenMP (using -mp). C++ OpenMP has limitations when being used with the GNU3.3 based front-end, and is only fully supported when on a GNU4 based system. You can invoke the newer front-end by the option -gnu4 on a GNU3 based system. While compiling this particular file may work with -gnu4 on a GNU3 based system, it is generally not safe to use this option for C++ on a GNU3 based system due to incompatibility issues. The ideal fix would be to try your compilation on a GNU4 based linux distribution. ---END SNIP--- -Joshua Bernstein Software Engineer Penguin Computing Jeff Squyres wrote: > FWIW, I'm able to duplicate the error. Looks definitely like a[nother] > pathscale bug to me. > > Perhaps David's suggestions to disable some of the optimizations may > help; otherwise, you can disable that entire chunk of code with the > following: > >--enable-contrib-no-build=vt > > (as Ralph mentioned, this VampirTrace code is an add-on to Open MPI; > it's not part of core OMPI itself) > > > On May 15, 2009, at 9:17 AM, David O. Gunter wrote: > >> Pathscale supports -O3 (at least as of the 3.1 line). Here are some >> suggestions from the 3.2 Users Manual you may also want to try. >> >> -david >> >> >> If there are numerical problems with -O3 -OPT:Ofast, then try either >> of the >> following: >> >> -O3 -OPT:Ofast:ro=1 >> -O3 -OPT:Ofast:div_split=OFF >> >> Note that ’ro’ is short for roundoff. >> >> -Ofast is equivalent to -O3 -ipa -OPT:Ofast -fno-math-errno - ffast-math >> so similar cautions apply to it as to -O3 -OPT:Ofast. >> >> To use interprocedural analysis without the "Ofast-type" optimizations, >> use either of the following: >> -O3 -ipa >> -O2 -ipa >> >> Testing different optimizations can be automated by pathopt2. This >> program >> compiles and runs your program with a variety of compiler options and >> creates a sorted list of the execution times for each run. >> >> -- >> David Gunter >> Los Alamos National Laboratory >> >> > Last I checked when we were building here, I'm not sure Pathscale >> > supports -O3. IIRC, O2 is the max supported value, though it has been >> > awhile since I played with it. >> > >> > Have you checked the man page for it? >> > >> > It could also be something in the VampirTrace code since that is where >> > you are failing. That is a contributed code - not part of OMPI itself >> > - so we would have to check with those developers. >> > >> > >> > On May 14, 2009, at 2:49 PM, Åke Sandgren wrote: >> > >> >> On Thu, 2009-05-14 at 13:35 -0700, Joshua Bernstein wrote: >> >>> Greetings All, >> >>> >> >>> I'm trying to build OpenMPI 1.3.2 with the Pathscale compiler, >> >>> version 3.2. A >> >>> bit of the way through the build the compiler dies with what it >> >>> things is a bad >> >>> optimization. Has anybody else seen this, or know a work around for >> >>> it? I'm >> >>> going to take it up with Pathscale of course, but I thought I'd >> >>> throw it out here: >> >>> >> >>> ---SNIP--- >> >>> /opt/pathscale/bin/pathCC -DHAVE_CONFIG_H -I. -I../.. -I../../ >> >>> extlib/otf/otflib >> >>> -I../../extlib/otf/otflib -I../../vtlib/ -I../../vtlib - >> >>> D_GNU_SOURCE -mp >> >>> -DVT_OMP -O3 -DNDEBUG -finline-functions -pthread -MT vtfilter- >> >>> vt_tracefilter.o >> >>> -MD -MP -MF .deps/vtfilter-vt_tracefilter.Tpo -c -o vtfilter- >> >>> vt_tracefilter.o >> >>> `test -f 'vt_
Re: [OMPI users] MPI processes hang when using OpenMPI 1.3.2 and Gcc-4.4.0
Simone Pellegrini wrote: sorry for the delay but I did some additional experiments to found out whether the problem was openmpi or gcc! The program just hangs... and never terminates! I am running on a SMP machine with 32 cores, actually it is a Sun Fire X4600 X2. (8 quad-core Barcelona AMD chips), the OS is CentOS 5 and the kernel is 2.6.18-92.el5.src-PAPI (patched with PAPI). I use a N of 1024, and if I print out the value of the iterator i, sometimes it stops around 165, other times around 520... and it doesn't make any sense. If I run the program (and it's important to notice I don't recompile it, I just use another mpirun from a different mpi version) the program works fine. I did some experiments during the weekend and if I use openmpi-1.3.2 compiled with gcc433 everything works fine. So I really think the problem is strictly related to the usage of gcc-4.4.0! ...and it doesn't depends from OpenMPI as the program hangs even when I use gcc 1.3.1 compiled with gcc 4.4! I finally got GCC 4.4, but was unable to reproduce the problem. How small can you make np (number of MPI processes) and still see the problem? How reproducible is the problem? When it hangs, can you get stack traces of all the processes? We're trying to hunt down some similar behavior, but I think yours is of a different flavor.
Re: [OMPI users] OpenMPI deadlocks and race conditions ?
François PELLEGRINI wrote: users-requ...@open-mpi.org wrote: Date: Thu, 14 May 2009 17:06:07 -0700 From: Eugene Loh Subject: Re: [OMPI users] OpenMPI deadlocks and race conditions ? To: Open MPI Users Fran?ois PELLEGRINI wrote: I sometimes run into deadlocks in OpenMPI (1.3.3a1r21206), when running my MPI+threaded PT-Scotch software. So, are there multiple threads per process that perform message-passing operations? Yes. I use the MPI_THREAD_MULTIPLE level of MPI. You mentioned that the problem was pretty easy to reproduce. Could you send a simple test case (simple means few lines of code and doesn't take a large system to run)?
Re: [OMPI users] OpenMPI 1.3.2 with PathScale 3.2
Well, I spoke Gautam Chakrabarti at Pathscale. It seems the long and short of it is that using OpenMP with C++ with a GNU3.3 (RHEL4) frontend creates some limitations inside of pathCC. On a RHEL4 system, the compilier activates the proper frontend for GCC 3.3, this is what creates the crash. As suggested I forced the compilier to use the newer frontend with the -gnu4 option and the build completes without an issue. Sad though that they aren't trying to be backwards compatible, or even testing on RHEL4 systems. I imagine there is still large group of people using RHEL4. Perhaps this is an OMPI FAQ entry? The full response from Pathscale appears below: ---SNIP--- It appears you are using the compiler on a relatively old linux distribution which has a default GCC compiler based on version 3.3. Our compiler has a front-end that is activated on such systems, and a different newer improved front-end which is activated on the newer GCC4-based systems. Our compiler is tested on GCC-based systems with versions up to 4.2. I see that you are using OpenMP (using -mp). C++ OpenMP has limitations when being used with the GNU3.3 based front-end, and is only fully supported when on a GNU4 based system. You can invoke the newer front-end by the option -gnu4 on a GNU3 based system. While compiling this particular file may work with -gnu4 on a GNU3 based system, it is generally not safe to use this option for C++ on a GNU3 based system due to incompatibility issues. The ideal fix would be to try your compilation on a GNU4 based linux distribution. ---END SNIP--- -Joshua Bernstein Software Engineer Penguin Computing Jeff Squyres wrote: FWIW, I'm able to duplicate the error. Looks definitely like a[nother] pathscale bug to me. Perhaps David's suggestions to disable some of the optimizations may help; otherwise, you can disable that entire chunk of code with the following: --enable-contrib-no-build=vt (as Ralph mentioned, this VampirTrace code is an add-on to Open MPI; it's not part of core OMPI itself) On May 15, 2009, at 9:17 AM, David O. Gunter wrote: Pathscale supports -O3 (at least as of the 3.1 line). Here are some suggestions from the 3.2 Users Manual you may also want to try. -david If there are numerical problems with -O3 -OPT:Ofast, then try either of the following: -O3 -OPT:Ofast:ro=1 -O3 -OPT:Ofast:div_split=OFF Note that ’ro’ is short for roundoff. -Ofast is equivalent to -O3 -ipa -OPT:Ofast -fno-math-errno -ffast-math so similar cautions apply to it as to -O3 -OPT:Ofast. To use interprocedural analysis without the "Ofast-type" optimizations, use either of the following: -O3 -ipa -O2 -ipa Testing different optimizations can be automated by pathopt2. This program compiles and runs your program with a variety of compiler options and creates a sorted list of the execution times for each run. -- David Gunter Los Alamos National Laboratory > Last I checked when we were building here, I'm not sure Pathscale > supports -O3. IIRC, O2 is the max supported value, though it has been > awhile since I played with it. > > Have you checked the man page for it? > > It could also be something in the VampirTrace code since that is where > you are failing. That is a contributed code - not part of OMPI itself > - so we would have to check with those developers. > > > On May 14, 2009, at 2:49 PM, Åke Sandgren wrote: > >> On Thu, 2009-05-14 at 13:35 -0700, Joshua Bernstein wrote: >>> Greetings All, >>> >>> I'm trying to build OpenMPI 1.3.2 with the Pathscale compiler, >>> version 3.2. A >>> bit of the way through the build the compiler dies with what it >>> things is a bad >>> optimization. Has anybody else seen this, or know a work around for >>> it? I'm >>> going to take it up with Pathscale of course, but I thought I'd >>> throw it out here: >>> >>> ---SNIP--- >>> /opt/pathscale/bin/pathCC -DHAVE_CONFIG_H -I. -I../.. -I../../ >>> extlib/otf/otflib >>> -I../../extlib/otf/otflib -I../../vtlib/ -I../../vtlib - >>> D_GNU_SOURCE -mp >>> -DVT_OMP -O3 -DNDEBUG -finline-functions -pthread -MT vtfilter- >>> vt_tracefilter.o >>> -MD -MP -MF .deps/vtfilter-vt_tracefilter.Tpo -c -o vtfilter- >>> vt_tracefilter.o >>> `test -f 'vt_tracefilter.cc' || echo './'`vt_tracefilter.cc >>> Signal: Segmentation fault in Global Optimization -- Dead Store >>> Elimination phase. >>> Error: Signal Segmentation fault in phase Global Optimization -- >>> Dead Store >>> Elimination -- processing aborted >>> *** Internal stack backtrace: >>> pathCC INTERNAL ERROR: /opt/pathscale/lib/3.2/be died due to signal 4 >> >> Haven't seen it. But I'm only using -O2 when building openmpi. >> Report it quickly, if we're lucky they might get a fix into the 3.3 >> release that is due out very soon. (I just got the beta yesterday) >> >> -- >> Ake Sandgren, HPC2N, Umea University, S-90187 Umea, Sweden >> Internet: a...@hpc2n.umu.se Phone: +46 90 7866134 Fax: +46 90 7866126 >> Mobile: +46 70 7716134 WWW: h
[OMPI users] CP2K mpi hang
Hi all - I have a bizarre OpenMPI hanging problem. I'm running an MPI code called CP2K (related to, but not the same as cpmd). The complications of the software aside, here are the observations: At the base is a serial code that uses system() calls to repeatedly invoke mpirun cp2k.popt. When I run from my NFS mounted home directory, everything appears to be fine. When I run from a scratch directory local to each node, it hangs on the _third_ invokation of CP2K (the 1st and 3rd invokations do computationally expensive stuff, the 2nd uses the code in a different mode which does a rather different and quicker computation). These behaviors are quite repeatable. Run from NFS mounted home dir - no problem. Run from node-local scratch directory - hang. Hang is always in the same place (as far as the output of the code, anyway). The underlying system is Linux with a 2.6.18-128.1.6.el5 kernel (CentOS 5.3) on a dual single core Opteron system with Mellanox Infiniband SDR cards. One note of caution is that I'm running OFED 1.4.1-rc4, because I need 1.4.1 for compatibility with this kernel as far as I can tell. The code is complicated, the input files are big and lead to long computation times, so I don't think I'll be able to make a simple test case. Instead I attached to the hanging processes (all 8 of them) with gdb during the hang. The stack trace is below. Nodes seem to spend most of their time in the btl_openib_component_progress(), and occasionally in mca_pml_ob1_progress(). I.e. not completely stuck, but not making progress. Does anyone have any ideas what could be wrong? Noam P.S. I get a similar hang with MVAPICH, in a nearby but different part of the code (on an MPI_Bcast, specifically), increasing my tendency to believe that it's OFED's fault. But maybe the stack trace will suggest to someone where it might be stuck, and therefore perhaps an mca flag to try? #0 0x2ac2d19d7733 in btl_openib_component_progress () from /share/ apps/mpi/openmpi-1.3.2/intel-11.0.083/lib/openmpi/mca_btl_openib.so #1 0x2ac2cdd4daea in opal_progress () from /share/apps/mpi/ openmpi-1.3.2/intel-11.0.083/lib/libopen-pal.so.0 #2 0x2ac2cd887e55 in ompi_request_default_wait_all () from /share/ apps/mpi/openmpi-1.3.2/intel-11.0.083/lib/libmpi.so.0 #3 0x2ac2d2eb544f in ompi_coll_tuned_allreduce_intra_recursivedoubling () from /share/apps/ mpi/openmpi-1.3.2/intel-11.0.083/lib/openmpi/mca_coll_tuned.so #4 0x2ac2cd89b867 in PMPI_Allreduce () from /share/apps/mpi/ openmpi-1.3.2/intel-11.0.083/lib/libmpi.so.0 #5 0x2ac2cd6429b5 in pmpi_allreduce__ () from /share/apps/mpi/ openmpi-1.3.2/intel-11.0.083/lib/libmpi_f77.so.0 #6 0x0077e7db in message_passing_mp_mp_sum_r1_ () #7 0x00be67dd in sparse_matrix_types_mp_cp_sm_sm_trace_scalar_ () #8 0x0160b68c in qs_initial_guess_mp_calculate_first_density_matrix_ () #9 0x00a7ec05 in qs_scf_mp_scf_env_initial_rho_setup_ () #10 0x00a79fca in qs_scf_mp_init_scf_run_ () #11 0x00a659fd in qs_scf_mp_scf_ () #12 0x008c5713 in qs_energy_mp_qs_energies_ () #13 0x008d469e in qs_force_mp_qs_forces_ () #14 0x005368bb in force_env_methods_mp_force_env_calc_energy_force_ () #15 0x0053620e in force_env_methods_mp_force_env_calc_energy_force_ () #16 0x00742724 in md_run_mp_qs_mol_dyn_ () #17 0x00489c42 in cp2k_runs_mp_cp2k_run_ () #18 0x0048878a in cp2k_runs_mp_run_input_ () #19 0x00487669 in MAIN__ () #20 0x0048667c in main () #0 0x2b4d0b57bf09 in mca_pml_ob1_progress () from /share/apps/mpi/ openmpi-1.3.2/intel-11.0.083/lib/openmpi/mca_pml_ob1.so #1 0x2b4d08538aea in opal_progress () from /share/apps/mpi/ openmpi-1.3.2/intel-11.0.083/lib/libopen-pal.so.0 #2 0x2b4d08072e55 in ompi_request_default_wait_all () from /share/ apps/mpi/openmpi-1.3.2/intel-11.0.083/lib/libmpi.so.0 #3 0x2b4d0d6a044f in ompi_coll_tuned_allreduce_intra_recursivedoubling () from /share/apps/ mpi/openmpi-1.3.2/intel-11.0.083/lib/openmpi/mca_coll_tuned.so #4 0x2b4d08086867 in PMPI_Allreduce () from /share/apps/mpi/ openmpi-1.3.2/intel-11.0.083/lib/libmpi.so.0 #5 0x2b4d07e2d9b5 in pmpi_allreduce__ () from /share/apps/mpi/ openmpi-1.3.2/intel-11.0.083/lib/libmpi_f77.so.0 #6 0x0077e7db in message_passing_mp_mp_sum_r1_ () #7 0x00be67dd in sparse_matrix_types_mp_cp_sm_sm_trace_scalar_ () #8 0x0160b68c in qs_initial_guess_mp_calculate_first_density_matrix_ () #9 0x00a7ec05 in qs_scf_mp_scf_env_initial_rho_setup_ () #10 0x00a79fca in qs_scf_mp_init_scf_run_ () #11 0x00a659fd in qs_scf_mp_scf_ () #12 0x008c5713 in qs_energy_mp_qs_energies_ () #13 0x008d469e in qs_force_mp_qs_forces_ () #14 0x005368
Re: [OMPI users] scaling problem with openmpi
Hi Pavel This is not my league, but here are some CPMD helpful links (code, benchmarks): http://www.cpmd.org/ http://www.cpmd.org/cpmd_thecode.html http://www.theochem.ruhr-uni-bochum.de/~axel.kohlmeyer/cpmd-bench.html IHIH Gus Correa Noam Bernstein wrote: On May 18, 2009, at 12:50 PM, Pavel Shamis (Pasha) wrote: Roman, Can you please share with us Mvapich numbers that you get . Also what is mvapich version that you use. Default mvapich and openmpi IB tuning is very similar, so it is strange to see so big difference. Do you know what kind of collectives operation is used in this specific application. This code does a bunch of parallel things in various different places (mostly dense matrix math, and some FFT stuff that may or may not be parallelized). In the standard output there's a summary of the time taken by various MPI routines. Perhaps Roman can send them? The code also uses ScaLAPACK, but I'm not sure how CP2K labels the timing for those routines in the output. Noam ___ users mailing list us...@open-mpi.org http://www.open-mpi.org/mailman/listinfo.cgi/users
Re: [OMPI users] could oversubscription clobber an executable?
There is another option here -- Fortran compilers can aggressively move code around, particularly when it doesn't know about MPI inter- function dependencies. This usually only happens with non-blocking MPI communication functions, though. Are you using those, perchance? On May 18, 2009, at 11:51 AM, Iain Bason wrote: On May 14, 2009, at 3:20 PM, Valmor de Almeida wrote: > I guess another way to ask is: is it guaranteed that A and B are > contiguous? Yes. > and the MPI communication correctly sends the data? I'm not sure what you're asking, but the code looks as though it ought to work. Iain ___ users mailing list us...@open-mpi.org http://www.open-mpi.org/mailman/listinfo.cgi/users -- Jeff Squyres Cisco Systems
Re: [OMPI users] scaling problem with openmpi
On May 18, 2009, at 12:50 PM, Pavel Shamis (Pasha) wrote: Roman, Can you please share with us Mvapich numbers that you get . Also what is mvapich version that you use. Default mvapich and openmpi IB tuning is very similar, so it is strange to see so big difference. Do you know what kind of collectives operation is used in this specific application. This code does a bunch of parallel things in various different places (mostly dense matrix math, and some FFT stuff that may or may not be parallelized). In the standard output there's a summary of the time taken by various MPI routines. Perhaps Roman can send them? The code also uses ScaLAPACK, but I'm not sure how CP2K labels the timing for those routines in the output. Noam
Re: [OMPI users] scaling problem with openmpi
Roman, Can you please share with us Mvapich numbers that you get . Also what is mvapich version that you use. Default mvapich and openmpi IB tuning is very similar, so it is strange to see so big difference. Do you know what kind of collectives operation is used in this specific application. Pasha. Roman Martonak wrote: I've been using --mca mpi_paffinity_alone 1 in all simulations. Concerning "-mca mpi_leave_pinned 1", I tried it with openmpi 1.2.X versions and it makes no difference. Best regards Roman On Mon, May 18, 2009 at 4:57 PM, Pavel Shamis (Pasha) wrote: 1) I was told to add "-mca mpi_leave_pinned 0" to avoid problems with Infinband. This was with OpenMPI 1.3.1. Not Actually for 1.2.X version I will recommend you to enable leave pinned "-mca mpi_leave_pinned 1" sure if the problems were fixed on 1.3.2, but I am hanging on to that setting just in case. We had data corruption issue in 1.3.1 but it was resolved in 1.3.2. In 1.3.2 version leave_pinned is enabled by default. If I remember correct mvapich enables affinity mode by default, so I can recommend you to try to enable it too: "--mca mpi_paffinity_alone 1". For more details please check FAQ - http://www.open-mpi.org/faq/?category=tuning#using-paffinity Thanks, Pasha. ___ users mailing list us...@open-mpi.org http://www.open-mpi.org/mailman/listinfo.cgi/users
Re: [OMPI users] scaling problem with openmpi
Hi Roman Note that in 1.3.0 and 1.3.1 the default ("-mca mpi_leave_pinned 1") had a glitch. In my case it appeared as a memory leak. See this: http://www.open-mpi.org/community/lists/users/2009/05/9173.php http://www.open-mpi.org/community/lists/announce/2009/03/0029.php One workaround is to revert to "-mca mpi_leave_pinned 0" (which is what I suggested to you) when using 1.3.0 or 1.3.1. The solution advocated by OpenMPI is to upgrade to 1.3.2. You reported you used "1.3", besides 1.2.6 and 1.2.8. If this means that you are using 1.3.0 or 1.3.1, you may want to try the workaround or the upgrade, regardless of any scaling performance expectations. Gus Correa - Gustavo Correa Lamont-Doherty Earth Observatory - Columbia University Palisades, NY, 10964-8000 - USA - Roman Martonak wrote: I've been using --mca mpi_paffinity_alone 1 in all simulations. Concerning "-mca mpi_leave_pinned 1", I tried it with openmpi 1.2.X versions and it makes no difference. Best regards Roman On Mon, May 18, 2009 at 4:57 PM, Pavel Shamis (Pasha) wrote: 1) I was told to add "-mca mpi_leave_pinned 0" to avoid problems with Infinband. This was with OpenMPI 1.3.1. Not Actually for 1.2.X version I will recommend you to enable leave pinned "-mca mpi_leave_pinned 1" sure if the problems were fixed on 1.3.2, but I am hanging on to that setting just in case. We had data corruption issue in 1.3.1 but it was resolved in 1.3.2. In 1.3.2 version leave_pinned is enabled by default. If I remember correct mvapich enables affinity mode by default, so I can recommend you to try to enable it too: "--mca mpi_paffinity_alone 1". For more details please check FAQ - http://www.open-mpi.org/faq/?category=tuning#using-paffinity Thanks, Pasha. ___ users mailing list us...@open-mpi.org http://www.open-mpi.org/mailman/listinfo.cgi/users ___ users mailing list us...@open-mpi.org http://www.open-mpi.org/mailman/listinfo.cgi/users
Re: [OMPI users] scaling problem with openmpi
I've been using --mca mpi_paffinity_alone 1 in all simulations. Concerning "-mca mpi_leave_pinned 1", I tried it with openmpi 1.2.X versions and it makes no difference. Best regards Roman On Mon, May 18, 2009 at 4:57 PM, Pavel Shamis (Pasha) wrote: > >> >> 1) I was told to add "-mca mpi_leave_pinned 0" to avoid problems with >> Infinband. This was with OpenMPI 1.3.1. Not > > Actually for 1.2.X version I will recommend you to enable leave pinned "-mca > mpi_leave_pinned 1" >> >> sure if the problems were fixed on 1.3.2, but I am hanging on to that >> setting just in case. > > We had data corruption issue in 1.3.1 but it was resolved in 1.3.2. In 1.3.2 > version leave_pinned is enabled by default. > > If I remember correct mvapich enables affinity mode by default, so I can > recommend you to try to enable it too: > "--mca mpi_paffinity_alone 1". For more details please check FAQ - > http://www.open-mpi.org/faq/?category=tuning#using-paffinity > > Thanks, > Pasha. > ___ > users mailing list > us...@open-mpi.org > http://www.open-mpi.org/mailman/listinfo.cgi/users >
Re: [OMPI users] could oversubscription clobber an executable?
On May 14, 2009, at 3:20 PM, Valmor de Almeida wrote: I guess another way to ask is: is it guaranteed that A and B are contiguous? Yes. and the MPI communication correctly sends the data? I'm not sure what you're asking, but the code looks as though it ought to work. Iain
Re: [OMPI users] Openmpi -MacOSX-mpif90 won't compile
Thanks for that comment. I thought that is what I was doing when I used the full path name /usr/local/openmpi-1.3/bin/mpif90 Is that not true? John Boccio On May 18, 2009, at 11:31 AM, Jeff Squyres wrote: Check first to make sure you're using the mpif90 in /usr/local/ openmpi-1.3/bin -- OS X ships with an Open MPI installation that does not include F90 support. The default OS X Open MPI install may be in your PATH before the Open MPI you just installed in /usr/local. On May 18, 2009, at 10:13 AM, John Boccio wrote: Hi, I need to use mpif90 for some work on a parallel cluster for galaxy- galaxy collision research. I am certainly not an expert in using UNIX to compile big packages like openmpi. I have list below all (I hope) relevant information and included output files(compressed) as an attachment. Thanks for any help, John Boccio boc...@swarthmore.edu Department of Physics Swarthmore College Here is g95 and xcode info. Using openmpi-1.3 Mac OSX Leopard 10.5.7 g95 from www.g95.com g95 -v Using built-in specs. Target: Configured with: ../configure --enable-languages=c Thread model: posix gcc version 4.0.3 (g95 0.92!) Oct 18 2008 xcode311_2517_developerdvd.dmg openmpi-1.3 sudo ./configure --enable-mpi-f77 --enable-mpi-f90 F77="/usr/bin/ g95" FC="/usr/bin/g95" > config.out sudo make clean sudo make clean prefix=/usr/local/openmpi-1.3 sudo make > make.out sudo make install prefix=/usr/local/openmpi-1.3 > make-install.out /usr/local/openmpi-1.3/bin/mpif90 -- Unfortunately, this installation of Open MPI was not compiled with Fortran 90 support. As such, the mpif90 compiler is non-functional. -- files included in attachment ompi-output.tar.gz : config.out config.status config.log Makefile make.out make-install.out ___ users mailing list us...@open-mpi.org http://www.open-mpi.org/mailman/listinfo.cgi/users -- Jeff Squyres Cisco Systems ___ users mailing list us...@open-mpi.org http://www.open-mpi.org/mailman/listinfo.cgi/users
Re: [OMPI users] Openmpi -MacOSX-mpif90 won't compile
Check first to make sure you're using the mpif90 in /usr/local/ openmpi-1.3/bin -- OS X ships with an Open MPI installation that does not include F90 support. The default OS X Open MPI install may be in your PATH before the Open MPI you just installed in /usr/local. On May 18, 2009, at 10:13 AM, John Boccio wrote: Hi, I need to use mpif90 for some work on a parallel cluster for galaxy- galaxy collision research. I am certainly not an expert in using UNIX to compile big packages like openmpi. I have list below all (I hope) relevant information and included output files(compressed) as an attachment. Thanks for any help, John Boccio boc...@swarthmore.edu Department of Physics Swarthmore College Here is g95 and xcode info. Using openmpi-1.3 Mac OSX Leopard 10.5.7 g95 from www.g95.com g95 -v Using built-in specs. Target: Configured with: ../configure --enable-languages=c Thread model: posix gcc version 4.0.3 (g95 0.92!) Oct 18 2008 xcode311_2517_developerdvd.dmg openmpi-1.3 sudo ./configure --enable-mpi-f77 --enable-mpi-f90 F77="/usr/bin/ g95" FC="/usr/bin/g95" > config.out sudo make clean sudo make clean prefix=/usr/local/openmpi-1.3 sudo make > make.out sudo make install prefix=/usr/local/openmpi-1.3 > make-install.out /usr/local/openmpi-1.3/bin/mpif90 -- Unfortunately, this installation of Open MPI was not compiled with Fortran 90 support. As such, the mpif90 compiler is non-functional. -- files included in attachment ompi-output.tar.gz : config.out config.status config.log Makefile make.out make-install.out ___ users mailing list us...@open-mpi.org http://www.open-mpi.org/mailman/listinfo.cgi/users -- Jeff Squyres Cisco Systems
Re: [OMPI users] scaling problem with openmpi
1) I was told to add "-mca mpi_leave_pinned 0" to avoid problems with Infinband. This was with OpenMPI 1.3.1. Not Actually for 1.2.X version I will recommend you to enable leave pinned "-mca mpi_leave_pinned 1" sure if the problems were fixed on 1.3.2, but I am hanging on to that setting just in case. We had data corruption issue in 1.3.1 but it was resolved in 1.3.2. In 1.3.2 version leave_pinned is enabled by default. If I remember correct mvapich enables affinity mode by default, so I can recommend you to try to enable it too: "--mca mpi_paffinity_alone 1". For more details please check FAQ - http://www.open-mpi.org/faq/?category=tuning#using-paffinity Thanks, Pasha.
[OMPI users] Openmpi -MacOSX-mpif90 won't compile
Hi,I need to use mpif90 for some work on a parallel cluster for galaxy-galaxy collision research.I am certainly not an expert in using UNIX to compile big packages like openmpi.I have list below all (I hope) relevant information and included output files(compressed) as an attachment.Thanks for any help,John Boccioboc...@swarthmore.eduDepartment of PhysicsSwarthmore CollegeHere is g95 and xcode info.Using openmpi-1.3Mac OSX Leopard 10.5.7g95 from www.g95.comg95 -vUsing built-in specs.Target: Configured with: ../configure --enable-languages=cThread model: posixgcc version 4.0.3 (g95 0.92!) Oct 18 2008xcode311_2517_developerdvd.dmgopenmpi-1.3sudo ./configure --enable-mpi-f77 --enable-mpi-f90 F77="/usr/bin/g95" FC="/usr/bin/g95" > config.outsudo make cleansudo make clean prefix=/usr/local/openmpi-1.3sudo make > make.outsudo make install prefix=/usr/local/openmpi-1.3 > make-install.out/usr/local/openmpi-1.3/bin/mpif90--Unfortunately, this installation of Open MPI was not compiled withFortran 90 support. As such, the mpif90 compiler is non-functional.--files included in attachment ompi-output.tar.gz :config.outconfig.statusconfig.logMakefilemake.outmake-install.out ompi-output.tar.gz Description: GNU Zip compressed data