[OMPI users] OMPI tested on Solaris 10 for AMD CPUs?
Hi all, has Open MPI been tested on Solaris 10 for AMD processors? I seem to have a problem with MPI_Bcast. First broadcast (2 ints) works fine, second broadcast (char string) hangs. I will check with a simpler example, but this is right at the start of my program, so really looks like a Open MPI issue. The code runs fine on Solaris SPARC, Linux, Windows, Mac OS X... BTW: I tested with version 1.1.2 and 1.2b1 Regards Dirk -- Dr. Dirk J. Evers, Managing Director International Graduate School in Bioinformatics and Genome Research CeBiTec, Bielefeld University, 33594 Bielefeld, Germany Office: V10-108, Tel: +49 (521) 106-3793, Fax: +49 (521) 106-6490 <>
Re: [OMPI users] OMPI tested on Solaris 10 for AMD CPUs?
Dirk, I have tried the simple case you mention and have run successfully with the current v1.2 branch, Solaris 10 on AMD. Maybe the actual code snipet might help reproduce, how many processes and how many nodes are involved? I should probably try the the latest v1.2b download. -DON Dirk J. Evers wrote On 12/14/06 04:28,: Hi all, has Open MPI been tested on Solaris 10 for AMD processors? I seem to have a problem with MPI_Bcast. First broadcast (2 ints) works fine, second broadcast (char string) hangs. I will check with a simpler example, but this is right at the start of my program, so really looks like a Open MPI issue. The code runs fine on Solaris SPARC, Linux, Windows, Mac OS X... BTW: I tested with version 1.1.2 and 1.2b1 Regards Dirk ___ users mailing list us...@open-mpi.org http://www.open-mpi.org/mailman/listinfo.cgi/users
Re: [OMPI users] OMPI tested on Solaris 10 for AMD CPUs?
Hello, On Thursday 14 December 2006 10:28, Dirk J. Evers wrote: > has Open MPI been tested on Solaris 10 for AMD processors? it is currently not tested by the nightly build system for this OS. However, I have been running the trunk version of Open MPI before SC'06 with the mpich- and mpi_test_suite-suites, as well as IMB on Solaris 10 with a Ultra-40 workstation, i.e. dual-socket dual-core Opteron. Hope, this helps? With best regards, Rainer Keller > I seem to have a problem with MPI_Bcast. > First broadcast (2 ints) works fine, second broadcast (char string) > hangs. I will check with a simpler example, but this is right at the > start of my program, so really looks like a Open MPI issue. > The code runs fine on Solaris SPARC, Linux, Windows, Mac OS X... > > BTW: I tested with version 1.1.2 and 1.2b1 > > Regards > Dirk -- Dipl.-Inf. Rainer Keller http://www.hlrs.de/people/keller High Performance Computing Tel: ++49 (0)711-685 6 5858 Center Stuttgart (HLRS) Fax: ++49 (0)711-685 6 5832 POSTAL:Nobelstrasse 19 email: kel...@hlrs.de ACTUAL:Allmandring 30, R.O.030AIM:rusraink 70550 Stuttgart
[OMPI users] pgi compiler configure problems
good day all, i've been trying to build ompi with the 6.2-X version of the pgi compiler set (pgcc 6.2-4 64-bit target on x86-64 Linux). i've tried both 1.1.2 and the current nightly build 1.1.3b2r12766, both fail with this error from configure: *** Fortran 77 compiler checking whether we are using the GNU Fortran 77 compiler... no checking whether /opt/pgi/linux86-64/6.2/bin/pgf77 accepts -g... yes checking if Fortran 77 compiler works... yes checking /opt/pgi/linux86-64/6.2/bin/pgf77 external symbol convention... single underscore checking if Fortran 77 compiler supports LOGICAL... yes checking size of Fortran 77 LOGICAL... 4 checking for C type corresponding to LOGICAL... not found configure: WARNING: *** Did not find corresponding C type configure: error: Cannot continue i will attache the config.log file. i've looked around the maillist and i've seen a few issue similar but no real resolution. the env i used to build is: F90=/opt/pgi/linux86-64/6.2/bin/pgf90 F77=/opt/pgi/linux86-64/6.2/bin/pgf77 LM_LICENSE_FILE=/opt/pgi/license.dat CXX=/opt/pgi/linux86-64/6.2/bin/pgCC FC=/opt/pgi/linux86-64/6.2/bin/pgf90 CC=/opt/pgi/linux86-64/6.2/bin/pgcc F90FLAGS=-fastsse CXXFLAGS=-g -O3 -Minfo=all -fastsse -Mipa=fast -Knoieee -mp FCFLAGS=-fastsse CFLAGS=-g -O3 -Minfo=all -fastsse -Mipa=fast -Knoieee -mp FFLAGS=-fastsse is there something simple i'm messing up? thanks! -- michael This file contains any messages produced by compilers while running configure, to aid debugging if configure makes a mistake. It was created by Open MPI configure 1.1.3b2, which was generated by GNU Autoconf 2.59. Invocation command line was $ ./configure --prefix=/share/apps/opmi/113b2r12766 --with-gnu-ld ## - ## ## Platform. ## ## - ## hostname = bioinfo.lsd.ornl.gov uname -m = x86_64 uname -r = 2.6.9-42.0.2.ELsmp uname -s = Linux uname -v = #1 SMP Wed Aug 23 13:38:27 BST 2006 /usr/bin/uname -p = unknown /bin/uname -X = unknown /bin/arch = x86_64 /usr/bin/arch -k = unknown /usr/convex/getsysinfo = unknown hostinfo = unknown /bin/machine = unknown /usr/bin/oslevel = unknown /bin/universe = unknown PATH: /opt/pgi/linux86-64/6.2/bin PATH: /opt/pgi/linux86/6.2/bin PATH: /usr/kerberos/sbin PATH: /usr/kerberos/bin PATH: /opt/globus/bin PATH: /opt/globus/sbin PATH: /usr/local/sbin PATH: /usr/local/bin PATH: /sbin PATH: /bin PATH: /usr/sbin PATH: /usr/bin PATH: /usr/X11R6/bin PATH: /opt/Bio/ncbi/bin PATH: /opt/Bio/mpiblast/bin/ PATH: /opt/Bio/hmmer/bin PATH: /opt/Bio/Emboss/bin PATH: /opt/Bio/clustalw/bin PATH: /opt/Bio/t_coffee/bin PATH: /opt/Bio/phylip/exe PATH: /opt/Bio/mrbayes PATH: /opt/Bio/fasta PATH: /opt/Bio/glimmer/bin PATH: //opt/Bio/glimmer/scripts PATH: /opt/Bio/gromacs/bin PATH: /opt/ganglia/bin PATH: /opt/ganglia/sbin PATH: /usr/java/jre1.5.0_09/bin PATH: /opt/lam/gnu/bin PATH: /usr/NX/bin PATH: /opt/openmpi/bin/ PATH: /opt/rocks/bin PATH: /opt/rocks/sbin PATH: /root/bin ## --- ## ## Core tests. ## ## --- ## configure:1811: checking for a BSD-compatible install configure:1866: result: /usr/bin/install -c configure:1877: checking whether build environment is sane configure:1920: result: yes configure:1985: checking for gawk configure:2001: found /bin/gawk configure:2011: result: gawk configure:2021: checking whether make sets $(MAKE) configure:2041: result: yes configure:2330: checking Open MPI version configure:2332: result: 1.1.3b2r12766 configure:2334: checking Open MPI Subversion repository version configure:2336: result: r12766 configure:2432: checking Open Run-Time Environment version configure:2434: result: 1.1.3b2r12766 configure:2436: checking Open Run-Time Environment Subversion repository version configure:2438: result: r12766 configure:2534: checking Open Portable Access Layer version configure:2536: result: 1.1.3b2r12766 configure:2538: checking Open Portable Access Layer Subversion repository version configure:2540: result: r12766 configure:2597: builddir: /share/build/openmpi-1.1.3b2r12766 configure:2599: srcdir: /share/build/openmpi-1.1.3b2r12766 configure:2621: checking build system type configure:2639: result: x86_64-unknown-linux-gnu configure:2667: checking host system type configure:2681: result: x86_64-unknown-linux-gnu configure:2835: checking Whether to run code coverage configure:2862: result: no configure:2872: checking whether to debug memory usage configure:2884: result: no configure:2904: checking whether to profile memory usage configure:2916: result: no configure:2936: checking if want developer-level compiler pickyness configure:2948: result: no configure:2963: checking if want developer-level debugging code configure:2975: result: no configure:3000: checking if want Fortran 77 bindings configure:3008: result: yes configure:3022: checking if want Fortran 90 bindings configure:3030: result: yes configure:3039: checking desired Fortran 90 bindings "size" configure:3121: result: small con
Re: [OMPI users] pgi compiler configure problems
Michael, I just did a quick config of 1.1.2 using PGI 6.2-4 and didn't run into this problem. Here's the environment I used (similar to yours): CXX=/usr/projects/hpctools/packages/x86_64/PGI/pgi_6.2-4/ linux86-64/6.2/bin/pgCC F90=/usr/projects/hpctools/packages/x86_64/PGI/pgi_6.2-4/ linux86-64/6.2/bin/pgf90 PGI=/usr/projects/hpctools/packages/x86_64/PGI/pgi_6.2-4 FC=/usr/projects/hpctools/packages/x86_64/PGI/pgi_6.2-4/ linux86-64/6.2/bin/pgf95 F77=/usr/projects/hpctools/packages/x86_64/PGI/pgi_6.2-4/ linux86-64/6.2/bin/pgf77 CC=/usr/projects/hpctools/packages/x86_64/PGI/pgi_6.2-4/ linux86-64/6.2/bin/pgcc F90FLAGS=-fastsse CXXFLAGS=-g -O3 -Minfo=all -fastsse -Mipa=fast -Knoieee -mp FCFLAGS=-fastsse CFLAGS=-g -O3 -Minfo=all -fastsse -Mipa=fast -Knoieee -mp FFLAGS=-fastsse Here's my configure options: /configure --prefix=/usr/projects/hpctools/packages/x86_64/OpenMPI/ openmpi-1.1.2 --with-io_romio_flags=--disable-aio --with-bproc *** Fortran 77 compiler checking whether we are using the GNU Fortran 77 compiler... no checking whether /usr/projects/hpctools/packages/x86_64/PGI/pgi_6.2-4/ linux86-64/6.2/bin/pgf77 accepts -g... yes checking if Fortran 77 compiler works... yes checking /usr/projects/hpctools/packages/x86_64/PGI/pgi_6.2-4/ linux86-64/6.2/bin/pgf77 external symbol convention... sin gle underscore checking if Fortran 77 compiler supports LOGICAL... yes checking size of Fortran 77 LOGICAL... 4 checking for C type corresponding to LOGICAL... int checking alignment of Fortran LOGICAL... 4 ... -david -- David Gunter HPC-4: HPC Environments: Parallel Tools Team Los Alamos National Laboratory On Dec 14, 2006, at 12:01 PM, Michael Galloway wrote: good day all, i've been trying to build ompi with the 6.2-X version of the pgi compiler set (pgcc 6.2-4 64-bit target on x86-64 Linux). i've tried both 1.1.2 and the current nightly build 1.1.3b2r12766, both fail with this error from configure: *** Fortran 77 compiler checking whether we are using the GNU Fortran 77 compiler... no checking whether /opt/pgi/linux86-64/6.2/bin/pgf77 accepts -g... yes checking if Fortran 77 compiler works... yes checking /opt/pgi/linux86-64/6.2/bin/pgf77 external symbol convention... single underscore checking if Fortran 77 compiler supports LOGICAL... yes checking size of Fortran 77 LOGICAL... 4 checking for C type corresponding to LOGICAL... not found configure: WARNING: *** Did not find corresponding C type configure: error: Cannot continue i will attache the config.log file. i've looked around the maillist and i've seen a few issue similar but no real resolution. the env i used to build is: F90=/opt/pgi/linux86-64/6.2/bin/pgf90 F77=/opt/pgi/linux86-64/6.2/bin/pgf77 LM_LICENSE_FILE=/opt/pgi/license.dat CXX=/opt/pgi/linux86-64/6.2/bin/pgCC FC=/opt/pgi/linux86-64/6.2/bin/pgf90 CC=/opt/pgi/linux86-64/6.2/bin/pgcc F90FLAGS=-fastsse CXXFLAGS=-g -O3 -Minfo=all -fastsse -Mipa=fast -Knoieee -mp FCFLAGS=-fastsse CFLAGS=-g -O3 -Minfo=all -fastsse -Mipa=fast -Knoieee -mp FFLAGS=-fastsse is there something simple i'm messing up? thanks! -- michael ___ users mailing list us...@open-mpi.org http://www.open-mpi.org/mailman/listinfo.cgi/users
Re: [OMPI users] pgi compiler configure problems
On Thu, Dec 14, 2006 at 02:01:54PM -0500, Michael Galloway wrote: > good day all, i've been trying to build ompi with the 6.2-X version of the pgi > compiler set (pgcc 6.2-4 64-bit target on x86-64 Linux). i've tried both > 1.1.2 and > the current nightly build 1.1.3b2r12766, both fail with this error from > configure: > > *** Fortran 77 compiler > checking whether we are using the GNU Fortran 77 compiler... no > checking whether /opt/pgi/linux86-64/6.2/bin/pgf77 accepts -g... yes > checking if Fortran 77 compiler works... yes > checking /opt/pgi/linux86-64/6.2/bin/pgf77 external symbol convention... > single underscore > checking if Fortran 77 compiler supports LOGICAL... yes > checking size of Fortran 77 LOGICAL... 4 > checking for C type corresponding to LOGICAL... not found > configure: WARNING: *** Did not find corresponding C type > configure: error: Cannot continue > > i will attache the config.log file. i've looked around the maillist and i've > seen a few issue > similar but no real resolution. the env i used to build is: > > F90=/opt/pgi/linux86-64/6.2/bin/pgf90 > F77=/opt/pgi/linux86-64/6.2/bin/pgf77 > LM_LICENSE_FILE=/opt/pgi/license.dat > CXX=/opt/pgi/linux86-64/6.2/bin/pgCC > FC=/opt/pgi/linux86-64/6.2/bin/pgf90 > CC=/opt/pgi/linux86-64/6.2/bin/pgcc > > F90FLAGS=-fastsse > CXXFLAGS=-g -O3 -Minfo=all -fastsse -Mipa=fast -Knoieee -mp > FCFLAGS=-fastsse > CFLAGS=-g -O3 -Minfo=all -fastsse -Mipa=fast -Knoieee -mp > FFLAGS=-fastsse > > is there something simple i'm messing up? thanks! > > -- michael interesting, if i build with the 32bit pgi components it works. the env: [root@bioinfo openmpi-1.1.2]# env | grep pgi PGI=/opt/pgi PATH=/opt/pgi/linux86/6.2/bin:/opt/pgi/linux86-64/6.2/bin:/usr/kerberos/sbin:/usr/kerberos/bin:/opt/globus/bin:/opt/globus/sbin:/usr/local/sbin:/usr/local/bin:/sbin:/bin:/usr/sbin:/usr/bin:/usr/X11R6/bin:/opt/Bio/ncbi/bin:/opt/Bio/mpiblast/bin/:/opt/Bio/hmmer/bin:/opt/Bio/Emboss/bin:/opt/Bio/clustalw/bin:/opt/Bio/t_coffee/bin:/opt/Bio/phylip/exe:/opt/Bio/mrbayes:/opt/Bio/fasta:/opt/Bio/glimmer/bin://opt/Bio/glimmer/scripts:/opt/Bio/gromacs/bin:/opt/ganglia/bin:/opt/ganglia/sbin:/usr/java/jre1.5.0_09/bin:/opt/lam/gnu/bin:/usr/NX/bin:/opt/openmpi/bin/:/opt/rocks/bin:/opt/rocks/sbin:/root/bin F90=/opt/pgi/linux86/6.2/bin/pgf90 F77=/opt/pgi/linux86/6.2/bin/pgf77 LM_LICENSE_FILE=/opt/pgi/license.dat CXX=/opt/pgi/linux86/6.2/bin/pgCC FC=/opt/pgi/linux86/6.2/bin/pgf95 CC=/opt/pgi/linux86/6.2/bin/pgcc [root@bioinfo openmpi-1.1.2]# env | grep FLAG F90FLAGS=-fastsse CXXFLAGS=-g -O3 -Minfo=all -fastsse -Mipa=fast -Knoieee -mp CFLAGS=-g -O3 -Minfo=all -fastsse -Mipa=fast -Knoieee -mp FFLAGS=-fastsse and config'd with: ./configure --prefix=/share/apps/opmi-112/pgi --with-gnu-ld at least configure completes. if i change back to x86_64 it fails with: *** Fortran 77 compiler checking whether we are using the GNU Fortran 77 compiler... no checking whether /opt/pgi/linux86-64/6.2/bin/pgf77 accepts -g... yes checking if Fortran 77 compiler works... yes checking /opt/pgi/linux86-64/6.2/bin/pgf77 external symbol convention... single underscore checking if Fortran 77 compiler supports LOGICAL... yes checking size of Fortran 77 LOGICAL... 4 checking for C type corresponding to LOGICAL... not found configure: WARNING: *** Did not find corresponding C type configure: error: Cannot continue and the env: [root@bioinfo openmpi-1.1.2]# env | grep pgi PGI=/opt/pgi PATH=/opt/pgi/linux86-64/6.2/bin:/opt/pgi/linux86/6.2/bin:/opt/pgi/linux86-64/6.2/bin:/usr/kerberos/sbin:/usr/kerberos/bin:/opt/globus/bin:/opt/globus/sbin:/usr/local/sbin:/usr/local/bin:/sbin:/bin:/usr/sbin:/usr/bin:/usr/X11R6/bin:/opt/Bio/ncbi/bin:/opt/Bio/mpiblast/bin/:/opt/Bio/hmmer/bin:/opt/Bio/Emboss/bin:/opt/Bio/clustalw/bin:/opt/Bio/t_coffee/bin:/opt/Bio/phylip/exe:/opt/Bio/mrbayes:/opt/Bio/fasta:/opt/Bio/glimmer/bin://opt/Bio/glimmer/scripts:/opt/Bio/gromacs/bin:/opt/ganglia/bin:/opt/ganglia/sbin:/usr/java/jre1.5.0_09/bin:/opt/lam/gnu/bin:/usr/NX/bin:/opt/openmpi/bin/:/opt/rocks/bin:/opt/rocks/sbin:/root/bin F90=/opt/pgi/linux86-64/6.2/bin/pgf90 F77=/opt/pgi/linux86-64/6.2/bin/pgf77 LM_LICENSE_FILE=/opt/pgi/license.dat CXX=/opt/pgi/linux86-64/6.2/bin/pgCC FC=/opt/pgi/linux86-64/6.2/bin/pgf95 CC=/opt/pgi/linux86-64/6.2/bin/pgcc odd . -- michael
Re: [OMPI users] pgi compiler configure problems
On Dec 14, 2006, at 1:48 PM, Michael Galloway wrote: On Thu, Dec 14, 2006 at 02:01:54PM -0500, Michael Galloway wrote: good day all, i've been trying to build ompi with the 6.2-X version of the pgi compiler set (pgcc 6.2-4 64-bit target on x86-64 Linux). i've tried both 1.1.2 and the current nightly build 1.1.3b2r12766, both fail with this error from configure: *** Fortran 77 compiler checking whether we are using the GNU Fortran 77 compiler... no checking whether /opt/pgi/linux86-64/6.2/bin/pgf77 accepts -g... yes checking if Fortran 77 compiler works... yes checking /opt/pgi/linux86-64/6.2/bin/pgf77 external symbol convention... single underscore checking if Fortran 77 compiler supports LOGICAL... yes checking size of Fortran 77 LOGICAL... 4 checking for C type corresponding to LOGICAL... not found configure: WARNING: *** Did not find corresponding C type configure: error: Cannot continue It appears that this is being caused by a problem much earlier in the configure process. For some reason I don't totally understand, the configure script is finding the size of shorts, ints, longs, long longs, void *s, etc. to be 2. This is rather unexpected on a 64-bit machine, obviously. In testing with PGI 6.1-3 on x86_64, it looks like the IPA option is causing this to occur -- if I remove the IPA option from the flags, configure runs properly. In my (somewhat humble) opinion, turning on such aggressive options is dangerous when compiling Open MPI (or most MPI implementations). With share memory and RDMA interconnects, assumptions compiler writers make when running at such high optimization setting are frequently not true for codes like Open MPI. And the performance gain in compiling the MPI layer with such high optimizations is frequently near zero (that being said, there is a significant advantage to adding some optimization flags and not including debugging symbols in the MPI library if you don't need them). It should be safe to compile Open MPI without the IPA options and still use IPA when compiling your application. Brian -- Brian Barrett Open MPI Team, CCS-1 Los Alamos National Laboratory
