April 2014 - discuss - mailman3.cels.anl.gov

Re: [mpich-discuss] mpiexec
by Nicolas Rosner 09 Apr '14

09 Apr '14

You need to create a text file containing the hostnames of the machines you want to use, on separate lines. Optionally, you can also append ":N" (where N is an integer) to a line in order to specify how many ranks (processes) you want on that particular machine. This is explained and exemplified here: http://wiki.mpich.org/mpich/index.php/Using_the_Hydra_Process_Manager Cheers N On Wed, Apr 9, 2014 at 3:57 AM, Mojtaba S. Moghani <mshokrollahi(a)ut.ac.ir> wrote: > How could I define a list of machines? to issue the mpiexec command in this > step : mpiexec -n 5 -f machinefile ./examples/cpi > > could you tell me after installing step, what should I exactly do then > > thanks, > > -- > Mojtaba S. Moghani > Faculty of Environment > University of Tehran > > > _______________________________________________ > discuss mailing list discuss(a)mpich.org > To manage subscription options or unsubscribe: > https://lists.mpich.org/mailman/listinfo/discuss

1 0

Re: [mpich-discuss] mpiexec
by fereshteh komijani 09 Apr '14

09 Apr '14

as i know before machine file you nead ./ as bellow: mpiexec -n 5 -f ./machinefile ./examples/cpi On Wed, Apr 9, 2014 at 11:27 AM, Mojtaba S. Moghani <mshokrollahi(a)ut.ac.ir>wrote: > How could I define a list of machines? to issue the mpiexec command in > this step : mpiexec -n 5 -f machinefile ./examples/cpi > > could you tell me after installing step, what should I exactly do then > > thanks, > > -- > Mojtaba S. Moghani > Faculty of Environment > University of Tehran > > > _______________________________________________ > discuss mailing list discuss(a)mpich.org > To manage subscription options or unsubscribe: > https://lists.mpich.org/mailman/listinfo/discuss > -- ***Angel***

1 0

mpiexec
by Mojtaba S. Moghani 09 Apr '14

09 Apr '14

How could I define a list of machines? to issue the mpiexec command in this step : mpiexec -n 5 -f machinefile ./examples/cpi could you tell me after installing step, what should I exactly do then thanks, -- Mojtaba S. Moghani Faculty of Environment University of Tehran

1 0

Re: [mpich-discuss] lldb and mpich issue:
by Balaji, Pavan 08 Apr '14

08 Apr '14

Drew, Thanks for reporting this issue. I’m able to reproduce this on a mac with lldb. On some initial digging it looks like lldb is messing with UNIX sockets created between mpiexec’s proxy and the MPI process. lldb should be seeing that socket, but it doesn’t seem to hand it off to the application process. That’s nasty. I’ve created a ticket for this: https://trac.mpich.org/projects/mpich/ticket/2063 The simplest solution will be to give an option in mpiexec to avoid the UNIX sockets path completely and use TCP/IP sockets. We don’t want to do this by default because: (1) it requires the system to allow internal TCP connections and (2) UNIX sockets can be a bit faster in some cases. We’ll try to get this into mpich soon. Thanks, — Pavan On Apr 8, 2014, at 8:19 PM, Drew Lewis <drew90(a)vt.edu> wrote: > Dear MPICH discuss group, > > Sorry if this is a duplicate, I tried to send it before being subscribed, and am not sure if it ever made it to the list. > > I am encountering an issue when trying to use lldb with both mpich 3.1 and 3.0.4. > > The command that generates the error is > > computer:tests ME$ mpirun -np 1 lldb ./a.out > Current executable set to './a.out' (x86_64). > (lldb) run > run > Process 67202 launched: './a.out' (x86_64) > [cli_0]: write_line error; fd=6 buf=:cmd=init pmi_version=1 pmi_subversion=1 > : > system msg for write_line failure : Bad file descriptor > [cli_0]: Unable to write to PMI_fd > [cli_0]: write_line error; fd=6 buf=:cmd=get_appnum > : > system msg for write_line failure : Bad file descriptor > Fatal error in MPI_Init: Other MPI error, error stack: > MPIR_Init_thread(467): > MPID_Init(140).......: channel initialization failed > MPID_Init(422).......: PMI_Get_appnum returned -1 > Process 67202 exited with status = 1 (0x00000001) > > mpich 3.1 was configured with > $SRC_DIR/configure \ > --prefix=$INSTALL_ROOT/$BUILD_DIR \ > --disable-cxx \ > --disable-f77 \ > --disable-fc \ > CC=clang \ > CXX=clang++ \ > > The version of lldb is > lldb-310.2.36 > > The minimal example that shows the problem is > #include <mpi.h> > > int main(int argc, char**argv){ > MPI_Init(&argc, &argv); > MPI_Finalize(); > return 0; > } > compiled with mpicc > > Since I know it has been an issue in the past the shell I am using is > GNU bash, version 3.2.51(1)-release (x86_64-apple-darwin13) > Copyright (C) 2007 Free Software Foundation, Inc. > > any help you can offer would be much appreciated. > > Thank you, > -Drew Lewis > _______________________________________________ > discuss mailing list discuss(a)mpich.org > To manage subscription options or unsubscribe: > https://lists.mpich.org/mailman/listinfo/discuss

1 0

lldb and mpich issue:
by Drew Lewis 08 Apr '14

08 Apr '14

Dear MPICH discuss group, Sorry if this is a duplicate, I tried to send it before being subscribed, and am not sure if it ever made it to the list. I am encountering an issue when trying to use lldb with both mpich 3.1 and 3.0.4. The command that generates the error is computer:tests ME$ mpirun -np 1 lldb ./a.out Current executable set to './a.out' (x86_64). (lldb) run run Process 67202 launched: './a.out' (x86_64) [cli_0]: write_line error; fd=6 buf=:cmd=init pmi_version=1 pmi_subversion=1 : system msg for write_line failure : Bad file descriptor [cli_0]: Unable to write to PMI_fd [cli_0]: write_line error; fd=6 buf=:cmd=get_appnum : system msg for write_line failure : Bad file descriptor Fatal error in MPI_Init: Other MPI error, error stack: MPIR_Init_thread(467): MPID_Init(140).......: channel initialization failed MPID_Init(422).......: PMI_Get_appnum returned -1 Process 67202 exited with status = 1 (0x00000001) mpich 3.1 was configured with $SRC_DIR/configure \ --prefix=$INSTALL_ROOT/$BUILD_DIR \ --disable-cxx \ --disable-f77 \ --disable-fc \ CC=clang \ CXX=clang++ \ The version of lldb is lldb-310.2.36 The minimal example that shows the problem is #include <mpi.h> int main(int argc, char**argv){ MPI_Init(&argc, &argv); MPI_Finalize(); return 0; } compiled with mpicc Since I know it has been an issue in the past the shell I am using is GNU bash, version 3.2.51(1)-release (x86_64-apple-darwin13) Copyright (C) 2007 Free Software Foundation, Inc. any help you can offer would be much appreciated. Thank you, -Drew Lewis

1 0

Re: [mpich-discuss] mpiexec
by Reuti 08 Apr '14

08 Apr '14

Am 07.04.2014 um 14:06 schrieb Mojtaba S. Moghani: > After this commad : mpiexec -n 5 -f machinefile ./examples/cpi, it has some errors when run, errors like : > > > HYDU_parse_hostfile (/home/mojtaba/libraries/mpich-3.1/src/pm/hydra/utils/args/args.c:319): unable to open host file: machinefile Did you define a list of machines in this file, and is the file at the same location from where you issue the `mpiexec` command? -- Reuti > [mpiexec@Mojtaba] mfile_fn (/home/mojtaba/libraries/mpich-3.1/src/pm/hydra/ui/mpich/utils.c:336): error parsing hostfile > [mpiexec@Mojtaba] match_arg (/home/mojtaba/libraries/mpich-3.1/src/pm/hydra/utils/args/args.c:152): match handler returned error > [mpiexec@Mojtaba] HYDU_parse_array (/home/mojtaba/libraries/mpich-3.1/src/pm/hydra/utils/args/args.c:174): argument matching returned error > [mpiexec@Mojtaba] parse_args (/home/mojtaba/libraries/mpich-3.1/src/pm/hydra/ui/mpich/utils.c:1596): error parsing input array > [mpiexec@Mojtaba] HYD_uii_mpx_get_parameters (/home/mojtaba/libraries/mpich-3.1/src/pm/hydra/ui/mpich/utils.c:1648): unable to parse user arguments > [mpiexec@Mojtaba] main (/home/mojtaba/libraries/mpich-3.1/src/pm/hydra/ui/mpich/mpiexec.c:153): error parsing parameters > > > what is the problem? > > thanks, > > > -- > Mojtaba S. Moghani > Faculty of Environment > University of Tehran > > _______________________________________________ > discuss mailing list discuss(a)mpich.org > To manage subscription options or unsubscribe: > https://lists.mpich.org/mailman/listinfo/discuss

1 0

mpich 3 and apple's newest lldb
by Drew Lewis 08 Apr '14

08 Apr '14

Dear MPICH discuss group, I am encountering an issue when trying to use lldb with both mpich 3.1 and 3.0.4. The command that generates the error is computer:tests ME$ mpirun -np 1 lldb ./a.out Current executable set to './a.out' (x86_64). (lldb) run run Process 67202 launched: './a.out' (x86_64) [cli_0]: write_line error; fd=6 buf=:cmd=init pmi_version=1 pmi_subversion=1 : system msg for write_line failure : Bad file descriptor [cli_0]: Unable to write to PMI_fd [cli_0]: write_line error; fd=6 buf=:cmd=get_appnum : system msg for write_line failure : Bad file descriptor Fatal error in MPI_Init: Other MPI error, error stack: MPIR_Init_thread(467): MPID_Init(140).......: channel initialization failed MPID_Init(422).......: PMI_Get_appnum returned -1 Process 67202 exited with status = 1 (0x00000001) mpich 3.1 was configured with $SRC_DIR/configure \ --prefix=$INSTALL_ROOT/$BUILD_DIR \ --disable-cxx \ --disable-f77 \ --disable-fc \ CC=clang \ CXX=clang++ \ The version of lldb is lldb-310.2.36 The minimal example that shows the problem is #include <mpi.h> int main(int argc, char**argv){ MPI_Init(&argc, &argv); MPI_Finalize(); return 0; } compiled with mpicc Since I know it has been an issue in the past the shell I am using is GNU bash, version 3.2.51(1)-release (x86_64-apple-darwin13) Copyright (C) 2007 Free Software Foundation, Inc. any help you can offer would be much appreciated. Thank you, -Drew Lewis

1 0

Re: [mpich-discuss] mpich2 - checkpointing error
by Wesley Bland 08 Apr '14

08 Apr '14

I'm not sure of the last version in which it worked. I'd have to dig back through the archives to try it out. On Tuesday, April 8, 2014, Marcelo Paiva Ramos <marcelo.paiva(a)cptec.inpe.br> wrote: > It works in another version of MPICH2 and BLCR? > > Best regards, > Marcelo. > > > On 07-04-2014 09:16, Wesley Bland wrote: > > Unfortunately, this is a known problem at the moment. BLCR checkpointing > hasn't worked for a few versions now. It's something we're working to fix > in a future version. > > Thanks, > Wesley > > On Monday, April 7, 2014, Marcelo Paiva Ramos <marcelo.paiva(a)cptec.inpe.br> > wrote: > > Hi, > Can you help me to solve this problem? > > cat /etc/issue > CentOS release 6.5 (Final) > Kernel \r on an \m > > uname -a > Linux server 2.6.32-431.11.2.el6.x86_64 #1 SMP Tue Mar 25 19:59:55 UTC > 2014 x86_64 x86_64 x86_64 GNU/Linux > > *INSTALL: blcr-0.8.5* > tar xzvf blcr-0.8.5.tar.gz > cd blcr-0.8.5 > mkdir builddir > cd builddir > ../configure --prefix=/opt/blcr > make > make install > /sbin/insmod /opt/blcr/lib/blcr/2.6.32-431.11.2.el6.x86_64/blcr_imports.ko > /sbin/insmod /opt/blcr/lib/blcr/2.6.32-431.11.2.el6.x86_64/blcr.ko > uname -r > 2.6.32-431.11.2.el6.x86_64 > > _______________________________________________ > discuss mailing list discuss(a)mpich.org <javascript:_e(%7B%7D,'cvml','discuss(a)mpich.org');> > To manage subscription options or unsubscribe:https://lists.mpich.org/mailman/listinfo/discuss > > >

1 0

Re: [mpich-discuss] statically linked mpiexec
by Andy May 08 Apr '14

08 Apr '14

Ken, Thanks, that improves things such that the number of shared libraries being referenced decreases, but it's still not quite there yet: ./configure --with-device=ch3:nemesis --with-pm=hydra CFLAGS=-static --disable-shared ldd ../install/bin/mpiexec linux-vdso.so.1 (0x00007fff435fe000) libnsl.so.1 => /lib64/libnsl.so.1 (0x00007fc8c4d2a000) libm.so.6 => /lib64/libm.so.6 (0x00007fc8c4a27000) libpthread.so.0 => /lib64/libpthread.so.0 (0x00007fc8c4809000) libgcc_s.so.1 => /lib64/libgcc_s.so.1 (0x00007fc8c45f2000) libc.so.6 => /lib64/libc.so.6 (0x00007fc8c4243000) libdl.so.2 => /lib64/libdl.so.2 (0x00007fc8c403f000) /lib64/ld-linux-x86-64.so.2 (0x00007fc8c4f42000) I tried also adding 'LDFLAGS=-static' too, it doesn't change anything. Any ideas? Thanks, Andy On 07/04/14 15:58, Kenneth Raffenetti wrote: > Ah, sorry. I misread your first email and assumed you were only talking > about the MPICH libraries and compiler wrappers. > > The Intel compilers should accept a '-static' flag that will force link > all libraries statically. Try adding it to your CFLAGS to force this > behavior on mpiexec. > > Ken > > On 04/07/2014 09:50 AM, Andy May wrote: >> Ken, >> >> The '--disable-shared' flag removes some of the external references (to >> the Intel compiler in my case) but still leaves: >> >> ldd ./bin/mpiexec >> linux-vdso.so.1 (0x00007fff7a7b9000) >> libnsl.so.1 => /lib64/libnsl.so.1 (0x00007f3298e3f000) >> libpthread.so.0 => /lib64/libpthread.so.0 (0x00007f3298c21000) >> libpciaccess.so.0 => /usr/lib64/libpciaccess.so.0 >> (0x00007f3298a17000) >> libxml2.so.2 => /usr/lib64/libxml2.so.2 (0x00007f32986ac000) >> libdl.so.2 => /lib64/libdl.so.2 (0x00007f32984a8000) >> libz.so.1 => /lib64/libz.so.1 (0x00007f3298292000) >> liblzma.so.5 => /usr/lib64/liblzma.so.5 (0x00007f329806c000) >> libm.so.6 => /lib64/libm.so.6 (0x00007f3297d69000) >> libgcc_s.so.1 => /lib64/libgcc_s.so.1 (0x00007f3297b52000) >> libc.so.6 => /lib64/libc.so.6 (0x00007f32977a3000) >> /lib64/ld-linux-x86-64.so.2 (0x00007f3299057000) >> >> whereas what I did gives: >> >> ldd ./bin/mpiexec >> not a dynamic executable >> >> Is there any option to ensure fully statically linked mpiexec via >> configure? >> >> Thanks, >> >> Andy >> >> On 07/04/14 13:26, Kenneth Raffenetti wrote: >>> Andy, >>> >>> The official way is to add '--disable-shared' to your configure line. >>> >>> Ken >>> >>> On 04/07/2014 05:47 AM, Andy May wrote: >>>> Hi, >>>> >>>> Does anyone know if there is a supported or recommended way to build >>>> MPICH such that it's binaries (eg. mpiexec) are statically linked? >>>> >>>> At present, post building I'm running: >>>> >>>> cd src/pm/hydra >>>> eval ./configure `./config.status --config` --disable-pci >>>> --disable-libxml2 >>>> make CCLD='$(CC) -all-static' >>>> >>>> which just disables two cases where the static libraries aren't >>>> available, then links statically, but is there a better or correct way >>>> to do this? >>>> >>>> Many thanks, >>>> >>>> Andy >>>> _______________________________________________ >>>> discuss mailing list discuss(a)mpich.org >>>> To manage subscription options or unsubscribe: >>>> https://lists.mpich.org/mailman/listinfo/discuss >>> _______________________________________________ >>> discuss mailing list discuss(a)mpich.org >>> To manage subscription options or unsubscribe: >>> https://lists.mpich.org/mailman/listinfo/discuss >> _______________________________________________ >> discuss mailing list discuss(a)mpich.org >> To manage subscription options or unsubscribe: >> https://lists.mpich.org/mailman/listinfo/discuss > _______________________________________________ > discuss mailing list discuss(a)mpich.org > To manage subscription options or unsubscribe: > https://lists.mpich.org/mailman/listinfo/discuss

1 0

Re: [mpich-discuss] mpich2 - checkpointing error
by Marcelo Paiva Ramos 08 Apr '14

08 Apr '14

It works in another version of MPICH2 and BLCR? Best regards, Marcelo. On 07-04-2014 09:16, Wesley Bland wrote: > Unfortunately, this is a known problem at the moment. BLCR > checkpointing hasn't worked for a few versions now. It's something > we're working to fix in a future version. > > Thanks, > Wesley > > On Monday, April 7, 2014, Marcelo Paiva Ramos > <marcelo.paiva(a)cptec.inpe.br <mailto:[email protected]>> wrote: > > Hi, > Can you help me to solve this problem? > > cat /etc/issue > CentOS release 6.5 (Final) > Kernel \r on an \m > > uname -a > Linux server 2.6.32-431.11.2.el6.x86_64 #1 SMP Tue Mar 25 19:59:55 > UTC 2014 x86_64 x86_64 x86_64 GNU/Linux > > *INSTALL: blcr-0.8.5* > tar xzvf blcr-0.8.5.tar.gz > cd blcr-0.8.5 > mkdir builddir > cd builddir > ../configure --prefix=/opt/blcr > make > make install > /sbin/insmod > /opt/blcr/lib/blcr/2.6.32-431.11.2.el6.x86_64/blcr_imports.ko > /sbin/insmod /opt/blcr/lib/blcr/2.6.32-431.11.2.el6.x86_64/blcr.ko > uname -r > 2.6.32-431.11.2.el6.x86_64 > lsmod | grep blcr > blcr 115465 0 > blcr_imports 10715 1 blcr > ldconfig -p | grep blcr > libcr_run.so.0 (libc6,x86-64) => /opt/blcr/lib/libcr_run.so.0 > libcr_run.so (libc6,x86-64) => /opt/blcr/lib/libcr_run.so > libcr_omit.so.0 (libc6,x86-64) => /opt/blcr/lib/libcr_omit.so.0 > libcr_omit.so (libc6,x86-64) => /opt/blcr/lib/libcr_omit.so > libcr.so.0 (libc6,x86-64) => /opt/blcr/lib/libcr.so.0 > libcr.so (libc6,x86-64) => /opt/blcr/lib/libcr.so > chkconfig --list | grep blcr > blcr 0:off 1:off 2:on 3:on 4:on 5:on > 6:off > > *INSTALL: mpich-3.1* > tar xzvf mpich-3.1.tar.gz > cd mpich-3.1 > ./configure --disable-fast CFLAGS=-O2 FFLAGS=-O2 CXXFLAGS=-O2 > FCFLAGS=-O2 --prefix=/opt/mpich2/ CC=/opt/intel/bin/icc > FC=/opt/intel/bin/ifort F77=/opt/intel/bin/ifort > --enable-checkpointing --with-hydra-ckpointlib=blcr > --with-blcr=/opt/blcr --with-blcr-include=/opt/blcr/include > --with-blcr-lib=/opt/blcr/lib > make > make install > > *.bashrc* > export > PATH=$PATH:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/opt/blcr/bin:/opt/mpich2/bin > export > LD_LIBRARY_PATH=/usr/local/lib:/usr/local/lib64:/opt/intel/lib/intel64:/opt/blcr/lib:/opt/mpich2/lib > > *mpiexec -info* > HYDRA build details: > Version: 3.1 > Release Date: Thu Feb 20 11:41:13 > CST 2014 > CC: /opt/intel/bin/icc -O2 > CXX: g++ -O2 > F77: /opt/intel/bin/ifort -O2 > F90: /opt/intel/bin/ifort -O2 > Configure options: '--disable-option-checking' > '--prefix=/opt/mpich2' '--disable-fast' 'CFLAGS=-O2 -O0' > 'FFLAGS=-O2 -O0' 'CXXFLAGS=-O2 ' 'FCFLAGS=-O2 ' > 'CC=/opt/intel/bin/icc' 'FC=/opt/intel/bin/ifort' > 'F77=/opt/intel/bin/ifort' '--enable-checkpointing' > '--with-hydra-ckpointlib=blcr' '--with-blcr=/opt/blcr' > '--with-blcr-include=/opt/blcr/include' > '--with-blcr-lib=/opt/blcr/lib' '--cache-file=/dev/null' > '--srcdir=.' 'LDFLAGS= -L/opt/blcr/lib' 'LIBS=-lrt -lcr -lpthread > ' 'CPPFLAGS= -I/root/mpich-3.1/src/mpl/include > -I/root/mpich-3.1/src/mpl/include -I/root/mpich-3.1/src/openpa/src > -I/root/mpich-3.1/src/openpa/src > -I/root/mpich-3.1/src/mpi/romio/include -I/opt/blcr/include' > Process Manager: pmi > Launchers available: ssh rsh fork slurm ll > lsf sge pbs manual persist > Topology libraries available: hwloc > Resource management kernels available: user slurm ll lsf sge > pbs cobalt > Checkpointing libraries available: blcr > Demux engines available: poll select > > > *ERROR* > mpiexec -n 1 -ckpointlib blcr -ckpoint-interval 20 -ckpoint-prefix > /home/marcelo/TESTE/ ./teste > [proxy:0:0@server] requesting checkpoint > [proxy:0:0@server] checkpoint completed > [proxy:0:0@server] HYDT_ckpoint_blcr_checkpoint > (tools/ckpoint/blcr/ckpoint_blcr.c:241): Checkpointing failed. > Make sure BLCR kernel module is loaded. Unknown error 2356 > [proxy:0:0@server] ckpoint_thread (tools/ckpoint/ckpoint.c:76): > blcr checkpoint returned error > [proxy:0:0@server] requesting checkpoint > [proxy:0:0@server] checkpoint completed > [proxy:0:0@server] HYDT_ckpoint_blcr_checkpoint > (tools/ckpoint/blcr/ckpoint_blcr.c:241): Checkpointing failed. > Make sure BLCR kernel module is loaded. Unknown error 2356 > [proxy:0:0@server] ckpoint_thread (tools/ckpoint/ckpoint.c:76): > blcr checkpoint returned error > > Best regards, > Marcelo. > > > > _______________________________________________ > discuss mailing list discuss(a)mpich.org > To manage subscription options or unsubscribe: > https://lists.mpich.org/mailman/listinfo/discuss

1 0