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

Re: [mpich-discuss] urgent-malloc problem
by Wesley Bland 12 Feb '14

12 Feb '14

Can you give more information about how your program is run? I tried to execute it, but it seems to expect some sort of input.

1 0

urgent-malloc problem
by Soheil Hooshdaran 12 Feb '14

12 Feb '14

Hello. I have a memory allocation problem (using malloc). I can't figure out its cause. Could you help me please?

1 0

Re: [mpich-discuss] Parallel test hanging with mpich on rhel7
by Orion Poplawski 11 Feb '14

11 Feb '14

On 02/10/2014 09:56 PM, Balaji, Pavan wrote: > That’s really weird. Errno 1 is "permission denied”. I don’t know how > that’s happening with gethostbyname. > > Can you send your mpiexec command line and a small program that reproduces > this error? E.g., if a program that just does MPI_INIT/MPI_FINALIZE > reproduces this error, that’ll be best. > > — Pavan > Reproduced with current nightly and an mpi hello world program: Fatal error in MPI_Init: Other MPI error, error stack: MPIR_Init_thread(467)..............: MPID_Init(177).....................: channel initialization failed MPIDI_CH3_Init(70).................: MPID_nem_init(319).................: MPID_nem_tcp_init(171).............: MPID_nem_tcp_get_business_card(418): MPID_nem_tcp_init(377).............: gethostbyname failed, i-0000205b (errno 1) For gethostbyname you need to check h_errno - and 1 is HOST_NOT_FOUND, which is the case for the Fedora builders - they cannot resolve even their own hostname. -- Orion Poplawski Technical Manager 303-415-9701 x222 NWRA/CoRA Division FAX: 303-415-9702 3380 Mitchell Lane orion(a)cora.nwra.com Boulder, CO 80301 http://www.cora.nwra.com

1 0

Re: [mpich-discuss] Parallel test hanging with mpich on rhel7
by Balaji, Pavan 10 Feb '14

10 Feb '14

On 2/10/14, 10:47 PM, "Orion Poplawski" <orion(a)cora.nwra.com> wrote: >On 02/06/2014 09:10 PM, Balaji, Pavan wrote: >> >> Thanks. That’s very useful analysis. Would you be willing to try the >> attached patch to see if it solves this issue? >> >> — Pavan > >Well, it seems to prevent a hang (although I'm also updating from 3.0.4 >to 3.1rc3 so not sure what is all changing here), but it does not run: It might be easier to use the nightly snapshots to make sure you are not missing some fixes: http://www.mpich.org/static/tarballs/nightly/master/mpich/ The patch I sent, as well as a few other patches after 3.1rc3, are all included in the nightly snapshots. >============================ >Fatal error in MPI_Init: Other MPI error, error stack: >MPIR_Init_thread(467)..............: >MPID_Init(177).....................: channel initialization failed >MPIDI_CH3_Init(70).................: >MPID_nem_init(319).................: >MPID_nem_tcp_init(171).............: >MPID_nem_tcp_get_business_card(418): >MPID_nem_tcp_init(377).............: gethostbyname failed, i-00001ff8 >(errno 1) >Fatal error in MPI_Init: Other MPI error, error stack: >MPIR_Init_thread(467)..............: >MPID_Init(177).....................: channel initialization failed >MPIDI_CH3_Init(70).................: >MPID_nem_init(319).................: >MPID_nem_tcp_init(171).............: >MPID_nem_tcp_get_business_card(418): >MPID_nem_tcp_init(377).............: gethostbyname failed, i-00001ff8 >(errno 1) That’s really weird. Errno 1 is "permission denied”. I don’t know how that’s happening with gethostbyname. Can you send your mpiexec command line and a small program that reproduces this error? E.g., if a program that just does MPI_INIT/MPI_FINALIZE reproduces this error, that’ll be best. — Pavan

1 0

Re: [mpich-discuss] Parallel test hanging with mpich on rhel7
by Orion Poplawski 10 Feb '14

10 Feb '14

On 02/06/2014 09:10 PM, Balaji, Pavan wrote: > > Thanks. Thatï¿½s very useful analysis. Would you be willing to try the > attached patch to see if it solves this issue? > > ï¿½ Pavan Well, it seems to prevent a hang (although I'm also updating from 3.0.4 to 3.1rc3 so not sure what is all changing here), but it does not run: ============================ Fatal error in MPI_Init: Other MPI error, error stack: MPIR_Init_thread(467)..............: MPID_Init(177).....................: channel initialization failed MPIDI_CH3_Init(70).................: MPID_nem_init(319).................: MPID_nem_tcp_init(171).............: MPID_nem_tcp_get_business_card(418): MPID_nem_tcp_init(377).............: gethostbyname failed, i-00001ff8 (errno 1) Fatal error in MPI_Init: Other MPI error, error stack: MPIR_Init_thread(467)..............: MPID_Init(177).....................: channel initialization failed MPIDI_CH3_Init(70).................: MPID_nem_init(319).................: MPID_nem_tcp_init(171).............: MPID_nem_tcp_get_business_card(418): MPID_nem_tcp_init(377).............: gethostbyname failed, i-00001ff8 (errno 1) =================================================================================== = BAD TERMINATION OF ONE OF YOUR APPLICATION PROCESSES = PID 19673 RUNNING AT i-00001ff8 = EXIT CODE: 1 = CLEANING UP REMAINING PROCESSES = YOU CAN IGNORE THE BELOW CLEANUP MESSAGES =================================================================================== -- Orion Poplawski Technical Manager 303-415-9701 x222 NWRA/CoRA Division FAX: 303-415-9702 3380 Mitchell Lane orion(a)cora.nwra.com Boulder, CO 80301 http://www.cora.nwra.com

1 0

Re: [mpich-discuss] [petsc-maint] updated to 3.4.3 and mpiexec jobs no longer run in background
by Jed Brown 09 Feb '14

09 Feb '14

Edward Bueler <elbueler(a)alaska.edu> writes: > One more thing, to further clarify that I am not doing something obviously > wrong. Here is what I see about my mpiexec. I can reproduce this. I don't know if the change was intentional, so I'm cross-posting. MPICH developers: a shell script containing multiple mpiexec commands no longer makes it past the first one. Apparently one now needs to set stdin to /dev/null to continue. $ cat t.sh #!/bin/sh $MPIEXEC -n 1 hostname $MPIEXEC -n 1 date ## MPICH-3.1b1 (also 3.0 series), "date" output not printed $ MPIEXEC=/opt/mpich/bin/mpiexec ./t.sh & [5] 29761 $ batura [5]+ Stopped MPIEXEC=/opt/mpich/bin/mpiexec ./t.sh ## Open MPI - note that "date" output is printed $ MPIEXEC=/usr/bin/mpiexec ./t.sh & [6] 29780 $ batura Sun Feb 9 14:12:35 MST 2014 [6] Done MPIEXEC=/usr/bin/mpiexec ./t.sh

1 0

Re: [mpich-discuss] Turn off EAGER send protocol.
by Anatoly G 09 Feb '14

09 Feb '14

Thank you very much. Sorry, I missed README.envvar file. Anatoly. On Sun, Feb 9, 2014 at 4:15 PM, Jeff Hammond <jeff.science(a)gmail.com> wrote: > Those are IBM-specific env vars. > > Look in the text docs in the root dir of MPICH. I recall all env vars are > listed there. > > Jeff > > Sent from my iPhone > > On Feb 9, 2014, at 2:19 AM, Anatoly G <anatolyrishon(a)gmail.com> wrote: > > Hi. > I can't find environment variable or parameter which should switch between > Eager and Rendezvous protocol in Mpich2 and Mpich-3.0.4. > Can you please suggest? > I tried: > *DCMF_EAGER 1000000000* > *MP_EAGER_LIMIT **1000000000* > > > Regards, > Anatoly. > > _______________________________________________ > 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] Turn off EAGER send protocol.
by Jeff Hammond 09 Feb '14

09 Feb '14

Those are IBM-specific env vars. Look in the text docs in the root dir of MPICH. I recall all env vars are listed there. Jeff Sent from my iPhone > On Feb 9, 2014, at 2:19 AM, Anatoly G <anatolyrishon(a)gmail.com> wrote: > > Hi. > I can't find environment variable or parameter which should switch between Eager and Rendezvous protocol in Mpich2 and Mpich-3.0.4. > Can you please suggest? > I tried: > DCMF_EAGER 1000000000 > MP_EAGER_LIMIT 1000000000 > > > Regards, > Anatoly. > _______________________________________________ > discuss mailing list discuss(a)mpich.org > To manage subscription options or unsubscribe: > https://lists.mpich.org/mailman/listinfo/discuss

1 0

Turn off EAGER send protocol.
by Anatoly G 09 Feb '14

09 Feb '14

Hi. I can't find environment variable or parameter which should switch between Eager and Rendezvous protocol in Mpich2 and Mpich-3.0.4. Can you please suggest? I tried: *DCMF_EAGER 1000000000* *MP_EAGER_LIMIT **1000000000* Regards, Anatoly.

1 0

Re: [mpich-discuss] MPICH 3.1rc3 packages for Ubuntu, Debian and Slackware (x86 and amd64)
by Balaji, Pavan 07 Feb '14

07 Feb '14

Thanks. The downloads page has been updated with this. http://www.mpich.org/downloads/ — Pavan On 2/7/14, 3:55 PM, "Marcelo Soares Souza" <marcelo(a)cebacad.net> wrote: > > mpich-3.1rc3 packages > >* Ubuntu 13.10/14.04 > >- amd64 >http://www.cebacad.net/files/mpich/ubuntu/mpich-3.1rc3/mpich_3.1rc3-1ubunt >u_amd64.deb > >- x86 >http://www.cebacad.net/files/mpich/ubuntu/mpich-3.1rc3/mpich_3.1rc3-1ubunt >u_i386.deb > >* Debian 7.3.0 (Wheezy) > >- amd64 >http://www.cebacad.net/files/mpich/debian/mpich-3.1rc3/mpich_3.1rc3-1debia >n_amd64.deb > >- x86 >http://www.cebacad.net/files/mpich/debian/mpich-3.1rc3/mpich_3.1rc3-1debia >n_i386.deb > >* Slackware 14.1 > >- amd64 >http://www.cebacad.net/files/mpich/slackware/mpich-3.1rc3/mpich-3.1rc3-x86 >_64-goa-1.txz > >- x86 >http://www.cebacad.net/files/mpich/slackware/mpich-3.1rc3/mpich-3.1rc3-i48 >6-goa-1.txz > >Enjoy! > >-- >Marcelo Soares Souza >https://marcelo.juntadados.org > >Diaspora* https://diaspora.juntadados.org/u/marcelo > >Friendica http://friendica.ambientedigital.org/profile/marcelo > >_______________________________________________ >discuss mailing list discuss(a)mpich.org >To manage subscription options or unsubscribe: >https://lists.mpich.org/mailman/listinfo/discuss

1 0