Re: [OMPI users] how to build a relocatable OpenMPI RPM ?

2013-09-10 Thread caohaijun
thanks. :) On Tue, Sep 10, 2013 at 10:19 PM, Ralph Castain wrote: > Ha - needed to read one email further down :-) > > Sent from my iPhone > > On Sep 10, 2013, at 3:03 AM, caohaijun wrote: > > I got it. Thank you very much. > > > Best, > Jimmy > > > On Tue, Sep 10, 2013 at 5:02 PM, Jeff Squyr

Re: [OMPI users] how to build a relocatable OpenMPI RPM ?

2013-09-10 Thread Ralph Castain
Ha - needed to read one email further down :-) Sent from my iPhone On Sep 10, 2013, at 3:03 AM, caohaijun wrote: > I got it. Thank you very much. > > > Best, > Jimmy > > > On Tue, Sep 10, 2013 at 5:02 PM, Jeff Squyres (jsquyres) > wrote: >> Note that Open MPI hard-codes its installation d

Re: [OMPI users] how to build a relocatable OpenMPI RPM ?

2013-09-10 Thread Ralph Castain
Let me be sure I understand. You configure without a prefix and then tell rpm to install at a specified prefix? You may need to add something to envar - will look tonite Sent from my iPhone On Sep 10, 2013, at 2:09 AM, caohaijun wrote: > Hi Ralph > > I tried to put prefix in PATH and LD_LIBR

Re: [OMPI users] Max number of processes per host for an OMPI run?

2013-09-10 Thread Ralph Castain
You're hitting max files due to socket requirements Sent from my iPhone On Sep 10, 2013, at 7:10 AM, Francesco Simula wrote: > is:

[OMPI users] Max number of processes per host for an OMPI run?

2013-09-10 Thread Francesco Simula
Dear forum, I probably must apologize in advance for the very basic question but I wasn't able to find an answer elsewhere: how do I find the maximum number of processes that can be concurrently instantiated by mpirun on one single host of a cluster? If I launch (on an CentOS 6.3 cluster with

Re: [OMPI users] MPI_IN_PLACE in a call to MPI_Allreduce in Fortran

2013-09-10 Thread Hugo Gagnon
I only get the correct output when I use the more "conventional" syntax: ... call MPI_Allreduce(a_loc,a,2,MPI_INTEGER,MPI_SUM,MPI_COMM_WORLD,ierr) ... However, I get the wrong output when I use MPI_IN_PLACE: ... MPI_Allreduce(MPI_IN_PLACE,a,2,MPI_INTEGER,MPI_SUM,MPI_COMM_WORLD,ierr) ... hence m

Re: [OMPI users] how to build a relocatable OpenMPI RPM ?

2013-09-10 Thread caohaijun
I got it. Thank you very much. Best, Jimmy On Tue, Sep 10, 2013 at 5:02 PM, Jeff Squyres (jsquyres) wrote: > Note that Open MPI hard-codes its installation directory into some of its > libraries and executables (so that it can find its data files, etc.). So > we don't support relocatable RPM

Re: [OMPI users] MPI_IN_PLACE in a call to MPI_Allreduce in Fortran

2013-09-10 Thread Jeff Squyres (jsquyres)
On Sep 7, 2013, at 5:14 AM, Hugo Gagnon wrote: > $ openmpif90 test.f90 > $ openmpirun -np 2 a.out > 0 4 6 > 1 4 6 > > Now I'd be curious to know why your OpenMPI implementation handles > MPI_IN_PLACE correctly and not mine! I don't un

Re: [OMPI users] how to build a relocatable OpenMPI RPM ?

2013-09-10 Thread caohaijun
Hi Ralph I tried to put prefix in PATH and LD_LIBRARY_PATH, but the problem still occurs. Now my envs in ~/.bashrc are as follows (--prefix=/root/program/orte): export PATH=/root/program/orte:/root/program/orte/bin:$PATH export LD_LIBRARY_PATH=/root/program/orte:/root/program/orte/lib64:/root/pr

Re: [OMPI users] how to build a relocatable OpenMPI RPM ?

2013-09-10 Thread Jeff Squyres (jsquyres)
Note that Open MPI hard-codes its installation directory into some of its libraries and executables (so that it can find its data files, etc.). So we don't support relocatable RPMs, in the RPM sense of the word "relocatable". You can, however, override OMPI's hard-coded installation location wi