[OMPI devel] OMPI collectives

Hey folks I’ve got someone asking about any documentation and/or papers out there that might summarize the algorithms used in the bcast, allreduce, barrier, and alltoall collectives - and might describe the analytical cost of each algo (i.e., #steps, etc). The papers on our web site are gettin

Re: [OMPI devel] configure.m4 for pmix/s1 and pmix/s2 question

Hi Paul, Yes, that is the minor problem I was referring to. It does in fact reflect the oldness of CLE 4. The cray pmi 5 and higher is newer software which probably should never have been installed on CLE 4, since the alps packaging changed completely between CLE 4 and 5. Howard 2014-10-28 14

Re: [OMPI devel] configure.m4 for pmix/s1 and pmix/s2 question

By Howard's definition I guess NERSC's Hopper (XE6) qualifies as "very old" at PrgEnv 4.2.34 {hargrove@hopper06 ~}$ pkg-config --cflags cray-pmi Package cray-alpslli was not found in the pkg-config search path. Perhaps you should add the directory containing `cray-alpslli.pc' to the PKG_CONFIG_PAT

Re: [OMPI devel] configure.m4 for pmix/s1 and pmix/s2 question

Hi Ralph, Oh on the cray, you don't need to specify the --with-pmi, except to say you either want a particular directory (for instance if you wanted to try your luck with s2 on a cray nativized slurm), or you want to say --with-pmi=no. Howard 2014-10-28 14:14 GMT-06:00 Ralph Castain : > > On

Re: [OMPI devel] configure.m4 for pmix/s1 and pmix/s2 question

HI Ralph, I think I found the problem. Thanks. Howard 2014-10-28 12:58 GMT-06:00 Ralph Castain : > > On Oct 28, 2014, at 11:53 AM, Howard Pritchard > wrote: > > Hi Ralph, > > > 2014-10-28 12:26 GMT-06:00 Ralph Castain : > >> >> > On Oct 28, 2014, at 11:16 AM, Howard Pritchard >> wrote: >> >

Re: [OMPI devel] configure.m4 for pmix/s1 and pmix/s2 question

> On Oct 28, 2014, at 1:05 PM, Howard Pritchard wrote: > > Hi Folks, > > The simplest and best way on cray is to use the pkg-config command. > No looking for odd header file names, etc. There is a minor issue > with external login nodes running very old (like CLE 4.X) that one has > to workaro

Re: [OMPI devel] configure.m4 for pmix/s1 and pmix/s2 question

On Tue, Oct 28, 2014 at 12:20 PM, Ralph Castain wrote: > On Oct 28, 2014, at 12:17 PM, Paul Hargrove wrote: > > Ralph, > > The Cray's at NERSC have *both* pmi_cray.h and pmi.h (and pmi2.h as well). > > > I understand that - I was questioning if that is universally true or not. > IF we are guaran

Re: [OMPI devel] configure.m4 for pmix/s1 and pmix/s2 question

Hi Folks, The simplest and best way on cray is to use the pkg-config command. No looking for odd header file names, etc. There is a minor issue with external login nodes running very old (like CLE 4.X) that one has to workaround, but otherwise works well. pkg-config --cflags cray-pmi etc. etc.

Re: [OMPI devel] configure.m4 for pmix/s1 and pmix/s2 question

> On Oct 28, 2014, at 12:17 PM, Paul Hargrove wrote: > > Ralph, > > The Cray's at NERSC have *both* pmi_cray.h and pmi.h (and pmi2.h as well). I understand that - I was questioning if that is universally true or not. IF we are guaranteed that nobody with a Cray ever renames pmi_cray.h to pmi.

Re: [OMPI devel] configure.m4 for pmix/s1 and pmix/s2 question

Ralph, The Cray's at NERSC have *both* pmi_cray.h and pmi.h (and pmi2.h as well). That is why I said our configure logic checks for pmi_cray.h *first*. Sorry if that wasn't clear. On NERSC's XE6: {hargrove@hopper06 ~}$ ls /opt/cray/pmi/default/include/ pmi2.h pmi_cray_ext.h pmi_cray.h pmi.h

Re: [OMPI devel] configure.m4 for pmix/s1 and pmix/s2 question

> On Oct 28, 2014, at 11:59 AM, Paul Hargrove wrote: > > > On Tue, Oct 28, 2014 at 11:53 AM, Howard Pritchard > wrote: > > We may no longer require those as you have separated the Cray check out, but > the original problem is that we would pickup the Slurm compone

Re: [OMPI devel] configure.m4 for pmix/s1 and pmix/s2 question

On Tue, Oct 28, 2014 at 11:53 AM, Howard Pritchard wrote: > >> We may no longer require those as you have separated the Cray check out, >> but the original problem is that we would pickup the Slurm components on >> the Cray because we would find pmi.h >> >> Oh, I forgot about that . > In GASNet

Re: [OMPI devel] configure.m4 for pmix/s1 and pmix/s2 question

> On Oct 28, 2014, at 11:53 AM, Howard Pritchard wrote: > > Hi Ralph, > > > 2014-10-28 12:26 GMT-06:00 Ralph Castain >: > > > On Oct 28, 2014, at 11:16 AM, Howard Pritchard > > wrote: > > > > Hi Folks, > > > > I'm trying to figure out wh

Re: [OMPI devel] configure.m4 for pmix/s1 and pmix/s2 question

Hi Ralph, 2014-10-28 12:26 GMT-06:00 Ralph Castain : > > > On Oct 28, 2014, at 11:16 AM, Howard Pritchard > wrote: > > > > Hi Folks, > > > > I'm trying to figure out what broke for pmi configure since now the > pmix/cray component > > doesn't compile any longer in master. > > Ouch - sorry about

Re: [OMPI devel] configure.m4 for pmix/s1 and pmix/s2 question

> On Oct 28, 2014, at 11:16 AM, Howard Pritchard wrote: > > Hi Folks, > > I'm trying to figure out what broke for pmi configure since now the pmix/cray > component > doesn't compile any longer in master. Ouch - sorry about that. I thought the Cray component strictly used the new Cray PMI che

[OMPI devel] configure.m4 for pmix/s1 and pmix/s2 question

Hi Folks, I'm trying to figure out what broke for pmi configure since now the pmix/cray component doesn't compile any longer in master. I was happening to look in the s1 and s2 configure.m4's and noticed a AC_REQUIRE for OPAL_CHECK_UGNI. This doesn't make sense to me. Maybe these were accidenta

Re: [OMPI devel] [mpich-discuss] ROMIO+Lustre problems in OpenMPI 1.8.3

On 10/28/2014 06:00 AM, Paul Kapinos wrote: Dear Open MPI and ROMIO developer, We use Open MPI v.1.6.x and 1.8.x in our cluster. We have Lustre file system; we wish to use MPI_IO. So the OpenMPI's are compiled with this flag: > --with-io-romio-flags='--with-file-system=testfs+ufs+nfs+lustre'

Re: [OMPI devel] 1.8.3 and PSM errors

Good to know. I will update the infinipath libraries on the next occasion and report back. This will probably take a few days (or weeks). Adrian On Mon, Oct 27, 2014 at 10:22:08PM +, Friedley, Andrew wrote: > Hi Adrian, > > I'm unable to reproduce here with OMPI v1.8.3 (I ass

Re: [OMPI devel] fixing a bug in 1.8 that's not in master

I just updated the wiki: NOTE: Pull requests on ompi-release must include a hash reference in the body/comments corresponding to the commit(s) on ompi:master from which it is derived, OR indicate that this is solely a release branch bug (i.e., there's no corresponding commit on ompi:master beca

[OMPI devel] ROMIO+Lustre problems in OpenMPI 1.8.3

Dear Open MPI and ROMIO developer, We use Open MPI v.1.6.x and 1.8.x in our cluster. We have Lustre file system; we wish to use MPI_IO. So the OpenMPI's are compiled with this flag: > --with-io-romio-flags='--with-file-system=testfs+ufs+nfs+lustre' In our newest installation openmpi/1.8.3 we fou