Re: [OMPI devel] devel Digest, Vol 2917, Issue 1

I've downloaded the OpenMPI master as suggested and rerun all my aggregate tests across my system with QDR IB and 10 Gbps RoCE. The attached unidirectional.pdf graph is the ping-pong performance for 1 core on 1 machine to 1 core on the 2nd. The red curve for OpenMPI 1.8.3 shows lower pe

[OMPI devel] RFC: merge opal_free_list_t and ompi_free_list_t

What: Merge the opal_free_list_t and ompi_free_list_t implementations and add explicit interfaces for single and multi-threaded use. Why: Historically these two lists were different due to ompi_free_list_t dependencies in ompi (mpool). Those dependencies have since been moved to opal so it is saf

Re: [OMPI devel] Fortran issue

Sorry but I miss the connection between this test and the issue of TestAny in Fortran? On Thu, Feb 19, 2015 at 2:00 PM, Dave Goodell (dgoodell) wrote: > On Feb 19, 2015, at 10:15 AM, George Bosilca wrote: > > > While looking the MPI_Testany issue, I came across a very unsettling > sentence in

Re: [OMPI devel] Fortran issue

On Feb 19, 2015, at 10:15 AM, George Bosilca wrote: > While looking the MPI_Testany issue, I came across a very unsettling sentence > in the MPI standard (3.0 page 58 line 36). > > > The array is indexed from zero in C, and from one in Fortran. > > This sentence seems to indicate that the inde

Re: [OMPI devel] git commit id in coverity

On Feb 19, 2015, at 1:41 PM, Ralph Castain wrote: > > It’s documented on the wiki for new developers, but otherwise isn’t listed > anywhere (not that it’s secret): > > http://www.open-mpi.org/mailman/listinfo.cgi/testing And just for completeness, here's that wiki page: https://github.com/ope

Re: [OMPI devel] git commit id in coverity

It’s documented on the wiki for new developers, but otherwise isn’t listed anywhere (not that it’s secret): http://www.open-mpi.org/mailman/listinfo.cgi/testing > On Feb 19, 2015, at 10:35 AM, Howard Pritchard wrote: > > HI Ralph, > > H

Re: [OMPI devel] git commit id in coverity

HI Ralph, How does one get this "MPI Create success" message? Is there a mailing list specifically for the nightly builds? Thanks, Howard 2015-02-16 21:48 GMT-07:00 Ralph Castain : > It's the git id of the nightly tarball - which you should get via the MPI > Create Success message each night

Re: [OMPI devel] MCA Aliases

Thanks!!! On Thu, Feb 19, 2015 at 11:58 AM, Ralph Castain wrote: > We call it a “synonym” - but yes, you certainly can, and we do it in many > places. From mca_base_var.h: > > /** > * Register a synonym name for an MCA variable. > * > * @param[in] synonym_for The index of the original variabl

Re: [OMPI devel] MCA Aliases

We call it a “synonym” - but yes, you certainly can, and we do it in many places. From mca_base_var.h: /** * Register a synonym name for an MCA variable. * * @param[in] synonym_for The index of the original variable. This index * must not refer to a synonym. * @param[in] project_name The pro

[OMPI devel] MCA Aliases

Folks, Is it possible to define an alias for an MCA parameter? Grepping around the interwebs, it seems there was an RFC along these lines in 2008. http://www.open-mpi.org/community/lists/devel/2008/04/3613.php It doesn't appear that the functionality was added or, if it was, it has since been re

[OMPI devel] Fortran issue

While looking the MPI_Testany issue, I came across a very unsettling sentence in the MPI standard (3.0 page 58 line 36). > The array is indexed from zero in C, and from one in Fortran. This sentence seems to indicate that the index returned by the TestAny and TestSome (as well as the correspondin