[Forgot the attachment.]
On Thu, Dec 22, 2011 at 15:16, Jed Brown wrote:
> I wrote a new communication layer that we are evaluating for use in mesh
> management and PDE solvers, but it is based on MPI-2 one-sided operations
> (and will eventually benefit from some of the MPI-3 one-sided proposal
I wrote a new communication layer that we are evaluating for use in mesh
management and PDE solvers, but it is based on MPI-2 one-sided operations
(and will eventually benefit from some of the MPI-3 one-sided proposals,
especially MPI_Fetch_and_op() and dynamic windows). All the basic
functionality
Well, no single answer on this, try to revive the question.
Anybody: does someone have OpenMPI up'n'running using DAPL 2.0 on Linux?
Why: on some computers, using native openib (Intel says 'ofa') interface
leads to 50% performance loss (Badwidth 1.3 Gb/s instead of 3Gb/s with
DAPL+IntelMPI in
Hi,
Working on building on windows with MinGW. Have built libopen-pal,
opal-restart, opal-wrapper and am at the open-rte stage of the build.
The OBJ_CLASS_INSTANCE hits errors about 'initializer element is not
constant'. Open-pal also uses OBJ_CLASS_INSTANCE and doesn't throw this
error.
Tri
Hi,
On a Windows 7 64 bit box for mingw-w64 the CMake (version 2.8.6 and
openmpi-1.5.5rc1) selection for MinGW generation doesn't put -D in front
of preprocessor definitions such as _USRDLL, OMPI_IMPORTS, etc.
OMPI_C_DEF_PRE=""
SuSE supports an application mpi-selector which installs when you load
the (broken) SuSE version of OpenMPI. According to your choice of mpi
implementation it will correctly set PATH, MANPATH and LD_LIBRARY_PATH
for you.
The initial installation drops the usual pair (.sh, .csh) of scripts
into /e