On Jun 5, 2007, at 6:46 PM, Jeff Squyres wrote:
On Jun 5, 2007, at 4:51 PM, Lie-Quan Lee wrote:
Using the OPEN_MPI #define is probably the safest bet to identify the
MPI as Open MPI, which then opens the door to using mpi --
showme to find the proper linker flags. I would *NOT* recommend
h
On Jun 5, 2007, at 4:51 PM, Lie-Quan Lee wrote:
Using the OPEN_MPI #define is probably the safest bet to identify the
MPI as Open MPI, which then opens the door to using mpi --
showme to find the proper linker flags. I would *NOT* recommend hard-
coding specific flags just because you know that
Hi,
I am trying a program in which I have 2 MPI nodes and each MPI node has
2 threads:
Main node-thread Receive Thread
-
MPI_Init_Thread(MPI_THREAD_MULTIPLE);
.
.
LOOP: LOOP:
THRE
Jeff is right-- if you've already confirmed ompi,
just use the ompi specific arguments to get the
MPI LDFLAGS out. I withdraw the comment about adding
a feature to ompi_info.
It is unfortunate, but true, that the mpi compiler wrappers
give no hint of the existence of -show, --showme:link, etc
befo
Using the OPEN_MPI #define is probably the safest bet to identify the
MPI as Open MPI, which then opens the door to using mpi --
showme to find the proper linker flags. I would *NOT* recommend hard-
coding specific flags just because you know that you're linking
against Open MPI (because we can/h
On Jun 5, 2007, at 4:22 PM, Ben Allan wrote:
What you really want, after configure has confirmed openmpi with the
macro check, is to extract the libraries listed in the output of
mpif90 -v test.f
that are needed.
Ideally someone could update ompi_config to output the link flags
for 3 cases:
C
What you really want, after configure has confirmed openmpi with the
macro check, is to extract the libraries listed in the output of
mpif90 -v test.f
that are needed.
Ideally someone could update ompi_config to output the link flags for 3 cases:
C++ linking without using mpicxx
F90 linking with
Hi Rich -
All the releases back to the 0.9 pre-release included a #define of
OPEN_MPI to 1 in mpi.h, so that would be a good way to find out if
you are using Open MPI or not.
Hope this helps,
Brian
On Jun 5, 2007, at 1:36 PM, Lie-Quan Lee wrote:
I was thinking of this way. Is macro OPEN
I was thinking of this way. Is macro OPEN_MPI always defined in the openmpi
release? I search through FAQ and other docs and cannot confirm it. Or there
is another macro which is unique to openmpi?
--Rich
On 6/5/07, Anthony Chan wrote:
Never tried this myself, but this test could work
AC_CO
this is neat but it will need to identify whether the MPI implementation is
openmpi. -- Not all the implementations accept --showme flag.
--Rich
On 6/5/07, Nuno Sucena Almeida wrote:
Hi,
there's an autoconf macro you can find at
http://autoconf-archive.cryp.to/acx_mpi.html
Never tried this myself, but this test could work
AC_COMPILE_IFELSE( [
AC_LANG_PROGRAM( [
#include "mpi.h"
], [
#if defined( OPEN_MPI )
return 0;
#else
#error
#endif
] )
], [
mpi_is_openmpi=yes
], [
mpi_is_openmpi=no
] )
A.Chan
On Tue, 5 Jun 2007, Lie-Quan Lee wrote:
Hi,
there's an autoconf macro you can find at
http://autoconf-archive.cryp.to/acx_mpi.html
it might be useful to you. I prefer to use it just to test the
existence of
MPI and then inside configure.ac something like (ugly):
(...)
MPI_CXXLIBS=`mpicxx --showme:link
I am a project written in C/C++ but used 3rd-party libraries written in
Fortran. In the linking stage, I use mpic++ andl need to put -lmpi_f90
-lmpi_f77 into the link flags.
I am wondering if anyone has already had a way to identify openmpi in
configure script using autoconf macro. By identifying
hi everybody,
i'm new on this list and started using OpenMPI for my parallel jobs. first step
was to measure the latency for blocking communication functions. now my first
question: is it possible that ordained communication pairs will be optimized?
background:
latency for special processnumbe
Ralph,
Ralph H Castain wrote:
Hmmm...I think I know what may be happening. Could you send me:
1. what Open MPI version you are using?
Open MPI 1.2.1
2. any MCA parameters you might be setting in your environment (remember
that we may be picking up some system configuration file for those)
Hmmm...I think I know what may be happening. Could you send me:
1. what Open MPI version you are using?
2. any MCA parameters you might be setting in your environment (remember
that we may be picking up some system configuration file for those)
This isn't related to the problem, but I also note
FWIW, we do something similar in the openib BTL -- we use the subnet
ID to determine if two IB ports are connected (we have the rule in
OMPI that physically disconnected subnets must have different ID's --
this is more stringent than the IB spec calls for). See:
http://www.open-mpi.org/faq
Ralph,
Please do not bother about the output containing "src is (null) and orte
type is 0" in my previous email. It is just some printf I added to
dss_copy.c to make some sense of what is going wrong.
Prakash
>>> prakash.velayut...@cchmc.org 06/05/07 6:16 AM >>>
Hi,
Sorry about that. Two lines
Hi,
Sorry about that. Two lines got cut out from the program. Here is the
full program and error messages again. No Resource Manager involved,
just ssh/rsh.
Hostfile contains
bmi-opt2-01
bmi-opt2-02
bmi-opt2-03
bmi-opt2-04
#include
#include
#include
#include"mpi.h"
19 matches
Mail list logo
