Hey all,
Not sure that this is the right location to send this report to, might
be more appropriate for the dev list. Let me know if so.
The pre-release OpenMPI 1.7rc5 does not built correctly if you attempt
to compile it with a non-standard CUDA location. The culprit is in
ompi/contrib/vt/vt/vtl
FWIW, we have seen bugs in the intel compiler suite before. We usually advise
people to get the latest version of the particular intel compiler suite version
that they have a license to obtain.
I don't know if that is the exact cause of the issue here, but it's definitely
something we've seen
On 11/01/2012 09:56 AM, J R Jones wrote:
Hi
I am trying to build OpenMPI 1.6.3 with the Intel compilers, but one of the
tests in make check fails with a segmentation fault. I wondered if anyone
else has come across this and if so, how they got around it?
$ icc --version
icc (ICC) 12.0.4 201104
On 10/30/2012 05:22 PM, Jeff Squyres wrote:
The Open MPI Team, representing a consortium of research, academic, and
industry partners, is pleased to the next release in the stable release series:
Open MPI version 1.6.3.
Version 1.6.3 is a minor bugfix release. All users are encouraged to upgra
More as a note for other cygwin users than a general patch, but it
should work in any case.
The attached patch allow to enable shmem_sysv.
as SHM_R | SHM_W are not defined on cygwin (and on posix) [1]
Additional advise :on cygwin the SYSV shared memory requires the
cygserver service running, o
Very cool. Please let us know when you publish the package and give us a link
to it - would be nice to include that on our web page so Windows users have a
migration path now that native support is no longer available.
On Nov 1, 2012, at 12:08 PM, marco atzeri wrote:
> On 11/1/2012 7:55 PM, R
Try configuring --without-psm
That should solve the problem. We are probably picking up that you have PSM
libraries on the machine, but it looks like you aren't actually running it.
And yes - it should gracefully disable itself. You might check the 1.6 series
to see if it behaves better - if n
I am getting a problem where something called "PSM" is failing to start and
that in turn is preventing my job from running. Command and output are below.
I would like to understand what's going on. Apparently this version of OpenMPI
decided to build itself with support for PSM, but if it's no
There is 8GB memory on each node, with 6GB available, swap is off by commenting
it out in the /etc/fstab
I cannot try the alternate mechanisms right now, thanks for the info, will try
it when we move up to 1.6.1
-Original Message-
From: users-boun...@open-mpi.org [mailto:users-boun...@o
What's the memory usage on your nodes -- are you invoking swap, perchance?
Can you try one of the other shared memory mechanisms (sysv or posix)? (I just
described how in my previous email)
On Nov 1, 2012, at 11:24 AM, Hodge, Gary C wrote:
> George,
>
> We move 40K and 160K size messages fro
I don't know how to explain this any more than I have: Open MPI only uses those
files for an initial shared memory rendezvous point (and they're not really
"files", either). After that, all communication is done through shared memory.
Open MPI 1.6.x actually offers using multiple different type
Hi,
I have some problem in MPI Function MPI_Issend. That is non-blocking
send in synchronous mode. That is, when start this function, the send
process send the request-to-send to the destination process. Then the
receiver send back a permission-to-send message . And then the sender now
sends t
12 matches
Mail list logo
