Hi all,
I notice in the last couple of weeks there was a patch with
ALL_DEPENDENCIES to fix CMake 2.8 builds with on Windows. With CMake
2.8 I'm getting exactly the same build errors in r22504 as in the 1.4.1
release. Has that patch made it into the snapshots yet, or is there a
regression?
Hi.
Let's say I have an MPI application that runs on several hosts. I want to
suspend the application. I do that by sending to the mpirun process the
signal TSTP. Is there any way to measure how long does it take to the
application to completely suspend? Doing this "time kill -TSTP PID" will
measu
Its definitely not a bug in Lustre - its an essential part of any
"coherent" cluster filesystem (CXFS, QFS, GFS, GPFS, ). The
whole point is that some people actually want to have meaningful
(non-garbage) files accessed using properly managed parallel IO
techniques. The locking is necessary
Agreed that it is not good (and I am recoding some programs to avoid
this), but (and here life gets interesting) is this a "bug" in Lustre?
On Tue, Feb 2, 2010 at 5:59 PM, David Singleton
wrote:
>
> But its a very bad idea on a "coherent", "POSIX" filesystem like Lustre.
> Locks have to bounce ar
But its a very bad idea on a "coherent", "POSIX" filesystem like Lustre.
Locks have to bounce around between the nodes for every write. This can
be VERY slow (even for trivial amounts of "logging" IO) and thrash the
filesystem for other users. So, yes, at our site, we include this sort
of "par
I know it's wrong, but I don't think it is forbidden (which I
guess is what you are saying).
On Tue, Feb 2, 2010 at 5:31 PM, Jeff Squyres wrote:
> +1 on Nick's responses.
>
> AFAIK, if you don't mind getting garbage in the output file, it should be
> fine to do. Specifically: it should not
+1 on Nick's responses.
AFAIK, if you don't mind getting garbage in the output file, it should be fine
to do. Specifically: it should not cause OS issues (crash, reboot, corrupted
filesystem, etc.) to do this -- but the file contents will likely be garbage.
That being said, this situation like
Thanks for the bug report!
I puzzled over your text for a while before I figured out what I think the
problem is -- I think the argument you've given to --with-valgrind is the
*source* tree for Valgrind. You actually want to give the *installation*
directory as the argument to --with-valgrind.
Hi Laurence,
I don't know whether it's as bad as a deadly sin, but for us parallel writes
are a huge problem and we get complete garbage in the file. Take a look at:
Implementing MPI-IO Atomic Mode and Shared File Pointers Using MPI One-Sided
Communication, * Robert Latham,*Robert Ross*, *Rajeev
hmmm, that's really strage, as the orted could be launched on remote
machine, that means your WMI is working properly, and of course your
settings are correct, but I don't know why the launched orted couldn't
talk to the head node, and why the mpirun crashes. There could be other
things caus
I get the same behavior regardless of which machine I execute mpirun from.
The security settings I altered were to add the user to both DCOM
Access and Launch with all options set to allow.
In WMI, I add the user to the CIMV2 namespace and grant all permissions
except edit security.
Windows firew
Shiqing
mpirun launches local apps fine from both machines.
I'll have to work though the configuration again on the current head
node so it can be connected to from the other machine.
Thanks!
Charles
On 2/2/2010 12:49 PM, Shiqing Fan wrote:
>
> Hi Charles,
>
> I've downloaded a fresh Open MPI
Hi Charles,
I've downloaded a fresh Open MPI 1.4.1 tarball, and built a clean
installer. After installing it onto two machines, everything worked for
me, I tested to launch notedpad and also other simple MPI applications
like barrier, they all worked. So there still might be some settings got
Thanks, I'll give it a try!
Lee Manko
On Tue, Feb 2, 2010 at 10:01 AM, Ralph Castain wrote:
> Probably the easiest solution is to tell OMPI not to use the second NIC.
> For example, if that NIC is eth1, then you could do this:
>
> mpirun -mca oob_tcp_if_exclude eth1 -mca btl_tcp_if_exclude eth
Probably the easiest solution is to tell OMPI not to use the second NIC. For
example, if that NIC is eth1, then you could do this:
mpirun -mca oob_tcp_if_exclude eth1 -mca btl_tcp_if_exclude eth1 ...
This tells both the MPI layer and the RTE to ignore the eth1 interface.
On Tue, Feb 2, 2010 at
The connection between nodes should be TCP/IP. I am currently using
Open MPI 1.4.1.
I've attached the output of ompi_info as a text file.
Thanks!
Charles
On 2/2/2010 11:46 AM, Shiqing Fan wrote:
>
> Hi Charles,
>
> It seems not a WMI problem, because the remote orted has already been
> launch
Hi Charles,
It seems not a WMI problem, because the remote orted has already been
launched, and only that part was done by WMI.
What connection do you have between the nodes, tcp? Could you provide
the version information of Open MPI, or just the output of ompi_info? so
that I can take a cl
No messages on the command prompt.
When I executed mpirun to launch notepad on the remote machine, It
crashes again.
No information is ever printed to the command line unless I enter a bad
password.
The very first time I attempt to use mpirun to launch a process on the
remote machine, I get an i
Hi Charles,
On the local machine, which can be considered also as the "head node",
no orted will be launched, the mpirun itself works as it locally.
Did you see any error message on the command prompt? That would be very
helpful.
To do a simple test, just try to launch notepad on remote n
Thank you Jody and Ralph. Your suggestions got me up and running (well sort
of). I have run into another issue that I was wondering if you have had any
experience with. My server has one NIC that is static and a second that is
DHCP on a corp network (the only way to get to the outside world). M
This is my first attempt at configuring a Beowulf cluster running MPI. ALL
of the nodes in the cluster are PS3s running Yellow Dog Linux 6.2 and the
host (server) is a Dell i686 Quad-core running Fedora Core 12. The cluster
is running openMPI v1.4.1 configured (non-homogeneous), compiled and
inst
Thanks Shiqing!
Unfortunately, it still doesn't work, but I've got more info.
I can use wmic to start an application on the remote machine, but that
application does not start in the current login process (notepad.exe
starts, but I have to ask task manager to show all processes to find it,
even t
I have a question concerning having many processors in a mpi job all
write to the same file -- not using mpi calls but with standard
fortran I/O. I know that this can lead to consistency issues, but it
can also lead to OS issues with some flavors of nfs.
At least in fortran, there is nothing "wron
Hi DevL, what compiler and options are you using to build OMPI. I am
seeing something similar (Warning messages and the Message Queue window
having bizarre values) when building with the Pathscale compiler but I
don't see this with SunStudio, gcc, Intel or PGI.
However, I do see pending recei
Can you send all the information listed on the getting help page on the ompi
web site? Also, information about your application would be helpful.
-jms
Sent from my PDA. No type good.
From: users-boun...@open-mpi.org
To: us...@openmpi.org
Sent: Tue Feb 02 0
Hi -
configure failed the VALGRIND_CHECK_MEM_IS_ADDRESSABLE test, for
openmpi-1.3.3 or -1.4 with Valgrind 3.5.0 -
I ran configure with
! /bin/csh
#! run configure
#
../configure --prefix=/home/pshare/lib/lf95/openmpi-Vg-1.3.3 \
FC=`which lf95` F77=`which lf95`
Hi,
I experience some stange behaviour on multi-core node of our cluster that I
presume is linked to openmpi.
When running for a long time, and several pseudo-nodes of a single multicore
node are concerned, one process freezes in a uninterrutible mode (D status)
and the others seem to wait fo
27 matches
Mail list logo
