On Fri, Jan 18, 2008 at 08:33:10PM -0500, Jeff Squyres wrote:
> Barry --
>
> Could you check what apps are still running when it hangs? I.e., I
> assume that all the uptime's are dead; are all the orted's dead on the
> remote nodes? (orted = our helper process that is launched on the
> rem
On Jan 3, 2008, at 11:38 AM, Jim Kusznir wrote:
> error: line 300: Dependency tokens must begin with alpha-numeric,
'_'
> or '/': Requires: %{_name}-runtime
Huh..this is strange. Here's the chunk from my spec file and rpm
version. I've now built 3 sets of multi-rpm openmpi, each with a
diff
You're right. This fix is already on the trunk; it looks like we
missed bringing it to the v1.2 branch. I'll file a CMR against the
v1.2 branch in case we ever release 1.2.6.
Thanks!
On Dec 27, 2007, at 10:34 PM, Doug Reeder wrote:
Hello,
The attachment contains a short explanation of a
Barry --
Could you check what apps are still running when it hangs? I.e., I
assume that all the uptime's are dead; are all the orted's dead on the
remote nodes? (orted = our helper process that is launched on the
remote nodes to exert process control, funnel I/O back and forth to
mpirun
Are all three machines running the same OS and version, perchance? If
the machines are heterogeneous in terms of OS, glibc version, etc.,
weird things like these hangs can occur.
Additionally, are you running a firewall on any of these machines?
Ensure that iptables isn't running. It doe
FWIW, you might want to ask the ROMIO maintainers if this is a known
problem. I unfortunately have no idea. :-\
On Jan 16, 2008, at 12:12 PM, Brock Palen wrote:
Hello,
Using OpenMPI-1.2.3+pgi-7.0+hdf5 parallel + lustre is giving me the
error:
File locking failed in ADIOI_Set_lock. If the
I will be out of the office starting 19-01-2008 and will not return until 04-02-2008.
I will respond to your message when I return.
Do you have the Linux firewall running on either of your machines,
perchance? This can either block random socket connections between
nodes (which Open MPI's TCP communication will use) or eat the
connection requests in a black-hole fashion such that the connections
will timeout.
On Ja
You can do it, it might or might not make sense, depending on your
application. Load imbalance in regular MPI applications kills
performance. Therefore if your cluster is very heterogeneous, you
might prefer some different programming paradigm that take care of
this by nature (let say RPC).
Hi everyone,
I am new to open-mpi and parallel computing so I hope I won't
bore/offend you with obvious/off-topic questions.
We are running scientific simulations (using meep from mit) on small
bi-processors pcs and to fully use both processors on each machine, we
had to compile a mpi version
10 matches
Mail list logo
