On Feb 20, 2008, at 9:53 AM, jessie puls wrote:
Secifically Jobs are not being handed to other nodes ever. Running
mpirun -mca btl openib,self -np 20 /bin/hostname
will return the same hostname 20 times, even if I specify -bynode as
an argument.
This is normal, and not an InfiniBand issu
On Feb 20, 2008, at 9:45 AM, Ben Allan wrote:
Our assumption was that if some other package defined these values,
they
would either likely be coming from the same standard autoconf tests
or
use the same #define conventions as the autoconf tests. As such, the
values that they are #defined to
On Feb 21, 2008, at 12:36 PM, George Bosilca wrote:
Here are the MCA parameters that you can use:
MCA btl: parameter "btl_openib_if_include" (current value:
)
Comma-delimited list of HCAs/ports to be used (e.g.
"mthca0,mthca1:2"; empty value means to
Thanks George!
I've added:
--mca btl_openib_if_include mthca0
--mca btl_openib_warn_default_gid_prefix 0
and hopefully it'll do the right thing without any warnings.
Bill
George Bosilca wrote:
Here are the MCA parameters that you can use:
MCA btl: parameter "btl_openib_if_include" (
Here are the MCA parameters that you can use:
MCA btl: parameter "btl_openib_if_include" (current value:
)
Comma-delimited list of HCAs/ports to be used (e.g.
"mthca0,mthca1:2"; empty value means to use all
ports found).
Mutuall
In trying to get openmpi up and running on a new cluster, I came across
this error about having both of my IB switches set to the same
subnet-gid. Snooping around on my hosts which run the opensm daemon, I
indeed found this to be the case in the /var/log/osm-ib[0-1].log files,
giving up find
Thanks a lot Kevin,
It seemed to me that something like your suggestion would
launch two unrelated "appname", ..., luckily I was wrong ;-).
Indeed it does what I want.
Thanks again,
Roberto
On Thu, 21 Feb 2008, Kevin Durda wrote:
> Hi Roberto,
>
> I think that you can do what you want i
I am not quite sure. It seems that your AR (advance reservation)
snapshot3 build is a bit new, and it may be a problem coming from it. I
am not quite familiar with this new SGE feature. I'd ping the gridengine
list to check on that error message coming from execd.
Neeraj Chourasia wrote:
Hell
Hi Roberto,
I think that you can do what you want if you use an appfile with something
like this:
-host node1,node2,node3 -np 6 -wdir /WorkingDir/ appname
-host node4 -np 2 -wdir /DifferentWorkingDir/ appname
Then run your program using "mpirun --app appfilename".
Kevin
On Wed, Feb 20, 2008 a
