Prev: plotting problem
Next: How to train a neural network with an empty input?
From: Matt J on 2 Aug 2010 13:44
Edric M Ellis <eellis(a)mathworks.com> wrote in message <ytwd3u12dca.fsf(a)uk-eellis-deb5-64.mathworks.co.uk>...

> That is true - but note that the MDCS workers will be drawn from a
> different "pool" than the "local" workers from PCT. With MDCS, you need
> to set up workers (possibly using the "jobmanager"), and you can place
> them where you wish. For example, you could run 16 MDCS workers across
> two machines, one of which is a desktop machine. You could then run
> PARFOR across them all.
=========

Great. Thanks a lot.
From: Alex Poe on 3 Aug 2010 19:38
Edric,

> To run across multiple nodes, you need "MATLAB Distributed Computing
> Server" on the cluster. See:
>
> http://www.mathworks.com/products/distriben/
>
> This allows you to submit a job to a remote cluster - you can then
> submit a job to multiple nodes to run your SPMD block across those
> nodes. (FYI - we use ScaLAPACK behind the scenes to solve "Ax = b" on
> the cluster).

Does MDCS need to be installed on the cluster?

> You mention "qsub" - are you using SGE or Torque? We have built-in
> integration with Torque; setting up SGE takes a little more work, but
> there are instructions telling you exactly what you need to do here.

It's Torque.

> > Do I submit the program from within MATLAB? Do I have to stay logged in (and
> > have MATLAB running) until the job is executed?

> If you use MDCS, then you do not need to stay logged in and you can
> collect your results later.

Is not at all possible to just write a simple .pbs script launching MATLAB -r <fun.m> across N-many nodes, and then qsub it?

Thank you,
--alex
From: Edric M Ellis on 5 Aug 2010 04:02
"Alex Poe" <wasteoff.nospam(a)gmail.com> writes:

>> To run across multiple nodes, you need "MATLAB Distributed Computing
>> Server" on the cluster. See:
>>
>> http://www.mathworks.com/products/distriben/
>>
>> This allows you to submit a job to a remote cluster - you can then
>> submit a job to multiple nodes to run your SPMD block across those
>> nodes. (FYI - we use ScaLAPACK behind the scenes to solve "Ax = b" on
>> the cluster).
>
> Does MDCS need to be installed on the cluster?

Yes, that's right. MDCS is basically a full MATLAB installation - but it
uses different licences.

>> You mention "qsub" - are you using SGE or Torque? We have built-in
>> integration with Torque; setting up SGE takes a little more work, but
>> there are instructions telling you exactly what you need to do here.
>
> It's Torque.

Good news, the setup should hopefully be relatively straightforward.

>> > Do I submit the program from within MATLAB? Do I have to stay logged in (and
>> > have MATLAB running) until the job is executed?
>
>> If you use MDCS, then you do not need to stay logged in and you can
>> collect your results later.
>
> Is not at all possible to just write a simple .pbs script launching
> MATLAB -r <fun.m> across N-many nodes, and then qsub it?

If you want communication enabled within the MATLAB instances on the
cluster, then you need to use our PCT/MDCS infrastructure to set up the
parallel job. If you only wish to run simple non-communicating tasks,
and you have sufficient MATLAB licences on the cluster, then that
approach will work. (MDCS worker licences are cheaper than standard
MATLAB licences, and allow you to effectively "transfer" licences for
toolboxes that you have on your desktop over to the cluster).

Cheers,

Edric.
First  |  Prev  | 
Pages: 1 2
Prev: plotting problem
Next: How to train a neural network with an empty input?