[Users] Cactus build problems
Dumsani Ndzinisa
g14n8326 at campus.ru.ac.za
Wed Aug 20 07:56:24 CDT 2014
Hi Ian,
As far as I know, those limits are set in the machine definition file,
the .ini file specifically. Once you have added them on, do a fresh
build, making sure to pass the ini file as an (optional) argument to
you new build. That is, simfactory/bin/sim build .....
--machine=simfactory/mdb/machines/<machinename.ini>
If not set, they will take on some default values which might limit
you on submission when you now try to ask for more cores
(processors) than the set (default) maximum. This seems to be
the trouble in your case.
It is also possible for the ini file that you are currently using not to
have these variables. In that case, you will need to add them. See
an example ini file in the files shipped with the code (in
simfactory/mdb/machines/).
I hope that helps shed some light.
Regards,
Dumsani
On 20/08/2014 14:35, Ian Smith wrote:
> Hi Guys,
>
> I tried to run the resource-reduced example (static_tov_small.par) in
> the simplified guide, but the simulation only uses one CPU core. I
> tried a couple of things to fix this:
>
> ian at sumo Wed Aug 20 01:26 PM /opt/EinsteinToolkit/Cactus
> $ ./simfactory/bin/sim submit static_tov
> --parfile=par/static_tov_small.par --procs=2 --walltime=8:0:0
> Assigned restart id: 3
> Error: Too many nodes specified: nodes=2 (maxnodes is 1)
> Aborting Simfactory.
>
> ian at sumo Wed Aug 20 01:27 PM /opt/EinsteinToolkit/Cactus
> $ ./simfactory/bin/sim submit static_tov
> --parfile=par/static_tov_small.par --ppn=2 --procs=2 --walltime=8:0:0
> Assigned restart id: 4
> Error: Illegal number of requested cores per node specified: specified
> ppn=2 (min-ppn is 1, max-ppn is 1)
> Aborting Simfactory.
>
> I can't see where these max & min limits are set, I checked in
> simfactory/etc, can anyone enlighten me?
>
> Cheers,
>
> Ian.
>
> _______________________________________________
> Users mailing list
> Users at einsteintoolkit.org
> http://lists.einsteintoolkit.org/mailman/listinfo/users
More information about the Users
mailing list