On 1 February 2012 19:42, Brian Smith <[email protected]> wrote:
> I've started a github page for some tools I've put together from various
> bits of code, how-tos, etc. to simplify the setup of parallel
> environments so that they work universally for all MPI implementations
> (on x86_64 Linux) w/ tight-integration support (no support for ssh yet).
>  The syntax for submitting parallel jobs becomes more similar to
> LSF/PBS/Torque and provides for easy configuration of your task layout
> (ppn,nodes,pcpus,pcpus_min,pcpus_max).  We use a JSV to make the magic
> happen.  We create PEs tied to queues since our queues often delineate
> changes in the underlying communication fabrics available.
>
It seems to me that you have 2 projects there not one.
i)Universal PE which simplifies a lot of things.
ii)A PBS emulation layer.

i)Is certainly useful and we're moving in that direction here as well
although some quirks of our system mean
I don't think we'll be able to just use your version.
ii)I think is less useful.    I find the Torque nodes nodes,ppn syntax
to be less expressive than Grid Engine's request the resources you
want and let the scheduler
figure it out method particularly in a heterogenous environment.
Also "If you want Torque you know where to find it".  I think this
would be more useful as an optional extra for those used to Torque.
eg: module load PBSemulation could add modify the path so that just
typing qsub invoked something like this
#!/bin/bash
$SGE_ROOT/bin/lx26-amd64/qsub -jsv $SGE_ROOT/util/torque-emulation.pl "$@"

_______________________________________________
users mailing list
[email protected]
https://gridengine.org/mailman/listinfo/users

Reply via email to