I am going to move ahead and work with calling 'fsl_sub randomise_parallel' directly.
If anyone feels this is not the correct way to run randomise on a multi-node cluster, I would appreciate some correction.
Thanks,
- BettyAnn
> Hello all,
>
> Could someone help me understand the mechanics of running randomise in
> a parallel fashion?
>
> I am running on a PBS cluster. I have copied and modified fsl_sub in
> my own directory. How do I run randomise with my PBS-modified fsl_sub
> script?
>
> Since randomise is 'auto-aware' of the cluster, does it just 'find'
> fsl_sub? If so, how? How will it find my PBS-modified version?
>
> Or do I need to explicitly run randomise_parallel thru fsl_sub like
> so:
>
> ./fsl_sub randomise_parallel -i <input> ...
>
> After searching around, I am still confused. I don't understand who
> calls what in what order.
>
> Thanks,
> - BettyAnn
|