[Wien] Pathscale+OpenMPI support

[Scott Beardsley]

>> A 32cpu job would always look like this:
>>
>> lapw0: compute-0-0:8 compute-0-1:8 compute-0-2:8 compute-0-3:8
>> 1:compute-0-0:8
>> 1:compute-0-1:8
>> 1:compute-0-2:8
>> 1:compute-0-3:8
>> granularity:1
>> extrafine:1
> 
> Please note: If a user runs a job with only 2 k-points (or with 5 !!), 
> lots of resources get waisted, because only half of the machines are 
> actually used (but all 32 nodes are reserved).

Do you happen to have a somewhat short (time wise - say <15mins for 
32cpus) "real world" input that I can test out to demonstrate the 
differences? I did some runs using strictly mpi parallelization and 
didn't notice much of a difference (effectively zero) between the above 
machines file and the below machines file. Results here[1].

lapw0: compute-0-0:8 compute-0-1:8 compute-0-2:8 compute-0-3:8
1:compute-0-0:8 compute-0-1:8 compute-0-2:8 compute-0-3:8
granularity:1
extrafine:1

>> Also, I'm curious if w2web is batch queue aware? If not, I'll have to 
>> discourage people from using it.
> 
> In w2web one can configure the "execution type". Predefined is just 
> "interactive" and "background", but one can configure its own modes,
> eg sge8, sge16, ... under "configuration" (directly in w2web).

Thanks for the info!
Scott
-----------
[1] http://wiki.cse.ucdavis.edu/support:hpc:software:wien2k#benchmarks