[torquedev] [torqueusers] qsub -l nodes=x+procs=y VS qsub -l nodes=x, procs=y

Ken Nielson knielson at adaptivecomputing.com
Tue Mar 22 14:17:07 MDT 2011


On 03/22/2011 02:02 PM, Martin Siegert wrote:
> Hi John,
>
> On Tue, Mar 22, 2011 at 01:50:33PM -0600, John Rosenquist wrote:
>> Currently the code allows for both a +procs in a nodes value and a
>> distinct -l procs value.
>> Is there anyone using the -l nodes=x+procs style command line syntax?
>>
>> If not would there be any issue removing the duplicate functionality?
>> (i.e. -l procs=? would be support and support would be dropped for -l
>> nodes=x+procs=y)
>> If anyone is using this +procs style, would then be willing to change it
>> to the -l procs style?
>>
>> Partially for simplifying the code, partially for removing duplicate
>> functionality.
>>
>> Thanks.
>>
>> -- 
>> John Rosenquist | Torque Developer
>> Direct Line: 801.717.3--- | Fax: 801.717.3738
>> 1656 S. East Bay Blvd. Suite #300 | Provo, Utah 84601 | USA
>> Adaptive Computing, Ent.
>> _______________________________________________
>> torqueusers mailing list
>> torqueusers at supercluster.org
>> http://www.supercluster.org/mailman/listinfo/torqueusers
> I would love to use it, but since moab does not support it I can't.
> With respect to using either
>
> 1) -l nodes=x:ppn=y+procs=z
> or
> 2) -l nodes=x:ppn=y -l procs=z
>
> I really don't care, but I believe (2) should always work. Thus,
> the question really is whether (1) should work in addition (and
> synonymously) to (2). As far as I am concerned that is not really
> necessary.
>
> Cheers,
> Martin
>
As we have been pondering this question it seems to us that by 
eliminating #1 we simplify the syntax and hopefully make it easier to 
understand. Having multiple ways to do the same thing is not only more 
difficult to support, it gets confusing for the end user. I vote for 
supporting #2 only.

Ken


More information about the torquedev mailing list