[Beowulf] Question: Who has apps that regularly swap?

[John Hearns]

On Tue, 2004-08-31 at 14:38, Sean Dilda wrote:
> On Tue, 2004-08-31 at 00:41, Mark Hahn wrote:
> 
> > actually, that leads to another survey question: do you schedule
> > your job by loadaverage or according to a resource-allocation scheme?
> > when a job runs on our systems, it's pretty much guaranteed a 
> > free CPU.  this actually causes some strife with people who want 
> > to run non-compute-bound jobs, or jobs which have dynamic amounts
> > of parallelism...
> 
> For my cluster its somewhat of both.  We use SGE which by default will
> schedule the job on the machine with the lowest load average, with some
> special allocation rules used for parallel jobs.  
> it.  All in all I think it works pretty well once you've twisted all the
> knobs the right way.

Absolutely. 
Though I'd like to add that it is possible to use other metrics than the
default load average if necessary, and tune this. 
In the jargon 'load sensors' - and this could even be floating licenses
for eg. Verilog or something.
(Not that I've ever had call to do it.)

For instance, there was a discussion on the Gridengine mailing list
today from someone who wanted to run many, many processes per node.
One of the suggestion was to switch off the default load average
behaviour, and then jobs will be dispatched to the nodes.

Apologies if this has been too SGE specific.