[VoxBo] Breaking regressions into parts?

[Thad Polk]

On Fri, 23 Mar 2007, Thad Polk wrote:

> I note that vbregress includes a -p flag that makes it possible to break the 
> regression into parts. Is there a reason the glm widget doesn't exploit this 
> feature to parallelize regressions across machines? Is there an easy way to 
> fit a glm and get it to exploit multiple CPUs? For example, is there a vbprep 
> script that will generate a job sequence comparable to the job sequence 
> created by the glm widget (and in which vbregress could be told to use 
> parts)?

Not an hour after asking why the glm widget doesn't split regressions into 
parts, I see a sequence on my cluster, submitted by the glm widget, doing 
just that!  So let me rephrase the questions: Under what circumstances 
does the glm break the regression up? Into how many parts? Is this 
something the user can control?

Thanks,

Thad