There is an ancient (undocumented?) pthread interface that can distribute a 
threadloop onto several threads. It is not automagic though but requires 
setting on a per-piddle basis AFAIR and also works only if the number of 
threads is a divisor of the threadloop dim. This is from memory and may only be 
partially correct.

Christian
 
On 6/08/2010, at 5:39 AM, Craig DeForest wrote:

> The threading engine was always intended to support multiple CPUs, but  
> nobody has yet taken the time to hack in and do it.  It "shouldn't be  
> too hard" to break up the outermost threadloop and hand off part of  
> the loop to multiple processors for each atomic threaded operation,  
> but nobody has yet done it.   The SMOP is duplicating the inner loop  
> variables between threads so they don't collide with one another.
> 
> 
> 
> 
> On Aug 5, 2010, at 10:34 AM, Daniel Carrera wrote:
> 
>> On Thu, Aug 5, 2010 at 6:17 PM, Ingo Schmid <[email protected]> wrote:
>>> On an unrelated note, is there a plan to support using multiple CPUs,
>>> parallelisation of some kind, or did I miss something and it is
>>> available already?
>> 
>> I don't know if PDL supports this, but if it doesn't, I'd imagine that
>> the threading feature is a good candidate for using multiple CPUs.
>> 
>> -- 
>> Intolerant people should be shot.
>> 
>> _______________________________________________
>> Perldl mailing list
>> [email protected]
>> http://mailman.jach.hawaii.edu/mailman/listinfo/perldl
>> 
> 
> 
> _______________________________________________
> Perldl mailing list
> [email protected]
> http://mailman.jach.hawaii.edu/mailman/listinfo/perldl

--
Christian Soeller PhD   Dept. of Physiology  +64 9 3737599 x82770
University of Auckland  Auckland, New Zealand  fax +64 9 3737499



_______________________________________________
Perldl mailing list
[email protected]
http://mailman.jach.hawaii.edu/mailman/listinfo/perldl

Reply via email to