[Scilab-users] parallelization in scilab

[Paul Carrico]

Dear all,

I'm currently reflecting in a routine development with Scilab where the use
of parallelization is possible and highly useful/powerful.

The context is the following:

-          I've a mesh with N elements composed of 8 nodes each

-          For each node (completely independent from the others), I've to
perform a set of iterative calculus (through functions)

 

I had a look to "parallel_run" function, but  :

-          is it possible to use parallelization (I'm a complete newbie) ?

-          is it possible to call the same (own) functions & sub-functions
at the same time ?

-          how to proceed ? by block ?

-          where can I find useful documentation/tutorial ?

-          how to manage the number of processors ? (the code depends on
that number, isn't it ?)

 

>From my first thought, the flowchart is similar to :

-          processor 1 -> node 1

-          .

-          Processor x -> node x

(the number of processor is adjusted depending on the device I use : up to 8
processors on my working stations (Windows & Linux) - only 2 on my PC
(Windows & Linux))

 

Thanks for any advice

 

Paul

 

 

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://lists.scilab.org/pipermail/users/attachments/20121104/66aa4397/attachment.htm>