Re: NG Lush and concurrency

Yann LeCun <[email protected]> Thu, 30 Mar 2006 22:40:15 -0500
Newsgroups gmane.lisp.lush.devel
Organization Courant Institute, NYU
Message-ID <[email protected]>
Hi Ralf,

For one thing, the low-level numerical functions could all
be implemented on top of an optimized library that can take
advantage of multiple cores/processors (e.g. Intel mkl,
Atlas,....).

For the higher level, we could provide parallel IDX
iterators. There is already idx-pbloop which is supposed 
to be a parallel version of idx-bloop (but currently just 
calls idx-bloop). 

For higher-level granularity, we could use MPI, and for 
coarse-grained parallelism, we can run multiple Lush processes 
through the remotelush class.

  -- Yann


> Multi-core architectures seem to be becoming the norm in desktop
> machines. To take advantage of multi-processor and multi-core
> architectures we need language constructs for expressing that
> certain operations may be carried out in parallel or change the
> semantics of existing operations where this can safely be done
> (many array operations may be carried out in parallel).
> On the one hand.
>
> On the other hand, a next generation lush runtime system should
> support parallel execution. How would that fit in here?
> Would we have multiple stack machines, running in different
> threads, plus a thread for the garbage collector, and a scheduler
> that assigns chunks of computations to idle stack machines?
>
> Ralf



-------------------------------------------------------
This SF.Net email is sponsored by xPML, a groundbreaking scripting language
that extends applications into web and mobile media. Attend the live webcast
and join the prime developer group breaking into this new coding territory!
http://sel.as-us.falkag.net/sel?cmd=lnk&kid=110944&bid=241720&dat=121642