[Openmcl-devel] Speed, compilers and multi-core processors

[Paul Krueger]

I completely agree with this. Before I retired from Cray last summer I  
was primarily investigating alternative computational paradigms for  
both massively multi-threaded and multi-core architectures. A couple  
of years ago, as Cray's representative, I attended a small NSF- 
sponsored conference  of about 6 industry and 6 university  
representatives to discuss what sorts of research should be done to be  
ready for the multi-core architectures that we are beginning to see  
now. Microsoft's representative (Burton Smith, who was previously the  
architect of Cray's multi-threaded supercomputers) called this the  
preeminent current challenge for the software industry.

Subsequently I identified a number of potential applications and  
computational approaches that are (or could be made to be) consistent  
with massively multi-threaded and/or multi-core architectures. These  
approaches included graph algorithms, genetic algorithms, fuzzy logic,  
memory-resident databases, and several others. Most of these would  
require quite a bit of software infrastructure development before they  
would be easy to use. These research/development suggestions got an  
enthusiastic reception from the technical community within Cray, but  
they couldn't stand up to a business-case analysis. There is a chicken- 
and-egg problem for Cray (and presumably for most vendors other than  
Intel and Microsoft) in that they can't afford to invest in this until  
there is sufficient demand; and there won't be much demand until users  
have a chance to try out and discover good uses for the new paradigms.  
I toyed with the idea of trying to promote a CCL port to Cray's multi- 
threaded architecture, but initial discussion with other senior  
management convinced me that it would never fly for lack of current  
demand. I briefly had visions of having access to the world's largest  
and fastest lisp machine; oh well  ...

NSF and DARPA will undoubtedly fund research in this area and they are  
known to take some chances on new ideas. But given how I saw them  
being pulled by influential industry and university factions, I expect  
that much of the funding will be used for small extensions to current  
paradigms rather than doing anything too radical. I hope I'm wrong  
about that.

So the hope that I still have is that people  such as those who  
subscribe to this newsgroup will continue to push the boundaries on  
their own; that CCL will turn into a really good multi-core tool to  
facilitate such research, and that eventually others will see the  
value too. Good luck to you all!

Paul

On May 19, 2009, at 7:05 AM, Alexander Repenning wrote:

> not so fast ;-)
>
> The "how can we make use of multiple cores" is currently on the the  
> hottest funding topics supported by NSF, DOE, Microsoft, .....
>
> Perhaps it is the Lisp way to look at architectures such as the x86  
> and see mostly limitations when indeed there are plenty of  
> opportunities. This is not about registers but about enabling end  
> user programmers such as scientists to make use of parallelism. The  
> big question is how to reconceptualize programming. One of the main  
> problems is the need to overcome bad algorithmic assumptions  
> especially the use of unnecessary loops. For instance, in  
> Bioinformatics textbooks are full of loop based implementations of  
> algorithms dealing with huge data structures such as gene sequences.  
> In many cases one could replace sequential loops with parallel  
> execution.
>
> Zoom out of the low level view of things. What could multi core Lisp  
> do? Look at the computational challenges that users are dealing  
> with. Try to come up with new computational paradigms that could  
> help. Lisp could be a great platform to explore these issues.  
> Careful: if you can contribute to this you may actually receive  
> funding.
>
> alex
>
>
>
> On May 18, 2009, at 10:45 AM, Brian Mastenbrook wrote:
>
>> On Mon, 2009-05-18 at 10:13 -0400, Glen Foy wrote:
>>
>>> My ignorance of compiler design is breathtaking, but could multi- 
>>> core
>>> compiler techniques be used to compensate for Intel's register- 
>>> starved
>>> architecture?
>>
>> In a word, no.
>
> Prof. Alexander Repenning
>
> University of Colorado
> Computer Science Department
> Boulder, CO 80309-430
>
> vCard: http://www.cs.colorado.edu/~ralex/AlexanderRepenning.vcf
>
>
> _______________________________________________
> Openmcl-devel mailing list
> Openmcl-devel at clozure.com
> http://clozure.com/mailman/listinfo/openmcl-devel

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://lists.clozure.com/pipermail/openmcl-devel/attachments/20090519/2e068918/attachment.htm>