[erlang-questions] Tilera 64-core chip - let's help them help us!

[G Bulmer]

On 6 Sep 2007, at 03:59, Kip Macy wrote:
>> I'd offer the same suggestions to Intel; if they want to get their  
>> 80-
>> core widely used, Erlang is the way to go, so support the  
>> community !-)
>
> Its coming, but will be targeted largely at graphics.
...
> I meant to say "to start off with".
>

Targeting at graphics doesn't mean ignore Erlang.

Several things worth considering:
1. Tim Sweeney's POPL 2006 talk about the 'next language for game  
playing' http://www.cs.princeton.edu/~dpw/popl/06/Tim-POPL.ppt
I must be clear, much of that is about *statically* typed functional  
languages, like Haskell, and not Erlang-style dynamic typing, but  
part is also concerned with concurrency.

2. erlang-questions Digest, Vol 4, Issue 8,  4. Stand Alone Erlang or  
Equivalent (Benjamin Tolputt)
I believe this is for the games industry.

3. NVIDIA and ATI are targeted on graphics, but they can see the  
value of supporting other industries through CUDA and CTM. Both are  
putting resources into non-graphic application areas.

A view of challenges for non-graphics uses for many-core chips comes  
from "The view from Berkeley" http://www.eecs.berkeley.edu/Pubs/ 
TechRpts/2006/EECS-2006-158.pdf
and also http://view.eecs.berkeley.edu/wiki/Main_Page as relevant.
One key point for Berkeley is "To maximize programmer productivity,  
programming models should be independent of the number of  
processors.", which also seems to agree with Tim Sweeney, and is a  
pretty good fit to Erlang.

Erlang is a practical tool available *now* which should map onto many  
independent cores (which seems to be Tilera's architecture) better  
than it maps onto e.g. NVIDIA 8800.

Providing it's a good fit , Tilera and Intel could broaden their  
technologies appeal and usefulness, by helping the community port  
Erlang. A good Erlang port could bring production-quality technology  
with an outstanding reputation early in their products lifecycle.

IMHO, getting a book as good as Joe Armstrong's will take significant  
time (which is precious right now for Tilera); it'd almost be worth  
supporting an Erlang port just to leverage Joe's book and the new  
interest in Erlang stimulated by it. Further, if Tilera can do it  
right, they could cover there costs. Intel have benefits to gain from  
a good Erlang port too, and HiPE may even work on the i80-core!  
What's not to like?

GB