I think we must have been hanging out in the same places. George
Morrow (Morrow Computers, one of the early CP/M brands) said way back
in the 70's "Anyone who believes what programmers say deserves what
happens to them".
On Oct 23, 2008, at 7:19 AM, Larry Young wrote:
> PC's are already about as fast serially as a supercomputer, so the
> only answer is massively parallel. This makes programming a
> nightmare except for problems that are huge and easily split into
> many separate unrelated tasks. CFD is not one of them. I remember
> back in the early 90s I would hear every year "This is the Year for
> Parallel Computing". I don't think is will ever happen.
>
> Bill Babcock wrote:
>> Super computers are either extremely fast in a serial
>> sense (limited) or massively parallel. If they are massively parallel
>> than everything that runs on them needs to be optimized up the wazoo,
>> and it's terrifically difficult to get stuff to work well.
>>
Bill Babcock
Babcock & Jenkins
Billb@bnj.com
503.936.7660
www.bnj.com
Editor
Ke Nalu e-Magazine
Paddlesurfing's Web Journal
Bill@kenalu.com
www.kenalu.com
blog: www.ponohouse.com/ponoblog
_______________________________________________
Support Team.Net http://www.team.net/donate.html
http://www.fot-racing.com
Fot mailing list
Fot@autox.team.net
http://autox.team.net/mailman/listinfo/fot
|