Re: computing breakthrough - full text
Wed, 10 Feb 1999 10:29:21 -0800

My guess is that they haven't actually constructed the full sized
$26 million system, but a smaller version with fewer chips. That's
one nice thing about parallel processing systems; you can build the size to suit the client's budget.

I worked in parallel processing back in the 80s and we always announced the biggest possible configuration of our systems, but we couldn't afford to build them until we had orders for them.

The reconfigurable architecture sounds good in theory but in practice the question is how difficult it will be to program the machine for practical problems. That was the big problem with the Connection Machine; you had to use special programming languages to use it efficiently.

Hal (no relation)