[extropy-chat] Using Graphic controllers
extropy at unreasonable.com
Sun Apr 18 15:01:02 UTC 2004
Eugen Leitl wrote:
>GPUs have a usually rather impressive memory bandwidth and good parallelism
>(broad buses). On the minus side, they're still a von Neumann design (in
>general terms), and the memory is expensive, hence limited to small
>increments. They've peaked out on power density and power supply (demand
>ramps up higher than power supply can provide, even capacitor-buffered).
The term "von Neumann design" refers to the design outlined in his paper
"First Draft of a Report on the EDVAC."
Nicholas Metropolis (who worked closely with von Neumann) wrote:
>It is clear that the stored-program concept predates von Neumann's
>participation in the EDVAC design. That von Neumann is often given credit
>for this fundamental concept is likely due to the fact that he wrote a
>preliminary report which summarized the earlier work on the EDVAC design,
>including the stored-program concept. Von Neumann contributed
>significantly to the development of this concept, but to credit him with
>its invention is a historical error.
Certainly Eckert and Mauchley ferociously disputed von Neumann's authorship.
My father's version was that my grandfather (Samuel), who had a substantial
role in inventing computers, said von Neumann was always sniffing around,
looking for something to take credit for. The actual smart guys were more
concerned with getting the work accomplished, and ignored him.
The truth is hard to come by. My grandfather was brilliant, tactless, and a
stickler for honesty. My father was the first two, but didn't let the facts
get in the way of telling a good story. I'm scrutinizing Sam's papers for
more insight into what actually transpired.
In any event, the term "EDVAC design" is correct and "von Neumann design"
-- David Lubkin.
More information about the extropy-chat