[FoRK] "Golden Girls" architectures

Dr. Ernest Prabhakar drernie at radicalcentrism.org
Mon Jan 13 10:54:24 PST 2014


Hi Dave,

On Jan 12, 2014, at 4:16 AM, Dave Long <dave.long at bluewin.ch> wrote:

> "[FoRK] parallelism Re: Welcome to the era of radical innovation"
> http://xent.com/pipermail/fork/Week-of-Mon-20140106/064534.html
>> I'm starting to think the root of the problem may be the original von Neumann architecture, which was designed to impose serial programming on inherently parallel electronic circuits.
>> 
>> I've been fantasizing about what it would be like to program a "Golden Girls*" architecture? ...
> 
> You may recall that Alain Martin (along with many others!) was working on this sort of thing two decades ago, and two decades before that EWD had similar fantasies:

I’m not surprised.  The more I study computer history, the more I find tantalizing hints of all the roads “not taken”.

I’m still trying to figure out whether:

a) They were just bad ideas, so the solutions we have are really the optimal architectures

b) They were great ideas, but we never pursued them due to cultural/institutional factors

c) They were simply ahead of their time, so we wrote them off early, but should revisit them

But does anyone even attempt to do this sort of techno-historical-anthropology?

— Ernie P.

> 
> "[FoRK] not quite the future from Sleeper"
> http://xent.com/pipermail/fork/Week-of-Mon-20131111/064421.html
>> With the advent of what is called "large scale integration" ...  it seems to become technically feasible to build machines more like "clouds of arithmetic units" with information processing activities going on simultaneously all over the place, for shorter periods of time even independently of each other ...
> 
> 
> Going back another eight decades, a quick flip through Lueger, "Lexikon der Gesamten Technik" reveals that even in the year 46 BT, certain people ("Webereiingenieur") went to a great deal of effort to increase the number of picks per second per operator, or, failing that, at least to pack more threads in a warp ...
> 
> -Dave
> 
> (note that John D. Rockefeller acquired most of his fortune because of making a big early bet that oil would *not* be discovered outside of Pennsylvania.  of course, it eventually was, but much later than everybody else had predicted...)
> 
> _______________________________________________
> FoRK mailing list
> http://xent.com/mailman/listinfo/fork




More information about the FoRK mailing list