Discussion:
transputing without transputing
(too old to reply)
Derek Simmons
2006-06-23 20:02:49 UTC
Permalink
Interestingly, Aspex Semiconductor produce a massively-parallel
computation engine, and their preferred method for programming is the
patch.
It's not particularly transputer-like, though.
(http://www.aspex-semi.com/pages/technology/technology_asprocore.shtml)
I did go out and take a look at this...a long look. It seems to be more
like a Thinking Machine Connection Machine 2a on a chip than a
transputer. I wonder if anybody ever told Danny Hillis. I wonder if you
could program it using CmLisp?

Derek
August West
2006-06-24 15:22:30 UTC
Permalink
Post by Derek Simmons
Interestingly, Aspex Semiconductor produce a massively-parallel
computation engine, and their preferred method for programming is the
patch.
It's not particularly transputer-like, though.
(http://www.aspex-semi.com/pages/technology/technology_asprocore.shtml)
I did go out and take a look at this...a long look. It seems to be
more like a Thinking Machine Connection Machine 2a on a chip than a
transputer.
Yes, that's pretty much what is is (to a first approximation). The
nature of the bit-pattern-matching activation gives it a rather
different flavour, though - possibly more ICL DAP-like?
Post by Derek Simmons
I wonder if anybody ever told Danny Hillis. I wonder if you
could program it using CmLisp?
You could possible try. what it really needs, in my opinion was CM's
C* (or Data Parallel C). Not everyone shares that opinion, however.
--
I can't complain but sometimes I still do
Loading...