For those who like to tinker a bit - The new Transputer

Started by vanjast, June 03, 2009, 06:17:08 AM

Previous topic - Next topic

vanjast

Only found this the other day, and I'm happy as a 'pig in sh1t'..

I thought the Transputer was lost forever  :( when Inmos was shut down way back in the early 90s, but it looks like some of the same guys are making a come-back.  :U. Hopefully this time the market impact will be more extensive.

Linky's
http://www.xmos.com/technology
http://www.xlinkers.org/forum/
:8)

PBrennick

Very interesting. I am surprised to see it back. One thing I always hoped would catch on in a big way was the use of touch screens.

Paul
The GeneSys Project is available from:
The Repository or My crappy website

Mark Jones

Interesting indeed. I'm likewise both surprised (and not surprised) that the great Donald Knuth's MMIX is still around... somewhat. No manufacturer is willing to take on such a beast however, as it would simply crush all the competition. :boohoo: (AMD, here's your chance!) :toothy
"To deny our impulses... foolish; to revel in them, chaos." MCJ 2003.08

dedndave

RISC programming has always interested me but, never having a platform
on which to play with it, i don't really know if i would like it or not

i think if they used a carefully thought-out instruction set, with some of the speed enhancements
that have been developed for other processors, it could be very fast and efficient

of course, some of todays enhancements would not be applicable on a RISC machine

i wonder if it would be monotonous writing the same instructions over and over - lol
it seems like a natural way to develop programs that write programs, though
better watch out, or we will be in a "Terminator" situation

vanjast

All the groundwork was done in the late 80's with the Occam language. This latest C dev kit looks exactly the same  :red
This stuff is very much assembler friendly, which what I like - Requires a different way of thinking, but once you get it, it's so simple.

Before the crunch, Inmos was developing the T9000, with it's associated packet switching chips. Had this chip come out it would probably have destroyed Intel, AMD, and Msoft in one foul swoop. Not to mention revolutionise the world network.
It did then only have a small following, but this type of system has 'unlimited potential'. Only after the demise of Inmos have Intel, AMD gone the multicore/multiprocessor route, but they've never been as successful as Inmos's (and now Xmos's) technology.

This is only the beginning, and I reckon in a few years we'll see real nice stuff. Cannot wait for all the 'toys'
:bg