has anybody heard of INMOS Transputer?

[savan]

INMOS was one of the widely accepted transputer of 1990s!
It is one of the successful processors which used Stack Processor!
I was wondering what made it s much accepted?
Was it just because of its Parallel Processing?
Has its promotion anything to do with Stack Processor in it unlike other whose computation is based on registers!
I found one of the following video but this highlights just its parallel processing!
https://www.youtube.com/watch?v=o7byWAv6qCk


Kind Regards
Savan Vachhani
Research Assistant
Advanced Computer Architecture Group
Department of Computer Science
The University of York
+447508198897
Website

 

[rohitkhanna]

No. It had many great features, but also several bad features (like price & complexity & less support).

To be commercially viable & successful... ALL ass pectneed to be at least 'good', if not great.
Thats true for any product

cheers!

 

[betwixt]

I used to live near the factory that made them and used them briefly about 15 years ago. They were processors based on good ideas but poor marketing and an obscure programming language, Occam I think it was called. Most other manufacturers managed to implement better versions of their advanced features.

Brian.

 

[pjmelect]

The great advantage of the transputer was its very simple hardware and good performance, its main disadvantage and the reason why we are not using its descendants today was its price, very expensive, they killed it by charging to much money for it.

 

[betwixt]

And building a fabrication plant beside a major highway (M4 in South Wales) where ground vibration cause manufacturing problems. Although the Transputer has almost vanished, it was the forerunner of what we now regard as RISC techmology.

Brian.

 

[savan]

Hello Brian!
I am currently working on stack machines and main concern is about Transputer using Stack based computing!
Was using Stack advantage for transputer?
I am unable to extract much detail about transputer and Stack based computing!
I guess people have stopped working on stack after 1990s!

Kind Regards
Savan Vachhani
Research Assistant
Advanced Computer Architecture Group
Department of Computer Science
The University of York
+447508198897
http://www.savan.in

 

[betwixt]

Unfortunately when I worked with Transputers is was verifying operation of finished equipment rather than on development so I didn't have involbvement in software development.

As I see it, the concept of stack computing was based on the methods proposed by Alan Turing in the 1940's where he visualized a computer as being an infinite line of data which could be traversed left or right based on the instructions being executed. Where he visualized a horizontal paper tape, I see the Transputer using a vertical silicon stack instead. I think in the early days of LSI it was probably more efficient to use a stack for all computation and keep slower 'normal' memory for storage. These days, the speed of registers and memory is much faster and the advantage of a fast stack has been lost, especially as programming in registers is so much simpler.

Do some research on Alan Turing and see if you agree. Most of what you will find is about his personal life but there are references to his technical achievements as well.


Brian.

 

[savan]

I will check out Turing's Technical achievement and give a thought! Are stacks really worth using?

Many Thanks and Kind Regards
Savan Vachhani
Research Assistant
Advanced Computer Architecture Group
Department of Computer Science
The University of York
+447508198897
http://www.savan.in