Fun fact, I worked on a transputer board for both Amiga and PC way back when. It was doing real time video processing for both machine vision and also broadcast solutions.
For the PC version I wrote the SCSI "drivers" that made the storage talk to transputer card, and also the API that let the transputer board do the video capture and processing for both machine vision and a broadcast capable non-linear video editing system. Just the lower level API, not the NLVE system itself. That was a different team. For the Amiga I did the video capture API, that grabbed frames and fed them to the storage system. Which also let me complete my Bachelor's thesis/final year project. The Amiga design was a lot easier to work with than the PC. I recall the reason for this was that PC was trying to satisfy multiple operating systems; DOS, OS/2 Warp, Windows 3.1, some old version of UNIX, but the Amiga card just supported, IIRC, the AmigaOS.
I forget the name of the company now, "Video something-something?", or "Vision something-something?" they had an office (one of several) in Bristol, UK, though I think the company was London based.
I think they may have been a subsidiary of a much larger entity. This was around 1991/1992, so well in to the Amiga lifespan by then. I had an A2000 with a quad-transputer card (though usually it only had two transputers in it because they were in short supply) on my desk.
In the design, the Transputer sat in front of a chunk of spinning rust storage, and your PC or Amiga talked to the transputer, which talked to the hard drives for you. It had multiple video in & out for camera feeds and a mind-blowing (for the time) amount of RAM per transputer card. 8MB or 16MB I recall. Enough to hold several uncompressed frames of video. Multiple cards would have given you 32MB or 64MB of RAM, which was almost unheard of in anything costing less than $50,000.
The intention was to have a PC or an Amiga do the equivalent of a high-end broadcast capable machine for just 10,000 UKP (in 1990's money), handling, I think, up to eight or 12 video streams simultaneously, doing various video effects and manipulation. Unfortunately the project sold a few units, and was only ever really a glorified demo of "coming soon" but never recouped its development costs and got squished after a little while.
But I look it as that job lead to the opportunity to go do some work for SGI on the Onyx RE, where I eventually fell in love with an over-priced Indigo that still to this day, sits in my garage with its granite keyboard and monitor.
I think the title should be something like "Add-on Transputer boards for the Amiga" because there is no singular "project" in tfa.
fwiw I worked at Inmos during this era and there were many many such add-on boards built for pretty much every kind of computer, typically similar to the reference designs we sold for development purposes on the PC/ISA bus. They all suffered from the same problem which was lack of useful applications. They solved the problem "how do I get some transputers connected to my XXX computer?", but not "how do I do something useful with that?".
I didn't know that transputer were made for the Amiga. Some years ago I did some research about the Atari one and the conclusion was: inefficacy. The serial link linking each transputer being a bottleneck. And more transputers you add, more complexity in tasks distribution you add so the performance curve was flattening.
I'm curious to know if multi-core processors were designed with this lesson in mind.
The Atari Transputer system was unfortunately hampered by a number of design factors. Transputers did have a bottleneck with inter-chip communication, but generally you programmed to the transputer architecture, rather than a general purpose computer architecture. This meant you broke up your data in to chunks, and then let each transputer do its own thing, and then regroup at the end, perhaps using a separate transputer to reassemble the data, much like you would with a modern GPU. You could treat your functions either like a pipeline or like a matrix, but you had to create the algorithm to work with that architecture. Part of my job was showing companies who were deploying transputer based systems how to port their work to this architecture. That said, the transputer to transputer communications bus was pretty optimal, for the time, compared to how most multiprocessor computers handled their work.
I remember back then the Transputer for the Atari ST, always wanted one. It was my first contact with multi processors. As an Amiga owner back then I didn't know there was a transputer project for Amigas.
For the PC version I wrote the SCSI "drivers" that made the storage talk to transputer card, and also the API that let the transputer board do the video capture and processing for both machine vision and a broadcast capable non-linear video editing system. Just the lower level API, not the NLVE system itself. That was a different team. For the Amiga I did the video capture API, that grabbed frames and fed them to the storage system. Which also let me complete my Bachelor's thesis/final year project. The Amiga design was a lot easier to work with than the PC. I recall the reason for this was that PC was trying to satisfy multiple operating systems; DOS, OS/2 Warp, Windows 3.1, some old version of UNIX, but the Amiga card just supported, IIRC, the AmigaOS.
I forget the name of the company now, "Video something-something?", or "Vision something-something?" they had an office (one of several) in Bristol, UK, though I think the company was London based.
I think they may have been a subsidiary of a much larger entity. This was around 1991/1992, so well in to the Amiga lifespan by then. I had an A2000 with a quad-transputer card (though usually it only had two transputers in it because they were in short supply) on my desk.
In the design, the Transputer sat in front of a chunk of spinning rust storage, and your PC or Amiga talked to the transputer, which talked to the hard drives for you. It had multiple video in & out for camera feeds and a mind-blowing (for the time) amount of RAM per transputer card. 8MB or 16MB I recall. Enough to hold several uncompressed frames of video. Multiple cards would have given you 32MB or 64MB of RAM, which was almost unheard of in anything costing less than $50,000.
The intention was to have a PC or an Amiga do the equivalent of a high-end broadcast capable machine for just 10,000 UKP (in 1990's money), handling, I think, up to eight or 12 video streams simultaneously, doing various video effects and manipulation. Unfortunately the project sold a few units, and was only ever really a glorified demo of "coming soon" but never recouped its development costs and got squished after a little while.
But I look it as that job lead to the opportunity to go do some work for SGI on the Onyx RE, where I eventually fell in love with an over-priced Indigo that still to this day, sits in my garage with its granite keyboard and monitor.