XMOS XC-1 kits are shipping

Is the comparison with the Transputer supposed to imply this is a half thought out design with brain dead execution? :-\

Steve

Don't be surprised if people are skeptical.

Remember what happened to the company that made Field-programmable- object arrays?

The transputer was ahead of its time, and really pushed the technology that was available. I sold a lot of systems using it, mostly to universities and research establishments, because there was nothing else around with that sort of performance then. Inmos even had their own fab!

Leon

These are processors, not FPGAs.

Leon

I'm well aware of that, I'm just saying that we shouldn't always believe the hype. I'll believe it when I see it..

I met a guy from the transputer design team down in the Bryce Canyon (here in the US). Unfortunately that company seemed to have lacked marketing savvy just like Plessey and many others. Great products (well, most of them) but that doesn't get you anywhere unless you can nail the next step, the deal. It's probably similar to engineer-driven car manufacturers like Borgward. A seasoned car mechanic here in the US told me that these were among the most well designed and quality built cars ever. But ...

Of course I secured a small stash of Plessey SL6440 mixers before it all imploded.

Yes at the time it was way ahead. Trouble is that there was not many people who had given thought to paralel computing back then - even now it is experimental on the whole and not mainstream. Another machine of the time was made by Linn Products who make Hi-Fi..

It was a small-talk machine called Recusiv - fully Object Orientated

- again way ahead of its time. People had just started using C and were still using Fortran. Assembler was the only low level language available.

Hardy

After they married, my brother used his wife's Borgward Isabella. It wasn't bad, apart from the slippery bench seat in the front. I didn't close the door properly once, slid across the seat when he cornered rather fast and nearly fell out of the door when it opened. Cars didn't have seat belts in those days, of course.

Leon

That had a purpose back then. Not exactly one that had to do with driving and it was only useful if you had a girlfriend in the car ;-)

On my pa's old Chrysler the whole frame would torque and warp a little at high speeds. At around 100mph (in Germany where that was legal) the driver side door made *BANG* and popped open one notch. Later it turned out some really clever engineer had positioned the battery right above a frame member. The occasional acid drips from there slowly ate it up.

Looks interesting but doesn't look to be a number-crunching device. It's general purpose for Ethernet type applications.

Hardy

I remember the Rekursiv. I even asked them if they would sell me some chips. They would only sell me one of their boards, which were rather expensive.

Leon

Look at it as a very fast and cheap 32-bit scalable processor.

Leon

I couldn't find a datasheet on the XMOS CPU. What's its architecture? Does it have 16 bit datapaths? What size is the multiplier/ALU? Can it do, e.g., 1600 32x32->80 fixed-point multiplies per second?

and slippin' and

Record*, ELO

It's got a 32 x 32 bit -> 64-bit MAC

Leon

...

There were only ever about 30 chipsets (from LSI Logic); if one had been sold individually it would have left a board empty.

In any case, a prototype board for two 299 pin PGAs (at the time, LSI said, the largest in Europe) and a 223 pin would have been decidedly non-trivial.

The price of £25k for a board was regrettable, but without a much larger company behind it, there wasn't going to be any other option.

A second version would have been far more cost-effective but never made it past the initial design stages.

- Brian

The Transputer wasn't ahead of its time. It was brain dead. A chip only effective in substantial arrays selling for hundreds of pounds per device was a dead duck from the start. The only people who could seriously look at it for substantial arrays were military applications. However, when approached about military parts Transputer gave evasive answers.

There was nothing innovative about the design of the Transputer. The device as it was supposed to be (i.e. separate comms and execution planes), rather than the crippled one they shipped, was similar to designs several people in the UK (and presumably elsewhere) were toying with at that time. The others did not proceed because the economics looked so wrong.

It was used very widely ar the time, because there wasn't anything comparable for sale. I even sold several of my modules to hobbyists and university students - at =A3500 each!

Some of the customers for my 16 module system were prestigious outfits such as BAe, Plessey Roke Manor, GCHQ and Oxford University PRG. It delivered 320 MIPS with very good floating-point performance at a comparatively low price (about =A313,000 with T800 chips).

I even had inquiries from Russia, indirectly, but it was embargoed because of the performance. I did once apply for an export license for Russia, to see what happened, and actually got it! I contacted the DTI about it; they got very excited, admitted thay had made a mistake and insisted that I returned the document. I'm quite sure that Russia did get the technology, though. I heard on the grapevine that there were companies in Finland that acted as intermediaries.

Leon

I briefly saw XMOS at a trade show in London. They seem far too fixated on doing everything in software, things like ethernet where there is no point shoveling bytes in software if hardware can take care of it.

The quarter VGA touchscreen on the dev kit has novelty value.

I'm not sure the performance justifies the effort of using somthing so unusual in most cases.

I'd want to make a single lifetime buy of processors if I used these in a product. Far too likely to dissapear without trace.

Bob