NSFW
 

Not suitable for work at the bleeding edge, that is. It's markedly more capable than the systems I was using back in 1979-81.

The [URL="http://www.megaprocessor.com/"]Megaprocessor[/URL] project makes me go all soggy with nostalgia.

It was built in the next village to me, Great Shelford rather than Little Shelford where I live, and you can be sure l will be going to the up-coming [URL="https://sites.google.com/site/greatshelfordcommunitywebsite/home/events"]open day[/URL]. Other locals (Tom is the obvious candidate) may wish to consider doing so too.

That is an incredibly interesting project!

You are very fortunate to live close enough to see it in person.

I produced the obligatory "Hello world" program for it. The output options are seriously limited so it produces Morse code on the internal memory's LEDs.

Biggest assembly language program I've written in over thirty years.

[QUOTE=xilman;439233]I produced the obligatory "Hello world" program for it. The output options are seriously limited so it produces Morse code on the internal memory's LEDs.

Biggest assembly language program I've written in over thirty years.[/QUOTE]
:bow: Fascinating stuff!

Totally kewel !

It brings back memories, as in the 1960s when I was in college, we had to design an entire computer from basic components as this machine is. It was the final project of a 1year long required hardware course. Of course, we never actually built our physical machines ( that was for the EE students in year 2 of this course series ). However, it sure was fun and one learned a lot about logic and design.

Looking at the machine, I wonder how many 'race' problems they had with all that looooong wiring connecting the parts. Of course, if it runs slowly, that would mitigate the problem. I remember reading that Seymour Cray spent a large portion of his design time working out race problems.

[QUOTE=tServo;439616]Looking at the machine, I wonder how many 'race' problems they had with all that looooong wiring connecting the parts. Of course, if it runs slowly, that would mitigate the problem.[/QUOTE]"They" is "he". James Newman designed, built and debugged it entirely by himself.

At hsi open day last Saturday, attended only by myself and a chap from ARM whose name I didn't catch, James explained all sorts of design features, including the difficulty of distributing the clock over such a large space. The cabling from one end of the processor to the other is at least 20m (~ 70ns at the speed of light) and so, for example, he had to use master/slave latches in his registers to ensure that timing glitches were mitigated.

If anyone wishes to play with the simulator I'll happily share my code, which now includes a sieve of Eratosthenes. More programs are likely to be wrotten over the next few weeks when (if) round tuits become available again.

That's just incredibly neat- thanks for spreading the word about it.

Norm

[QUOTE=tServo;439616]I remember reading that Seymour Cray spent a large portion of his design time working out race problems.[/QUOTE]

Yeah. I read that he specified the length of wires so that all the signals would arrive exactly when needed.

When asked what design tools he used, he answered something like "fine lined grid paper and sharp pencils".

When he was told that Apple had bought a Cray to help design the next Mac, he replied that he had just bought a Mac to help design the next Cray.

Edit: The above was from memory. Already fully covered at [URL="https://en.wikipedia.org/wiki/Seymour_Cray"]https://en.wikipedia.org/wiki/Seymour_Cray[/URL].

[QUOTE=xilman;439622]"They" is "he". James Newman designed, built and debugged it entirely by himself.

At hsi open day last Saturday, attended only by myself and a chap from ARM whose name I didn't catch, James explained all sorts of design features, including the difficulty of distributing the clock over such a large space. The cabling from one end of the processor to the other is at least 20m (~ 70ns at the speed of light) and so, for example, he had to use master/slave latches in his registers to ensure that timing glitches were mitigated.[/quote]

I'm sorry I didn't make it to the open day - I got back from Hungary at 0415 that morning and was not in a state to do more than walk around town leisurely buying cheese.

The chap from ARM was I think called Artyom; he's in my group, and at the sprint-meeting this morning said that he'd gone to the Megaprocessor event and met 'someone called Paul who said he knew you from grad school' :)

This is pretty damn awesome. I've built simple logic circuits before, but never a CPU or a whole computer.

[QUOTE=Mark Rose;439631]This is pretty damn awesome. I've built simple logic circuits before, but never a CPU or a whole computer.[/QUOTE]

In my final year for my EE we designed and programmed a simple processor (using a Xilinx FPGA). It was definitely daunting, and our designs were only 4-bit (we were limited by the FPGA itself, but I'm glad for that...extending the design to 8-bit would have made us pull our hair out even more).

It was long enough ago that I don't remember much of the details, but I do know we had some basic techniques to avoid race conditions, using gray code to avoid multi bit flipping, etc.

Definitely a fun challenge working with the logic tables, gate designs, bus signalling, etc. And at the end all it did was light up LEDs. I have massive respect for the folks doing actual CPU design... it's incredible.