Interesting reading about Setun

[Mac Buster]

The ones who interested in history of ternary computing can read the following web-page - The creator of ternary computer.

[Mac Buster]

And yet another interesting document - Setun's reflection. The story looks incomplete, but there are some really rare photos and events

[jay4th]

I've read these articles with great interest. I would love to see the instruction set of the SETUN and the Setun-70 (especially the 70). I have 'read' descriptions in Googlish (google translated Russian) but would be much happier to read explanations that were clearer and in English, or a compilable language like C, Java, etc. The 'description' on the online simulator is completely impenetrable for me, though some of the Googlish (Original Russian) is almost understandable.

[Mac Buster]

I would love to see the instruction set of the SETUN and the Setun-70 (especially the 70).

The difference between Setun's and Setun-70's instructions set is really small, only one or two rarely used instructions of Setun were modified in Setun-70. As result the machine became stack-based (actually two stacks - command and data) and instead of using 24 base hardwired instruction programmer can use up to 81 new instruction: 27 base, 27 system (for I/O, etc) and 27 user defined. Sadly there are no implementation of these command sets in modern languages

But I have many docs and articles about these machines in russian language and will try to find and translate them into english for you. As far as I remember some article contain description and kind of simulation of Setun-70 architecture with examples written in Algol, so this might be usefull.

[Shaos]

The 'description' on the online simulator is completely impenetrable for me, though some of the Googlish (Original Russian) is almost understandable.

Do you mean this simulator?
http://en.trinary.ru/projects/setunws/

[jay4th]

Yes that one. For example, there is no hint on the pages there about how trit #9 works in calculating an effective address. The googlish I read made it clear that trit #9 does something like:

effective address = address + trit9*some-register

So address + some-register
or address - some-register
or address

is the effective address. Cool idea, nice clean way to encode an instruction set (with a minor issue). But not obvious from any English documents. (I hate the way national languages isolate people.)

Oh, the minor issue: The contents of a register are naturally stored in trinary, so naturally already support negative numbers... so seems like there is a bit of wasted effort. Both the register, and bit 9 can call for reverse indexing. And it introduces concerns about double-reversals. This is an example where binary might make more sense, or the third value could be used for something else like maybe

(if trit9 == 0)
effective address = address
(if trit9 == 1)
effective address = address + some_register
(if trit9 == -1)
effective address = address + 3*some_register

Anyway, I know I could learn a lot by looking over their instruction set. If only I could find something I could read and understand.

OK, who out there is working on a re-implementation of Setun or Setun 70?

[jay4th]

The difference between Setun's and Setun-70's instructions set is really small, only one or two rarely used instructions of Setun were modified in Setun-70.

Wow, I totally didn't understand that from the googlish I read. I got the impression that Setun-70 was not an 9/18 trit machine, but was a 6/12/18 trit machine. So a very different instruction set, word size, etc. I do remember the bit about "just changed a few instructions" but I didn't think that was Setun-70 vs Setun, but rather Setun-70 V1.0 compared to Setun-70 V2.0. But I was reading old technical Russian through a computer 'translation' and am *certain* I missed and mis-understood parts.

[Mac Buster]

OK, who out there is working on a re-implementation of Setun or Setun 70?

As far as I know there are no such projects now. But at least one was in past - http://pyramid.codeplex.com/ (no idea why I think this is all about original Setun, probably been reading in someone's blog).

If you wish to start such a project, I'd like to take part in in, at least I can suppy some docs and articles about the machine.

[Mac Buster]

I do remember the bit about "just changed a few instructions" but I didn't think that was Setun-70 vs Setun, but rather Setun-70 V1.0 compared to Setun-70 V2.0.

It could be that you're right here, I'll re-read docs. Sorry.

[jay4th]

Building a Setun reimplementation, or Setun-70 would be fun. I've been playing with logisim, maybe I'll start by designing a implementation/simulation using that tool. It's the right price for me -- free! New thread?

[Shaos]

Building a Setun reimplementation, or Setun-70 would be fun. I've been playing with logisim, maybe I'll start by designing a implementation/simulation using that tool. It's the right price for me -- free! New thread?

You may create new topic in "Ternary Hardware" forum:
http://www.ternary.info/modules/newbb/v ... hp?forum=5

[Shaos]

Forums on ternary.info are obsolete now, so new ternary topics in English must be created here:

viewforum.php?f=62