home .. forth .. misc mail list archive ..

RE: Support for UT


This is all good fun, but MISC may need to evolve
into a new name with the low fat computing idea
built in, since Minimal Instruction Set isn't 100%
of what UT's output is about.

John Griessen
Austin TX

> -----Original Message-----
> From: Keith Wootten [mailto:Keith@wootten.demon.co.uk]
> Sent: Thursday, March 18, 1999 4:58 AM
> To: MISC@pisa.rockefeller.edu
> Subject: Re: Support for UT
> 
> 
> From: Luis Commins 
> 
> [ a teeshirt ]
>   (Maybe with the slogan: 'Don't RISC it, MISC it'?)
> 
> Reply from: Keith Wootten
> 
> I'll buy one, they sound too good to MISC.
> 
> 
> 
> From: Luis Commins
> 
>   Looks like if we want some FORTH processors, DYOP is better,
>   but then, it won't bring either performance or price.
> 
> From: vic plichota
> 
>   I really *hate* to say this, but for a verrry long time, the only
>   satisfactory effort that I've ever seen has been the Harris RTX,
>   period.
> 
> 
> Reply from: Keith Wootten
> 
> Are you unaware of the Patriot Scientific PSC1000 (ShBoom) processor?
> It's available, (I use it with a simple Forth cross-compiler built with
> Win32Forth) and has many MISC features.  Its forerunner was designed by
> Chuck Moore and, while I don't know the full story, it looks as if
> someone took over the design and decided to make it work in a less
> quirky way.  It's a 32bit twin-stack (on-chip) machine which currently
> runs at 100MHz loading four 8 bit instructions at a time.  The 10ns
> instructions include *many* Forth primitives and it runs Forth very
> quickly.  In addition there is a set of general purpose registers, and
> the return stack can also be accessed as registers if required.  The
> memory interface is very flexible, and there is also a simple on-chip
> I/O processor with priority for time-critical tasks.  The device can
> also run (more slowly) with 8 bit memory, and a minimum system comprises
> uP, ROM, DRAM and Oscillator - no glue required. 
> 
> While I agree that we should all support Jeff - and you can count me in
> for Teeshirts or Newsletters or whatever - it seems to me that, if the
> PSC1000 were more widely used, there would be a smaller mental leap for
> potential MISC investors.  The FPGA approach is interesting, but a
> PSC1000 costs $10 in quantity - couple this with a small FPGA to handle
> the peripherals, and you have one hell of a system.  It's certainly not
> as powerful as MISC will hopefully be, but it's real, working,
> documented Silicon which can be bought now.  Use it or - maybe - lose
> it.
> 
> While supporting the development of what we would all like to see, let's
> use what we already have in real products and projects.  I don't have
> any connection with Patriot Scientific, but I've used the RTX2001
> extensively, and know a good Forth chip when I see one.
> 
> See www.ptsc.com for details of ShBoom,  Forth inc have a commercial
> compiler and show some very interesting benchmarks on their WebSite. 
> 
> Cheers
> -- 
> Keith Wootten
>