[comp.lang.forth] Forth Engines

ZMLEB@SCFVM.BITNET (Lee Brotzman) (08/06/88)

In article <8808051447.AA00873@rutgers.edu> twitch!grt@att.att.com writes:
>
>You mention Charles Moore's claim that the 68000 subroutine call and
>return take 20 clocks and you took that with a grain of salt.  I just
>looked up JSR in Motorola's MC68000 8-/16-/32-Bit Mircrprcessors
>Programmer's Reference Manual.  The JSR takes 16 to 22 clocks, depending
>on the adressing mode.  The quickest is address register indirect;
>absolute long takes 20.
>---
>        George Tomasevich, att!twitch!grt
>        AT&T Bell Laboratories, Holmdel, NJ

It's good to see the claim is true.  Not being a 68000 programmer, and since
Charles Moore *is* the inventor of a competing product, I wanted to give
the 68000 the benefit of the doubt.

It seems to me that even if Forth engines aren't the wave of the future,
manufacturers of traditional computer processors can learn a lot from the
designs of the Forth chips on the market.  The Forth community has been
rocking the "more is better" boat for years, and the application of
the Forth philosophy to hardware goes a long way toward showing that we
were right.

Now another question:  Just because a Forth chip has Forth as its assembly
language, does that mean that one has to program it in Forth?  I have
heard that someone was writing a C compiler for the NOVIX chip.  Is this
true?

I am sure that the resulting code is more efficient if written in native
Forth "assembly" for these chips, just like code written in traditional
assembly is more efficient than executables from compiled high-level
languages.

Are we going to see more Forth programmers because of the new hardware,
or simply faster C (Pascal, FORTRAN, etc.) compilers?  I wonder...

-- Lee Brotzman (FIGI-L Moderator)
-- My opinions are my own and not those of my employer, ST Systems Corp.,
-- or their employer, NASA Goddard Space Flight Center, or their employer,
-- Ronald Reagan, or his employer, Nancy Reagan, or her employer, the
-- planets and stars.

web@garnet.berkeley.edu (William Baxter) (11/18/88)

If you run a Forth Engine, I would like to receive your comments:

What hardware do you have, and how much did it cost?
What type of programming do you do?
What benefits does the Forth Engine provide?
What limitations have you encountered?
What else should I know when considering buying such hardware?

Please email your responses.  I will summarize to the net. 

William Baxter

ARPA: web@{garnet,brahms,math}.Berkeley.EDU   
UUCP: {sun,dual,decwrl,decvax,hplabs,...}!ucbvax!{garnet,brahms,math}!web