Unidentified subject!

[Wayne Morellini <waynem1@xxxxxxxxxxxxxxxxx>]

Forwarded message:
>From MISC-request@pisa.rockefeller.edu Wed Feb  7 07:00:28 1996
>Resent-Date: Tue, 6 Feb 1996 15:54:12 -0500
>Posted-Date: Tue, 6 Feb 96 14:57:06 CST
>Date: Tue, 6 Feb 96 14:57:06 CST
>From: lowry@src.honeywell.com (Dave Lowry)
>Message-Id: <9602062057.AA29532@terpsichore.src.honeywell.com>
>To: MISC@pisa.rockefeller.edu
>Resent-Message-ID: <"abpHC2.0.r31.n_x5n"@pisa>
>Resent-From: MISC@pisa.rockefeller.edu
>Subject: Unidentified subject!
>X-Mailing-List: <MISC@pisa.rockefeller.edu> archive/latest/336
>X-Loop: MISC@pisa.rockefeller.edu
>Precedence: list
>Resent-Sender: MISC-request@pisa.rockefeller.edu
>
>MISC Readers-
>
>On the subject of compilers for X21 chips...
>
>Does anyone know if the rumored commercial C compiler is a reality?
>Supposedly MPE in the UK have a C compiler for stack architectures,
>but I don't know if they have been paid to do an X21 port.  I suspect,
>anyway, the the cost would be beyond most MISCer's budgets.

At lst that I checked it was not (last year), I did suggest that some Misc
enthusiaste might like to do the port in exchange for royalities, but to no
avail.

>
>I've done a little work porting both Small-C v2.0 and Dunfield C to X21.  Both
>compilers assume a traditional register set architecture that maps poorly
>onto X21.  The compiled code would be many, many times slower than
>native X21 code.  It's not clear to me that having a C compiler that
>produces slow X21 code is in the best interests of MISC.
>

Yes, very, as long as you can gaurantee descent C code speed for the
compiled code.  You see there is a wealth of C code out there in programs
that support variouse hardware and standards (the type of components I have
been talking about).  So if somebody wanted to build a board that offered
compatability with an existing standard, support a hardware interface etc,
they could rewrite the software in Forth or licence a C version and compile
it thus saving time and money and comming to the market early.

Snip

>-Dave
>lowry@src.honeywell.com
>( not speaking for Honeywell, etc. )
>

Wayne
waynem1@cq-pan.cqu.edu.au

--