From ewt@SunSITE.unc.edu Mon Jun 28 17:37:47 1993
Return-Path: <ewt@SunSITE.unc.edu>
Received: from gossip.pyramid.com by dutecai.et.tudelft.nl (4.1/1.34JP)
          id AA26136; Mon, 28 Jun 93 17:37:41 +0200
Received: from sword.eng.pyramid.com 
	by gossip.pyramid.com (5.61/OSx5.1a Pyramid-Internet-Gateway)
	id AA22765; Mon, 28 Jun 93 08:37:30 -0700
Received: from goss.pyramid.com
	by sword.eng.pyramid.com (5.61/Pyramid_Internal_Configuration)
	id AA10379; Mon, 28 Jun 93 08:36:20 -0700
Received: from SunSite.unc.edu 
	by gossip.pyramid.com (5.61/OSx5.1a Pyramid-Internet-Gateway)
	id AA22542; Mon, 28 Jun 93 08:36:18 -0700
Received: by SunSITE.unc.edu (4.1/tas-gen/1-30-93)
	id AA01523; Mon, 28 Jun 93 11:35:06 EDT
Date: Mon, 28 Jun 1993 11:35:05 -0400 (EDT)
From: Erik Troan <ewt@SunSite.unc.edu>
Subject: 34010 X compiler, mailing lists
To: riscy@pyramid.com
Message-Id: <Pine.2.4.53.9306281125.A1445@calypso>
Status: RO
X-Status: 


On xxx, 27 Jun 1993, Pat Mackinlay wrote:

[lots of good stuff about 34010's deleted becuase we've all read it already

> BTW: I wouldn't expect that using a 34010 would make the graphics
> system outperform any of the VGAs out there, but it'd be quite
> interesting, and would leave the main CPU free to do some "real"
> work. It might even be practical to run PEX stuff <grin>

If we do decide on the 34010 path, I agree with Pat that a gcc port would
be a huge problem. I'm willing to write a small C compiler for the chip
from scratch. I'm not offering to port gcc, as I think that such an effort
is unnecessary. If the chip is going to be driving graphics, code doesn't
have to be written often, so I think we can live with putting the 
optimizations in the code instead of the compiler.

By small I mean I'm not going to write a floating point library (though
I'd be happy to integrate with one written by someone else), generate
anything but straight object code for the 34010 (we don't need dll's,
etc for a graphics processor), or even acknowledge C++'s existence any
further then allowing // for comments. Let me know if this is worth
my time.

> [...if there is no archive site, would someone be willing to mail me
> the last day or two of messages? I don't like having to repeat questions
> that people have already answered...]

If there is an archive site, would someone please let me know. If not, I'd
like to set one up on sunsite.unc.edu. /pub/Riscy maybe? Seems like we
should start thinking about an ftp site anyhow.

Erik

 
From PET@informatik.ee.edag.de  Ukn Jun 29 04:13:57 1993
Received: from iraun1.ira.uka.de by SunSITE.unc.edu (4.1/tas-gen/1-30-93)
	id AA02759; Tue, 29 Jun 93 04:13:52 EDT
Received: from xlink1.xlink.net by iraun1.ira.uka.de with SMTP (PP) 
          id <21535-0@iraun1.ira.uka.de>; Tue, 29 Jun 1993 10:13:10 +0200
Received: from edfd by xlink1.xlink.net id aa21561; 29 Jun 93 10:12 MET DST
Received: from charon (informatik) by edfd.edag.de with SMTP 
          id AA22077 (5.65c/IDA-1.5 for <ewt@SunSite.unc.edu>);
          Tue, 29 Jun 1993 09:04:15 +0200
Received: From INFORMATIK/WORKQUEUE by charon via Charon-4.0A-VROOM with IPX 
          id 100.930629085953.448; 29 Jun 93 08:54:41 +0100
Message-Id: <MAILQUEUE-101.930629085950.416@informatik.ee.edag.de>
To: SunSite.unc.edu!ewt@informatik.ee.edag.de,
        pyramid.com!riscy@informatik.ee.edag.de
From: peter katzmann <PET@informatik.ee.edag.de>
Date: 29 Jun 93 08:59:50 CET
Subject: Re: 34010 X compiler, mailing lists
Priority: normal
X-Mailer: Pegasus Mail v2.3 (R5).
Status: RO
X-Status: 

>
> On xxx, 27 Jun 1993, Pat Mackinlay wrote:
>
> [lots of good stuff about 34010's deleted becuase we've all read it already
>
> > BTW: I wouldn't expect that using a 34010 would make the graphics
> > system outperform any of the VGAs out there, but it'd be quite
> > interesting, and would leave the main CPU free to do some "real"
> > work. It might even be practical to run PEX stuff <grin>
It will be faster then the normal VGA solution's. We wrote a CAD
package for an PC with TIGA or VGA and the TIGA outperformes the VGA
of about 70% and it wasn't the bist design.
>
> If we do decide on the 34010 path, I agree with Pat that a gcc port would
> be a huge problem. I'm willing to write a small C compiler for the chip
> from scratch. I'm not offering to port gcc, as I think that such an effort
> is unnecessary. If the chip is going to be driving graphics, code doesn't
> have to be written often, so I think we can live with putting the
> optimizations in the code instead of the compiler.
>
You don't have to wrote the Compiler from the scratch. There's ans
old C projekt from Dr Dobb's when i'm think right. I have the source
code here for K&R or Turbo C. Several company's rewrote the generator
part for their different single Chip CPU's.

> By small I mean I'm not going to write a floating point library (though
> I'd be happy to integrate with one written by someone else), generate
> anything but straight object code for the 34010 (we don't need dll's,
> etc for a graphics processor), or even acknowledge C++'s existence any
> further then allowing // for comments. Let me know if this is worth
> my time.
If it's necessary i could do it, but not one that's IEEE conform.


----

EDAG-Fulda | P. Katzmann | E-Planung | Steinauer Str.| 6400 Fulda
tel. 0661/6000-454/464 | W. Germany | in: pet@informatik.ee.edag.de

 
From PET@informatik.ee.edag.de  Ukn Jun 29 04:19:50 1993
Received: from gossip.pyramid.com by SunSITE.unc.edu (4.1/tas-gen/1-30-93)
	id AA03078; Tue, 29 Jun 93 04:19:49 EDT
Received: from sword.eng.pyramid.com 
	by gossip.pyramid.com (5.61/OSx5.1a Pyramid-Internet-Gateway)
	id AA28774; Tue, 29 Jun 93 01:14:40 -0700
Received: from goss.pyramid.com
	by sword.eng.pyramid.com (5.61/Pyramid_Internal_Configuration)
	id AA04515; Tue, 29 Jun 93 01:14:40 -0700
Received: from iraun1.ira.uka.de 
	by gossip.pyramid.com (5.61/OSx5.1a Pyramid-Internet-Gateway)
	id AA28770; Tue, 29 Jun 93 01:14:31 -0700
Received: from xlink1.xlink.net by iraun1.ira.uka.de with SMTP (PP) 
          id <21537-0@iraun1.ira.uka.de>; Tue, 29 Jun 1993 10:13:10 +0200
Received: from edfd by xlink1.xlink.net id aa21566; 29 Jun 93 10:12 MET DST
Received: from charon (informatik) by edfd.edag.de with SMTP 
          id AA22077 (5.65c/IDA-1.5 for <riscy@pyramid.com>);
          Tue, 29 Jun 1993 09:04:15 +0200
Received: From INFORMATIK/WORKQUEUE by charon via Charon-4.0A-VROOM with IPX 
          id 100.930629085953.448; 29 Jun 93 08:54:41 +0100
Message-Id: <MAILQUEUE-101.930629085950.416@informatik.ee.edag.de>
To: SunSite.unc.edu!ewt@informatik.ee.edag.de,
        pyramid.com!riscy@informatik.ee.edag.de
From: peter katzmann <PET@informatik.ee.edag.de>
Date: 29 Jun 93 08:59:50 CET
Subject: Re: 34010 X compiler, mailing lists
Priority: normal
X-Mailer: Pegasus Mail v2.3 (R5).
Status: RO
X-Status: 

>
> On xxx, 27 Jun 1993, Pat Mackinlay wrote:
>
> [lots of good stuff about 34010's deleted becuase we've all read it already
>
> > BTW: I wouldn't expect that using a 34010 would make the graphics
> > system outperform any of the VGAs out there, but it'd be quite
> > interesting, and would leave the main CPU free to do some "real"
> > work. It might even be practical to run PEX stuff <grin>
It will be faster then the normal VGA solution's. We wrote a CAD
package for an PC with TIGA or VGA and the TIGA outperformes the VGA
of about 70% and it wasn't the bist design.
>
> If we do decide on the 34010 path, I agree with Pat that a gcc port would
> be a huge problem. I'm willing to write a small C compiler for the chip
> from scratch. I'm not offering to port gcc, as I think that such an effort
> is unnecessary. If the chip is going to be driving graphics, code doesn't
> have to be written often, so I think we can live with putting the
> optimizations in the code instead of the compiler.
>
You don't have to wrote the Compiler from the scratch. There's ans
old C projekt from Dr Dobb's when i'm think right. I have the source
code here for K&R or Turbo C. Several company's rewrote the generator
part for their different single Chip CPU's.

> By small I mean I'm not going to write a floating point library (though
> I'd be happy to integrate with one written by someone else), generate
> anything but straight object code for the 34010 (we don't need dll's,
> etc for a graphics processor), or even acknowledge C++'s existence any
> further then allowing // for comments. Let me know if this is worth
> my time.
If it's necessary i could do it, but not one that's IEEE conform.


----

EDAG-Fulda | P. Katzmann | E-Planung | Steinauer Str.| 6400 Fulda
tel. 0661/6000-454/464 | W. Germany | in: pet@informatik.ee.edag.de

 
From jeremy@sw.oz.au  Ukn Jun 29 00:33:04 1993
Received: from munnari.oz.au by SunSITE.unc.edu (4.1/tas-gen/1-30-93)
	id AA24381; Tue, 29 Jun 93 00:32:37 EDT
Received: from sw.oz.au (via basser.cs.su.oz.au) by munnari.oz.au with MHSnet (5.83--+1.3.1+0.50)
	id AA09884; Tue, 29 Jun 1993 14:32:07 +1000 (from jeremy@sw.oz.au)
Received: from chao.sw.oz.au by swift.sw.oz.au with SMTP
	id AA00541; Tue, 29 Jun 93 14:09:39 AES (5.59)
	(from jeremy@sw.oz.au for ewt%SunSite.unc.edu@munnari.oz.au)
Received: by chao.sw.oz.au (4.1/SMI-4.1)
	id AA14493; Tue, 29 Jun 93 14:09:59 EST
From: jeremy@sw.oz.au (Jeremy Fitzhardinge)
Message-Id: <9306290409.AA14493@chao.sw.oz.au>
Subject: Re: 34010 X compiler, mailing lists
To: ewt@SunSite.unc.edu (Erik Troan)
Date: Tue, 29 Jun 1993 14:09:58 +1000 (EST)
In-Reply-To: <Pine.2.4.53.9306281125.A1445@calypso> from "Erik Troan" at Jun 28, 93 11:35:05 am
Organization: Softway Pty Ltd
X-Face: 
	 '6U=%Tv\k1<Ek%ql%PN^v`Db4bakr[v~y]\u7"GbO#I=]N{l1=#P,glz$9q>l-:?\$C[D@G
	 7(vl~w8&y}!f\bh#w<Y*S~bEBTI:s&.QR>L#n,TGKh>T.c7eT5-y)Hl'i;A1z$9?*lD.k}yqshddFb
	 l[EC}c=;uc%x'}uh3E91p&oE<q$w1r&U0yw.Sb3V&uw
X-Mailer: ELM [version 2.4 PL21]
Content-Type: text
Content-Length: 1570      
Status: RO
X-Status: 

Erik Troan bubbles:

> On xxx, 27 Jun 1993, Pat Mackinlay wrote:

> [lots of good stuff about 34010's deleted becuase we've all read it already

> > BTW: I wouldn't expect that using a 34010 would make the graphics
> > system outperform any of the VGAs out there, but it'd be quite
> > interesting, and would leave the main CPU free to do some "real"
> > work. It might even be practical to run PEX stuff <grin>

> If we do decide on the 34010 path, I agree with Pat that a gcc port would
> be a huge problem. I'm willing to write a small C compiler for the chip
> from scratch. I'm not offering to port gcc, as I think that such an effort
> is unnecessary. If the chip is going to be driving graphics, code doesn't
> have to be written often, so I think we can live with putting the 
> optimizations in the code instead of the compiler.

> By small I mean I'm not going to write a floating point library (though
> I'd be happy to integrate with one written by someone else), generate
> anything but straight object code for the 34010 (we don't need dll's,
> etc for a graphics processor), or even acknowledge C++'s existence any
> further then allowing // for comments. Let me know if this is worth
> my time.

There are at least 2 cheap or free 34k C compilers I know of.  One was
written by a friend of mind.  If people get really keen about a 34010
(curse the day), I'll ask him for terms.  It is adapted from a 68k
compiler, and it has a number of simple machine independant optimisations.
It generates assembly output.  I can also get an assembler and linker.

	J

 
