Summary: C compilers for DEC 2060, TOPS-20

[The Polymath]

About two weeks ago I posted a request for information on C compilers  that
run  on  a  DEC  2060  under TOPS-20.  At present we have only the Utah PCC
which we do not consider  satisfactory  for  our  purposes.  Below  is  the
summary  of responses I've received to my query.  My thanks to all who took
the time to contribute.

==============================================================================

I've had the pleasure of working with only one TOPS-20 implementation
of C, and it's probably the undocumented, unsupported one you alluded
to in your posting.  This compiler compiles your C code into yet-
another-PDP-10-assembly-language (MIDAS) and then invokes a non-DEC
assembler to produce non-DEC object modules, which are then linked by
a non-DEC loader (appropriately named STINKR) to produce a compatible
TOPS-20 .EXE file.  Sound at all familiar?

Anyway, if you get any leads on a production-quality C compiler for
the -20, I'd be interested in hearing about it!

------------------------------------------------------------------------------

You can check with Tartan Labs.  They will release one if there is enough
demand.  You can talk to Jackie Kelly, (412) 621-2210, or write to
	Tartan Labs Inc.
	477 Melwood Ave.
	Pitt, PA 15213

[NOTE: I did check with Tartan Labs.  They told me they only had C compilers
to run on VAX, not on DEC 2060's.  --JH ]

------------------------------------------------------------------------------

The ONLY C compiler I know of for tops20 is the port of the PCC
compiler.  It's slow, inconvenient to use, generates bad code,
incomplete, and a number of other things.  I suppose it just never
was a high priority item, it being much cheaper to buy vaxen and
small micros and get C free than port to a million++ dollar machine.

I don't know where to get it, but I can track it down if you like (if
that's not what you already have).

------------------------------------------------------------------------------

Here at Columbia University  CUCCA  (Columbia University Center for Computing
Activities) has 4 dec 20s which have a C compiler that is definitely documented
and supported. I have not used it so I am unable to give an opiniion as to 
quality, however CUCCA usually puts out high quality products.

I am not sure who you should contact at CUUCA, but their main # is
(212) 280 -3703.

[Unfortunately, not knowing who to talk to caused a problem here.  I called,
but the switchboard operator had no idea who to refer me to, so I had to let
it go.  --JH]

------------------------------------------------------------------------------

There are at least 3 C compilers for the 20's of which I am aware.

The first is a port of Alan Snyder's original Portable C compiler, and
is (was) available from MIT.  You need a valid UNIX license to get this.

The second is a port of the Bell Labs PCC which showed up first in the
7th Edition of UNIX and UNIX/32V.  This is available from the
University of Utah and comes with a library of routines which contain
many of the subroutines in sections 2 & 3 of Volume I of the UPM.
The compiler is a credible one in that it successfully compiles itself
and other other reasonably complex programs (such as yacc).  However,
it does things like use 36 bits for char's and does not (as far as I
know) support extended addressing.  I believe e Randy Frank at Utah's
CS department is the contact point.  You need a UNIX license from AT&T
and I think you also need a 4.1 license.

The third is a new compiler which is being developed at Stanford, it's
called KCC (for Kok Chen C, Kok Chen being its original author).  This
compiler is the one which I think will eventually be the C of choice
for the 20's, it produces good code, supports extended addressing,
is fast.  It is still in development, though, and although a few
people have it to play with, the bugs are still being shaken out.
It too has a library of UPM calls, and in addition a UNIX compatibility
environment (like the TOPS-10 compatibility package) is being developed
which will run with it.  This latter item is something I'm doing, it's
about 90% done but work goes slowly as it's something that I work on
with my own time, something I haven't had a lot of lately.  This
package is called PAUNIX (play on PA1050).  I'm not certain of the
availability of KCC, you should talk to Greg Satz at Stanford's EE
department about it.  PAUNIX will be made available to the public
either free or at cost of distribution when it is complete.  

------------------------------------------------------------------------------

You didn't say *which* TOPS-20 C compiler you already have.  I've been
using one I obtained from MIT, currently maintained by Eliot Moss
(EBM at MIT-XX).  Another new one is available from New Mexico Institute
of Technology: contact Bob Knight at 505-835-5735 - for $50...


[I just called NMIT and spoke to Greg Titus who is the developer of their C
for  TOPS-20.  He  said they plan to release their compiler in about a week
and described it as, in his opinion, better than the UNIX C compiler.  This
sounds  like  the  best bet so far and is probably the one we'll go with if
nothing better turns up (and if the project isn't cancelled). --JH]

==============================================================================

That's it.  Thanks again to all who took the time to contribute.

-- 
-_-_-_-_-_-_-_-_-_-_-_-_-_-_-_-_-_-_-_-_-_-_-_-_-_-_-_-_-_-_-_-_-_-_-_-_-_-_-_-
The Polymath (aka: Jerry Hollombe)
Citicorp TTI
3100 Ocean Park Blvd.
Santa Monica, CA  90405
(213) 450-9111, ext. 2483
{philabs,randvax,trwrb,vortex}!ttidca!ttidcc!hollombe