[net.micro.pc] Optimiczing CI-C86?

ctk@ecsvax.UUCP (12/10/83)

I see ads now for the Optimizing CI-C86 compiler. I've tried to
call them but always get a busy line. Does this really exist?
Has anyone out there got one? Is anyone from CI out there?
 I have the current version of their compiler and like it a
lot but, like most anyone, need MORE SPEED.
    C.T. Kelley
    Dept. of Math.
    N.C. State U.
    Raleigh, N.C. 27650
    decvax!mcnc!ecsvax!ctk

bae@astrovax.UUCP (Brian Ehrmantraut ) (12/11/83)

	I've got the beta version of their compiler, and it
works great, except for a few minor bugs which are already fixed
in the release version.  One note : it is very slow in compiling
when you have the optimizer and the large-model option cranked
up.

	The code it puts out is very tight, usually 2-4x the speed
produced by the older version.



		Brian A. Ehrmantraut

		allegra!{astrovax, fisher, twiggy}!bae

nathan@orstcs.UUCP (12/14/83)

#R:ecsvax:-170000:orstcs:8500007:000:1305
orstcs!nathan    Dec 13 18:35:00 1983


We obtained version 1.33d in April '83 (the latest non-optimizing version),
with the promise that for our $400 we would be getting the optimizing version
within a month.  We have yet to receive it.  About every month or so, they
say "in three or four weeks" when I call to check up.  Otherwise, I am very
pleased with the compiler and have had very little trouble contacting them.

I am particularly pleased that I can talk to the person who wrote the compiler
when I have difficulties.  (compiler bugs do surface on occasion...)

The optimizing version still won't have structure assignment or enum types.
These facilities have been in Unix C for several years now, and there is
little excuse for not putting them into a modern compiler.  Recently we
had occasion to port code produced by yacc to this compiler; the CI 
compiler rejected it. (with some hand-massaging we made CI accept it)

CI also does not ship a lint with their compiler.  Since nobody else but
Bell seems to, we can't really single them out; but they claim to supply
"full K&R C".  Lint is mentioned many times in K&R as accompanying the
C compiler, and is described as being as much a part of the C system as
"stdio".  

Can we consider a C compiler product complete K&R if it doesn't include lint?

	Nathan C. Myers 
	Alan Batie