Path: utzoo!utgpu!news-server.csri.toronto.edu!cs.utexas.edu!uunet!bfmny0!tneff
From: tneff@bfmny0.BFM.COM (Tom Neff)
Newsgroups: comp.sys.intel
Subject: Re: Intel-8086/80186-Assembler for System-V available ?
Message-ID: <15947@bfmny0.BFM.COM>
Date: 14 Oct 90 03:25:43 GMT
References: <15919@bfmny0.BFM.COM> <1990Oct11.084750.1183@orfeo.radig.de> <96@calcite.UUCP>
Reply-To: tneff@bfmny0.BFM.COM (Tom Neff)
Lines: 55

In article <96@calcite.UUCP> vjs@calcite.UUCP (Vernon Schryver) writes:
>
>Maybe PLM*86 would run faster as a native UNIX program.  Who cares?  It
>would generate the same code.

This is quite true, and in fact a virtue, the rest of Vern's message
notwithstanding.

>PLM86 was mind boggling bad even at peep-hole optimizing in MDS-311 days.
>(MDS-311 was the ISIS 8085 cross development package you could get for your
>very own low speed 8086 in 1978.)  

Ah, and what a system it was.  We still have one of the old Tower of
Power mds's sitting around somewhere.  It would make... an attractive
planter.

This, of course, dates from the long-forgotten days when micros were
required to do useful WORK to earn their supper; unlike our modern
all-electric era when they're expected to spew waste-cycles doing
pointless X Windows crap like all the other CPUs.

>                                   The "code" generated by PLM386 is not
>significantly different, at least in the versions I've used.  PLM386 still
>wastes registers and cycles with abandon--as if the *86 had any to spare.

There was one major revision of the PL/M code generator, about 1987.
Other than that it is largely still the same old cranky beast that
existed in 1981.  It doesn't do quite so many wild ass optimizations as
whatever precious academic effort has everyone's heart a-flutter this
week.  But it is very *predictable* and that is extraordinarily
comforting in a number of unglamorous, rent-paying types of situations.

>It still cannot strength reduce multiplies of constant powers of two.  It
>usually ignores and so recomputes condition codes generated by preceding
>statements.  And so on and on.

Yes, and so on... every PL/M programmer has at least one thing he HATES
about the compiler.  That's how I know it's a good one.  It spreads the
hate around rather evenly.

>Anyone with freedom of choice (i.e. no dusty albatrosses) would choose
>almost any C compiler.

Oh yeah?  I suggest looking at the original Mark Williams C compiler,
which was OEM'd as "Intel C" up through 3.0.  You want to talk about
BAAAAAAD code!  I'm amazed they offered an assembler listing option at
all: simple modesty ought to have intervened. :-)  Nowadays Intel C has
been totally rewritten with an ANSI parser -- and the PL/M code
generator!!!  So there's no way out.  If you want to generate OMF, that
is.  If you don't want OMF, you've invented some other job for yourself,
and you might as well add a Cray to the wishlist while you're at it.
-- 
Technology is a way of organizing       '   '     Tom Neff
the universe so that man doesn't have    '   '    tneff@bfmny0.BFM.COM
to experience it. -- Max Frisch           '   '   uunet!bfmny0!tneff