Path: utzoo!attcan!uunet!imspw6!bob
From: bob@imspw6.UUCP (Bob Burch)
Newsgroups: comp.sys.ibm.pc
Subject: Re: MicroSoft/Borland, execution speed least critical
Message-ID: <176@imspw6.UUCP>
Date: 10 Oct 88 12:59:40 GMT
Organization: IMS Inc., Rockville, MD
Lines: 75




From: Eelco van Asperen :
Erasmus University EF/AIV,Rotterdam,Netherlands
 
>The results clearly show that the Microsoft compiler produces superior
>code when compared to Borland's. In a number of cases the MS-C code
>outperformed TC's by a factor of ten, for example with the TLONG and ARRAY
>tests. The good optimization in MS-C also provides some problems; since
>benchmarks are artificial and try to measure the efficiency of a
>certain type of operation, they are extremely prone to being optimized
>away, ie. reduced to no code at all.  This is shown best by the
>LOOPTST, STORAGE and TRIG programs. We definitely need a new class of
>benchmarks for future tests.
 
Borland and Microsoft are leap-frogging eachother in this regard with each
new release;  Turbo 1.5 checked out slightly faster than MSC 5.0 in the
Turbo Tech Journal's comparison, something like 16 out of 30 categories.
It would surprise me if TurboC 2.0 doesn't generate faster code than MSC
5.1 but, when you get right down to it, you will never notice any differences
in execution speed between MSC and TurboC code in 999 out of 1000 applic-
ations.  Of all of the comparison factors between the two, this one ranks
at the very bottom of my list of importance by a considerable margin over
whatever's in second-to-last place.
 
Things which are far more important would include the following:
 
1.  Borland invented the idea of selling compilers for $100 or so instead
    of $500 or $1500.  Microsoft is being forced to play a game they don't
    enjoy playing and which they would cease to play tommorrow should
    anything happen to Borland.  The loyalty issue here is pretty real.
 
2.  Bugs.  The absense of serious bugs factor favors Borland heavily.
 
3.  Compile speed / fast prototyping.  Real world programs get written
    in finite sums of time and with finite amounts of psychic energy.
    The difference between what I can accomplish with Borland's
    integrated environment (which now includes serious debugging) and
    3 second compiles versus what I might could accomplish using the
    old-fashioned system, MSC, (Quick C with it's one memory model and
    it's failure to recognize half the hardware out there is not a
    serious alternative) will show up FAR more than any minute differences
    in quality of code generation.
 
>Based on the data presented here and my experiences with both products,
>Microsoft C wins the battle; it generates by far the best code. Turbo
>C's one-pass compiler has shorter compile times and creates smaller
>executables but the code produced is inferior to MS-C's.
 
   You don't win the war by winning the least critical battle.
 
>Furthermore, when it comes to writing a reference manual for a language
>the boys (and girls) at Borland could learn something from the
>Unix-community........
 
   I like UNIX as much as the next guy and possibly more (cf. PCWeek
   letters to editor, Oct 3 issue).  But UNIX documentation???????
   Do you mean the documentation for Curses, for sockets...??????
   I am just now starting to see some more or less rexpectable UNIX
   documentation from one or two major firms such as UNISYS  (the
   red books), but the stuff I've been seeing forever is bullshit
   pure and simple, based on the notions that if it isn't arcane, it
   isn't fun, real men don't NEED documentation, etc. etc.  I hope to
   hell neither Borland Nor MicroSoft nor anyone else in the PC world
   ever learns anything at all from whoever is responsible
   for standard UNIX documentation.
 
 
I am now getting first looks at TurboC 2.0, Turbo Debugger, TASM etc. and,
from what I'm seeing, it looks like MicroSoft has a good deal of catching
up to do.
 
Ted Holden
HTE