Path: utzoo!utgpu!water!watmath!clyde!ima!haddock!suitti
From: suitti@haddock.ima.isc.com (Steve Uitti)
Newsgroups: comp.sys.ibm.pc
Subject: Re: Microsoft Vs. Borland; benchmarks!
Message-ID: <9088@haddock.ima.isc.com>
Date: 10 Oct 88 02:50:49 GMT
References: <876@galaxy>  <8254@haddock.ima.isc.com>  <788@euraiv1.UUCP>
Reply-To: suitti@haddock.ima.isc.com (Steve Uitti)
Organization: Interactive Systems, Boston
Lines: 65

In article <788@euraiv1.UUCP>  evas@euraiv1.UUCP (Eelco van Asperen) writes:
> [Here's a comparison of MSC and TurboC as my contribution to the 
> "Microsoft vs. Borland" discussion.
> ...
> CONCLUSIONS
> 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. 

My data for sieve matches yours pretty well (I haven't played
with the other benchmarks).  My experience differs.  My
conclusions differ.  For nontrivial programs, MSC beats TC
(both with optimization).  However, for anything real
(largish), I've found that I can't use /Ox with MSC.  Something
breaks (who knows what).  Thus, I feel I have to compare MSC
with only /Gs (remove stack probes) with TC (whose optimizer
has never failed me).  When you do this, you find that TC wins
most of the time (but not all the time).

MSC (5.0, 5.1) would be a much better compiler if it (the
optimization part) worked.

I also test a version of "sieve" that has been hand crafted
with all sorts of pointer stuff.  Theory has it that a good
compiler will produce real good code with this version.  What
happens is that compilers with big optimizers (MSC 5.1 included)
actually do worse with the optimizer than without (though still
better than the non-pointer version).

> 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; start each reference on a separate
> page !  In its current form, the TC reference manual is a real
> pain to use.  I guess this is a Borland "feature" used to save
> paper and thus money on the cost of the manual. 

It also saves mass.  It is such a pain to read the MSC manauls
while riding the train to work...

TC's maual has an index.  It rarely fails me (***everything***
eventually fails me.  This makes me an awesome Beta tester.)
I find it easier to look something up in the TC reference
manual than to find the index (or which book to use) for MSC.

The TC User's Guide is quite well written.  It told me in
minutes how to get going.  By contrast, I'm still bewildered by
the options required by MSC.  I have to look them up for nearly
every project.

Much of this is reflected by the fact that the Turbo
Environment is simply thousands of times more friendly.  It
gets you working sooner.  MSC is infinitely more frustrating.
You have to use the manuals all the time.  You have to wait for
the beast to compile.  The compile breaks (sometimes even
without the optimizer turned on!).  What did I do to deserve
this?  (I paid good money, and lots of it.).

> One of the things missing from both compilers (and from most PC
> C-compilers for that matter) is profiling.

Good point.  Think C for the Mac has had profiling for
awhile now.  I actually use it now & again.

	Stephen.