Xref: utzoo comp.sys.mac:48889 comp.sys.mac.programmer:12563
Path: utzoo!utgpu!jarvis.csri.toronto.edu!rutgers!tut.cis.ohio-state.edu!zaphod.mps.ohio-state.edu!uakari.primate.wisc.edu!aplcen!uunet!usna!pjw
From: pjw@usna.MIL (Prof. Peter Welcher)
Newsgroups: comp.sys.mac,comp.sys.mac.programmer
Subject: MPW vs Think C performance
Keywords: performance,C,C compilers
Message-ID: <369@usna.MIL>
Date: 18 Feb 90 15:55:36 GMT
Followup-To: poster
Distribution: usa
Organization: U.S. Naval Academy, Annapolis, MD
Lines: 26


I've got a large-ish program that I'm polishing. I'm interested in
performance enhancements. It was suggested to me by two very 
knowledgeable sources (pros with several commercial products) that
MPW C produces better code than LightSpeed (Think) C, by up to 30%.
So I bought MPW, ported, and 450K of executable went to 500K. There
seems to be a 5-10% performance hit too.

So I wrote some small quick-and-dirty benchmarks. MPW seemed to do worse than
LSC at integer and floating arithmetic, but a bit better at string and
pointer manipulation. This seems to confirm my porting experience.

I am using 2-byte ints and 10-byte doubles in MPW. That doesn't seem to be
the problem, based on benchmarking.

My question: is the above borne out by general experience ? Could I
be missing a compiler flag or something (I doubt it) ? Are there
after-the-fact code optimizers available ? Is there a better C compiler
than MPW or Think C (i.e. faster, tighter code) ?

Since I read the news groups irregularly, and may lose a reply
in the clutter, I'd appreciate email responses:

	pjw@sma.usna.navy.mil

Thanks !