Path: utzoo!utgpu!jarvis.csri.toronto.edu!mailrus!tut.cis.ohio-state.edu!zaphod.mps.ohio-state.edu!uakari.primate.wisc.edu!aplcen!haven!uvaarpa!murdoch!krebs.acc.Virginia.EDU!wrp
From: wrp@krebs.acc.Virginia.EDU (Bill Pearson)
Newsgroups: comp.sys.mac.programmer
Subject: Think 'C' vs MPW file i/o
Message-ID: <1990Feb18.021804.18148@murdoch.acc.Virginia.EDU>
Date: 18 Feb 90 02:18:04 GMT
Sender: news@murdoch.acc.Virginia.EDU
Reply-To: wrp@krebs.acc.Virginia.EDU (Bill Pearson)
Distribution: usa
Organization: Academic Computing Center, University of Va.
Lines: 20


	I recently converted a program that scans and analyzes the contents
of at 4 mbyte file from Think 'C' to MPW 'C', in order to take advantage
of 32 bit ints.  To my surprise, the program was about 1/2 as fast under
MPW 'C' (running under the MPW shell).  Even when the program was compiled
to use 68020 codes, it still took 50% longer to run.

	My Think version used to take about twice as long, until I increased
the file read buffer size with setvbuf().  I have also done this with
the MPW version, but I am suspicious that setvbuf() may not really work
under MPW C 3.0.  I note that in stdio.h, the FILE data structure only
uses an unsigned short for the buffer size.  Since MPW C does not
supply source for the libraries, I cannot check.

	Thus, I have two questions:  Does the MPW shell slow things down
substantially?  Does setvbuf() work properly under MPW 'C'?  (Really I
only have one question, how do I make my program run as fast under MPW
as it does under Think 'C'?)

Bill Pearson