From: Shawn Hargreaves <Shawn AT talula DOT demon DOT co DOT uk>
Newsgroups: comp.os.msdos.djgpp
Subject: Re: Allegro performance
Date: Fri, 14 Mar 1997 08:25:05 +0000
Organization: None
Distribution: world
Message-ID: <xePvQAAhtQKzEwDh@talula.demon.co.uk>
References: <Pine DOT GSO DOT 3 DOT 95 DOT 970313113331 DOT 29601A-100000 AT pesqueira>
NNTP-Posting-Host: talula.demon.co.uk
MIME-Version: 1.0
Lines: 22
To: djgpp AT delorie DOT com
DJ-Gateway: from newsgroup comp.os.msdos.djgpp

Saulo Medeiros de Araujo writes:
>       I am developing a simple graphic library and I would like compare
>its performance versus allegro. I am using the results of the test program
>included in allegro. Are these results the best allegro can produce or the
>test program adds some unnecessary delay?

They are very approximate: no correction is made for the delays
introduced by generating the random numbers, looping, checking for
keypresses, etc. Also, the test program draws everything directly to
video memory, while you can usually get dramatic speed increases by
rendering onto a memory bitmap instead.

The main reason for displaying the timings was as a profiling tool
(while optimising a routine, I can easily see if a change improved it or
not), rather than as a 100% accurate measurement. I'd estimate the
figures are in error by up to 5 or 10 percent...


/*
 *  Shawn Hargreaves - shawn AT talula DOT demon DOT co DOT uk - http://www.talula.demon.co.uk/
 *  Beauty is a French phonetic corruption of a short cloth neck ornament.
 */