Path: chuka.playstation.co.uk!news
From: "Jon Prestidge  (Jon@surfed.to)" <jon@surfed.to>
Newsgroups: scee.yaroze.freetalk.english
Subject: Re: Daft Question [long]
Date: Mon, 27 Aug 2001 21:30:17 +0100
Organization: PlayStation Net Yaroze (SCEE)
Lines: 53
Message-ID: <9meajl$1sb17@www.netyaroze-europe.com>
References: <9loc6a$p3k2@www.netyaroze-europe.com> <9loeu8$p3k3@www.netyaroze-europe.com> <01c1297e$8cb6e8a0$LocalHost@pal-s-omnibook> <9lrqqf$3j63@www.netyaroze-europe.com> <01c129ca$a5a6f700$d81fe4d5@pal-s-omnibook> <9m1d9c$b2i7@www.netyaroze-europe.com> <01c12ceb$fba64880$9709e4d5@pal-s-omnibook> <9m9c88$ioj5@www.netyaroze-europe.com> <01c12edc$cc2bf440$991fe4d5@pal-s-omnibook> <9me1g1$1sb13@www.netyaroze-europe.com>
NNTP-Posting-Host: host213-122-71-74.btinternet.com
X-Priority: 3
X-MSMail-Priority: Normal
X-Newsreader: Microsoft Outlook Express 5.00.2615.200
X-MimeOLE: Produced By Microsoft MimeOLE V5.00.2615.200

> I have absolutely no idea how fast an integer divide is

... it's only one CPU instruction... it may take an extra clock cycle or two
longer to execute than a logical shift operation  for example, but it's
still very, very quick.

> cant be any slower then the several if's

.... one integer divide should be a lot quicker that several IFs.  The Risc
chip can only compare against zero so unless you're IF compares a value
against zero the IF statment will be two CPU instructions ( a subtract and a
compare-and-branch ).

A big over head is the loading and unloading of variables to CPU registers
from memory and back again...if you can arrange small performance critical
routines to use just a few variables (or just two if you can manage it) then
it's a lot quicker because there's more chance of the variable staying in a
CPU register and not being loaded back and forth to memory.

Jon


John ( QuietBloke ) <quietbloke@btinternet.com> wrote in message
news:9me1g1$1sb13@www.netyaroze-europe.com...
> aye.. To be honest I have absolutely no idea how fast an integer divide
is.
> Im just kinda assuming it cant be any slower then the several if's and
> shifts' etc of the other solution.
> Anyway.. I shall keep you informed.. and thanx for the initial code.. was
a
> real time saver.
>
> Cheere
> John ( Quietbloke )
>
> "pal" <pal_news@hotmail.com> wrote in message
> news:01c12edc$cc2bf440$991fe4d5@pal-s-omnibook...
> > I agree that this new method is far more elegant than the shift's one,
but
> > are you sure it'll be quick enough? It uses three divide's, which I
> believe
> > will take a lot of time, won't they?
> >
> > Anyway, please let us know the results and your final implementation, as
> > this angle problem is a recurrent one...
> >
> >
> > pal
> >
>
>