From: "Damian Yerrick" <web DOT poison AT pineight DOT 8m DOT com>
Newsgroups: comp.os.msdos.djgpp
Subject: Re: Pointers vs subscripting.
Date: Wed, 29 Sep 1999 20:33:52 -0500
Organization: Rose-Hulman Institute of Technology
Lines: 56
Message-ID: <7suepg$bip$1@solomon.cs.rose-hulman.edu>
References: <QIbyN9jI2fdBqTToI8WDui1c9Rdn AT 4ax DOT com>
NNTP-Posting-Host: 137.112.205.146
X-Trace: solomon.cs.rose-hulman.edu 938655344 11865 137.112.205.146 (30 Sep 1999 01:35:44 GMT)
X-Complaints-To: news AT cs DOT rose-hulman DOT edu
NNTP-Posting-Date: 30 Sep 1999 01:35:44 GMT
X-Priority: 3
X-MSMail-Priority: Normal
X-Newsreader: Microsoft Outlook Express 5.00.2314.1300
X-MimeOLE: Produced By Microsoft MimeOLE V5.00.2314.1300
To: djgpp AT delorie DOT com
DJ-Gateway: from newsgroup comp.os.msdos.djgpp
Reply-To: djgpp AT delorie DOT com


Kevin <krugman AT nospam DOT demon DOT co DOT uk> wrote:

If you're trying to mung your address so that spambots don't get
it, changing the domain to nospam.* doesn't work; your mail will
probably still get through. Try changing demon to all caps and
inserting a cute word (exorcist or something?) between two of
the letters.

krugman AT DEexorcistMON DOT co DOT uk

This will confuse spambots more thoroughly.

> Ive always gone to great lengths to avoid subscripting
> arrays and have gone to lots of trouble to use pointers (at
> least in the most frequently accessed loops, I don't usually
> bother with program initialisation things).
>
> I have understood that this is the quicker way of doing things.

Was.

> Is this still the case or do modern compilers optimise this
> automagically now ?

Compilers use a technique known as "strength reduction" to
turn multiplies into adds and adds into increments. If you are
accessing a multidimensional array, try storing a pointer to the
first element.

int foo[6][6];
int *fooLine;

for(y = 0; y < 6; y++)
{
  fooLine = foo[y];
  for(x = 0; x < 6; x++)
  {
    DoSomethingWith(fooLine[x]);
  }
}

Play with it for a while. Remember to compile with at least -O3
for best results.

> I realise that a little experimentation would probably answer this
> question for me, but I'm posting another message here at the same

If you are wondering about the speed of one technique versus
another, you always have gprof. Read the FAQ at
http://www.delorie.com/djgpp/v2faq/

Damian Yerrick
http://come.to/yerrick