Re: lazy execution optimization

["Steven W. Orr" <steveo at world dot std dot com>]

I'm definitely not a qualified person to answer this, but have you tried
the following construct to see if the outcome is affected:

int f (double a, double b)
{
  double c;
  c = a+b;
  if (a > 0.0)
    return 0;
  else if (c > 0.0)
    return 1;
  else
    return 2;
}

I'm just speculating that maybe this in conjunction with -O2 might just
reorder in a better way.


-- 
-Time flies like the wind. Fruit flies like a banana. steveo@world.std.com-
-Stranger things have happened but none stranger than this. Steven W. Orr-
Does your driver's license say Organ Donor?Black holes are where God \
-------divided by zero. Listen to me! We are all individuals!---------

On Sun, 26 Sep 1999, Sylvain Pion wrote:

=>Let's consider the following function:
=>
=>int f (double a, double b)
=>{
=>  double c;
=>  c = a+b;
=>  if (a > 0.0) return 0;
=>  if (c > 0.0) return 1;
=>  return 2;
=>}
=>
=>I would like GCC to optimize it by computing "c" only after the first test,
=>just as if the code was:
=>
=>int f (double a, double b)
=>{
=>  double c;
=>  if (a > 0.0) return 0;
=>  c = a+b;
=>  if (c > 0.0) return 1;
=>  return 2;
=>}
=>
=>It seems 2.95.1 as well as the current CVS do not perform such an
=>optimization.
=>
=>Would it be reasonnable and possible ?
=>
=>