Genetic Algorithms and Compiler Optimizations

mfago writes "Scott Robert Ladd has written an enlightening article and accompanying software package that utilizes genetic algorithms to optimize the optimization flags fed to a compiler. Those who have tried to squeeze the last drop of performance from a code know that it can be very difficult to determine the set of optimization flags beyond -O3 (with gcc for example) that yields the best performance. Lots of background links included in the article."

Comments

Isn't this the Windows technique?

[Anonymous Coward]

That is, fiddle with settings at random until it's acceptable?

Re:Beware the simple, elegant solution.

[tessaiga]

In Go, there is a computer player that utilizes neural networks, yet it is near the middle of the pack.

Wait, I'm confused; I thought on Slashdot we're supposed to wait for someone to mention Chess before we bring up Go?

Re:Beware the simple, elegant solution.

[koekepeer]

anyways think about the overkill of developing such an elegant solution for such a minor problem. i mean those few percentages of performance enhancement...

anyway, i think it *is* really cool this guy finds it interesting to tackle this 'problem' with a genetic algorithm. why not try some bayesian networks next? support vector machines? improbability drives...

ooops, i'm getting carried away. sorry 'bout that. /me bows in admiration

(not intended as flamebait but mod me to hell anyway *evil grin*)

Re:Beware the simple, elegant solution.

[zanerock]

Neural networks != genetic algorithms

Slow Go player != inelegant solution

I was going to dissagree with you, but I'm not even sure what your arguments are. Do you have any reason to believe that the solution for the problem discussed in the article is inelegant or innefficient, or did you just get a bad grade in your AI class?