http://www.jsoftware.com/jwiki/Essays/Memoization describes two implementations of "fib" that differ only by one having M. and one not. The one with M. is faster than the one without by a factor of 100,000 on an argument of 32.
----- Original Message ----- From: Terrence Brannon <[EMAIL PROTECTED]> Date: Sunday, May 20, 2007 12:24 pm Subject: Re: [Jprogramming] gerund creation failure > On 5/20/07, Devon McCormick <[EMAIL PROTECTED]> wrote: > > > > > > Of course, this is because they reduce recursion. Which, to my > mind, brings > > into > > question the whole project: why compare recursive versions for > speed when > > the > > greatest speed-up comes from doing away with recursion? > > Yes, the point of the Shootout is not to see how well a language can > solve a problem. But how well the language can do certain things when > all languages are forced through the same hoops > > So here they are testing the ability of the language to execute > recursive calls. Not the ability of skilled authors to come up with > the optimal solution vis-a-vis space/time. ---------------------------------------------------------------------- For information about J forums see http://www.jsoftware.com/forums.htm
