On Monday, April 29, 2002, at 02:03 , Bob Showalter wrote: [..] http://www.wetware.com/drieux/CS/lang/Perl/Beginners/BenchMarks/randTest.txt has been updated to take in more data... > > splice() will be slower as the size of the array grows. If I take > your benchmark and change to array from 1..100 to 1..10000, I get > the following results for 100 iterations (on an old Pentium 266): WOW! talk about size matters.... some of the gambits work better at a given size than others. I think the decLoop trick because of it's overall reasonable performance across the spectrum seems to clearly have merit. I need to find a more effective way to show the strength's of felix's fisher_yates_shuffle from the perldoc's since I rather doubt that I would use it quite the way that I tested it - and saw no real significance from 'inlining' the code.... ciao drieux ---Thread Previous | Thread Next