Author: Gerd Isenberg
Date: 06:52:35 07/08/03
Go up one level in this thread
On July 08, 2003 at 08:28:07, Uri Blass wrote: >On July 08, 2003 at 08:16:01, Tim Foden wrote: > >>On July 08, 2003 at 05:31:04, Gerd Isenberg wrote: >> >>>Andrew's one - as expected faster than Tim's approach >>> >>>so far: >>>Andrew n = 1799993753, time = 5.828 >>>Tim n = 1799993753, time = 6.875 >>>Switch256 n = 1799993753, time = 2.671 >>> >>>so now i really have to work ;-) >> >>Ah... But what about with a data set where on average only 1.27405 bits are set >>(see http://www.talkchess.com/forums/1/message.html?305213)? I guess this may >>bring the numbers closer together... but my simple approach will probably still >>be slower. :) >> >>Cheers, Tim. > >I decided to use Andrew approach. >maybe switch 256 is faster(I doubt it) but I first need to do things correctly >and >other ideas that may be faster have the risk of bugs. > >I already had a bug when I tried to use the same idea in another place >because I wrote the functions wrong or I did the function in the wrong time. > >I do not need another source for bugs. > >Andrew approach seems to be simple enough to avoid bugs in implementing when >switch256 does not seem to be simple and maybe I should think about considering >switch256 only after I finish the job correctly with andrew's approach. > >Uri Hi Uri, for sure - one need to write a c-source generator for the switch256. And in "real" life i guess your handler for each bit require some bytes more as n += someNumber ;-) Gerd
This page took 0 seconds to execute
Last modified: Thu, 15 Apr 21 08:11:13 -0700
Current Computer Chess Club Forums at Talkchess. This site by Sean Mintz.