Author: Sune Fischer
Date: 04:45:17 12/09/01
Go up one level in this thread
On December 08, 2001 at 23:56:58, Robert Hyatt wrote: >On December 08, 2001 at 05:21:33, Sune Fischer wrote: > >>On December 07, 2001 at 23:52:37, Robert Hyatt wrote: >> >>Bob, I can tell you, that if you do not have more than 77000 pawn positions and >>still get collisions, then you do have a bug. >> >>-S. > > >For the tests I ran I had a _lot_ more than 77000 pawn positions. I used a >hash size of 6M, and it was filled quite quickly. Remember that I didn't >just search a single position. I annotated a full game. > >I don't see any way for a bug to exist here. The PRN's are known to be >good, hamming-distance-wise, and I always run "validate" tests on everything >for each new version. That takes that incremental signature and compares it >to a "computed-from-scratch" signature to make sure they match and catch any >errors in the incremental update code (among other things)... Okay, no bug then, you should be getting collisions. With more than 77000 "birthdays" and only 2^32 "days to a year", many students will celebrate on the same day ;) Thanks for testing, it makes sense now :) -S.
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.