Author: Robert Hyatt
Date: 15:43:19 01/20/98
Go up one level in this thread
On January 20, 1998 at 18:24:45, Stuart Cracraft wrote: >On January 20, 1998 at 17:13:06, Bruce Moreland wrote: > >> >>On January 20, 1998 at 16:42:12, Stuart Cracraft wrote: >> >>>On January 20, 1998 at 16:36:09, Bruce Moreland wrote: >>> >>>>When I wrote my other response, I misunderstood what you meant. >>>> >>>>What happens if you do this for 20 different positions? >>>> >>>>Maybe it works for you. >>>> >>>>bruce >>> >>>The results of the first 20 positions from Win-at-Chess each searched >>>to 6 ply, excepting mates, is posted elsewhere on this thread. I don't >>>know why you haven't been able to see it yet. >>> >>>Anyway, the result was 10,000 more nodes (out of a quarter-million), >>>but an improvement in search rate more than sufficient to offset this >>>and result in a 10% timing improvement for the whole set. >>> >>>As for rehashing comparing linear vs. random, I haven't had time to do >>>that one yet. >> >>I messed up, I didn't see the "20", and I just picked that number at >>random later, and it was the same one you'd actually picked. >> >>I think that I should admit that I didn't understand your original post >>and I don't understand the source code you provided later. I don't know >>why "quiescence" is a void function and I don't know what use it is to >>probe the hash table after calling it. >> >>But if it works, by all means keep doing it. >> >>bruce > >My quiescence function isn't void but for the sake of brevity was listed >as so. Sorry for the confusion brevity created. > >Anyway, I think the test results show that for short 6-ply searches, >doing >quiescence beyond the full-width depth before tranposition table probes >is better than after transposition table probes. > >I can run the full suite but think I'd find the same thing. I don't know >why >Bob Hyatt's result would be different. I'm sorry, but this is insane. What is the purpose of doing a hash probe? To eliminate searching the positions you already know the outcome of, right? But instead, you are searching *first* and then looking up *second*? Of what use is that lookup? You have already done the search.. So throw that lookup out and go even faster, because it can't possibly provide any info that the search you just completed did. If this is better, you *must* have a bug, because the purpose of doing this is to eliminate that call to Quiesce() totally, which should happen at least 10% of the times you call lookup. Ergo, if it works, something else is broken. *badly*... Try fine70 both ways... and I assume you are also doing the same thing before you call Search()? IE you call Search() first and then call Lookup()? If not, why are you treating 'em differently??
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.