Search found 18172 matches

by hgm
Fri Feb 21, 2020 6:01 pm
Forum: Computer Chess Club: Programming and Technical Discussions
Topic: hash collisions
Replies: 352
Views: 10132

Re: hash collisions

OK, I stand corrected. But that was long after you started dumping on me, and so can be hardly called an excuse for appearing here.
by hgm
Fri Feb 21, 2020 5:36 pm
Forum: Computer Chess Club: Programming and Technical Discussions
Topic: hash collisions
Replies: 352
Views: 10132

Re: hash collisions

Your whole appearence here had no other intention than to do just that; so far you have said nothing on-topic here.

So, like I said, just unfounded moderator bashing, and when you are challenged to account for it, you cannot deliver.
by hgm
Fri Feb 21, 2020 4:54 pm
Forum: Computer Chess Club: Programming and Technical Discussions
Topic: EPD destruction tests
Replies: 58
Views: 1062

Re: EPD destruction tests

Well, always searching the history all the way back to the beginning of the game is not wrong (and in fact what you have to do in Crazyhouse to not be wrong). But because for some of the history comparison is pointless, it is an unnecessary performance drain. But so is keeping track of a duplicat co...
by hgm
Fri Feb 21, 2020 12:55 pm
Forum: Computer Chess Club: Programming and Technical Discussions
Topic: EPD destruction tests
Replies: 58
Views: 1062

Re: EPD destruction tests

You can also hash the history. Then you only have to check one history item (or in case of a collison, which should be rare if the table is large enough, sometimes two). Then you don't need to update any counter, and effectively compare to the entire history. In Crazyhouse or Shogi you would have to...
by hgm
Fri Feb 21, 2020 11:57 am
Forum: Computer Chess Club: Programming and Technical Discussions
Topic: hash collisions
Replies: 352
Views: 10132

Re: hash collisions

No, it is not. It was a justified conclusion from your preceding apparent rejection of speed-enhancing techniques because they could occasionaly result in a very poor move. And it came with an extensive explanation of why in chess programming such an attitude would not work. The inclusion of the wor...
by hgm
Fri Feb 21, 2020 11:28 am
Forum: Computer Chess Club: Programming and Technical Discussions
Topic: Zobrist key independence
Replies: 22
Views: 642

Re: Zobrist key independence

Some more musings about Kings To be frank, I am not sure how much all this King stuff helps. The King really seems a special case, because you know there will always be one and only one. Suppose every King key would have a 6th-order dependency with the basis keys used for the other pieces. That mea...
by hgm
Fri Feb 21, 2020 10:56 am
Forum: Computer Chess Club: Programming and Technical Discussions
Topic: hash collisions
Replies: 352
Views: 10132

Re: hash collisions

So why do you quote this one, then? Don't eat your heart out, quote anything I posted here that you think violated the charter. Or is this just your way of launching an unmotivated personal attack on moderation? And better not come up with crap like that "Then writing chess programs is the wrong bus...
by hgm
Fri Feb 21, 2020 9:25 am
Forum: Computer Chess Club: Programming and Technical Discussions
Topic: hash collisions
Replies: 352
Views: 10132

Re: hash collisions

Rebel wrote:
Thu Feb 20, 2020 10:18 pm
hgm wrote:
Thu Feb 20, 2020 10:04 pm
Suite yourself. I would say it is prudent, and good strategy! :lol:
Moderator moderate yourself, if you can't resign.
In what way does the quoted posting violate the charter, then???
by hgm
Fri Feb 21, 2020 9:06 am
Forum: Computer Chess Club: Programming and Technical Discussions
Topic: Zobrist key independence
Replies: 22
Views: 642

Re: Zobrist key independence

Improving the King keys If we use the obvious numbering of board squares (a1=0, b1=1, ...) to derive the King keys for the squares, a1 and b1 would only differ (XOR-wise) by a single basis key, and could thus be confused if that basis key is involved in some high-order dependency with other basis k...
by hgm
Thu Feb 20, 2020 10:16 pm
Forum: Computer Chess Club: Programming and Technical Discussions
Topic: SIMD methods in TT probing and replacement
Replies: 15
Views: 426

Re: SIMD methods in TT probing and replacement

Yes, that was the initial idea. Originally I proposed the older MMX instructions for this; these also have an 8x8bit compare. Problem is that you have to use compiler intrinsics, and that it often takes extra instructions to shuttle the data into the SIMD registers. This is why the ordinary uint64 a...