Search found 370 matches

by op12no2
Fri Jan 15, 2021 9:53 am
Forum: Computer Chess Club: General Topics
Topic: New engine releases & news 2021
Replies: 34
Views: 4184

Re: New engine releases & news 2021

- Lozza 1.19 (also all previous versions available now) https://github.com/op12no2/lozza/tree/master/history Hi Guenther, My apologies, I've deleted 1.19 now and replaced it with 2.0 which is WIP - I'll be more careful with my commit comments! I'll add a final commit saying released but it will not...
by op12no2
Tue Jan 12, 2021 4:01 pm
Forum: Computer Chess Club: Programming and Technical Discussions
Topic: Weaker play with TT
Replies: 31
Views: 2523

Re: Weaker play with TT

can I somehow manually reproduce wrong hashes without running search to figure out what's going wrong? I'm not sure I totally follow your code, but, if your hash reflects the current player when you get to nmp, I think you need to tweak ep and the player turn, since you are going to search with the...
by op12no2
Tue Jan 12, 2021 3:43 pm
Forum: Computer Chess Club: Programming and Technical Discussions
Topic: Weaker play with TT
Replies: 31
Views: 2523

Re: Weaker play with TT

That would probably be a lot faster than messing with JavaScript objects containing a variety of types as fields. And also more memory efficient. Agree, (that's what Lozza does). Maksim, if you use typed arrays you know exactly what is happening and they are very fast. Using objects in this context...
by op12no2
Tue Jan 12, 2021 2:29 pm
Forum: Computer Chess Club: Programming and Technical Discussions
Topic: Weaker play with TT
Replies: 31
Views: 2523

Re: Weaker play with TT

maksimKorzh wrote:
Tue Jan 12, 2021 2:03 pm
Yup. I'm aware of BigInts and typed arrays - was thinking to consider them but rejected due to the same reasons you did.
Just to clarify, I do use typed arrays, I don't use big ints.

cj
by op12no2
Tue Jan 12, 2021 12:26 pm
Forum: Computer Chess Club: Programming and Technical Discussions
Topic: Weaker play with TT
Replies: 31
Views: 2523

Re: Weaker play with TT

Hi Colin, nope - no email from - only that regarding PRNG you've sent previously, could you please resend it? Done. PS As I recall you can get up to 54 reliable bits :roll: with a simple Javascript number - even though there are more floating (ha) around - p4wn does that I think. But I use 2 32 bit...
by op12no2
Mon Jan 11, 2021 10:55 pm
Forum: Computer Chess Club: Programming and Technical Discussions
Topic: Weaker play with TT
Replies: 31
Views: 2523

Re: Weaker play with TT

Hi Maksim, I did reply to your email re TT - did you see it?
by op12no2
Sat Jan 09, 2021 11:22 am
Forum: Computer Chess Club: General Topics
Topic: Website engines
Replies: 33
Views: 6024

Re: Website engines

I can confirm that replacing my old version of Node (5.0.0.0) against the one I am using now with WukongJS (11.0.0.0) gives a speedup of 401 kn/s vs. 352 kn/s ~ +14% => this was with lozza 1.18 from the startpos after around 150s Awesome :) I sat watching Lozza do a soak test yesterday in Cutechess...
by op12no2
Fri Jan 08, 2021 9:20 pm
Forum: Computer Chess Club: General Topics
Topic: Website engines
Replies: 33
Views: 6024

Re: Website engines

BTW will there be any speed difference or other advantages using a newer Node? I just realized I never tried to update Node for at least five years. Hi Guenther, Apologies, I must have been drinking when I replied to this before. Yes the Google V8 Javascript engine that node.js is built around - an...
by op12no2
Wed Dec 30, 2020 8:25 am
Forum: Computer Chess Club: Programming and Technical Discussions
Topic: Blunder option
Replies: 37
Views: 2609

Re: Blunder option.

Ajedrecista wrote:
Tue Dec 29, 2020 6:26 pm
Good luck with your implemetation!

Regards from Spain.

Ajedrecista.
Thanks Ajedrecista. I have not read any of these threads in full yet but they all look very interesting.
by op12no2
Tue Dec 29, 2020 5:13 pm
Forum: Computer Chess Club: Programming and Technical Discussions
Topic: Blunder option
Replies: 37
Views: 2609

Re: Blunder option

So after generating all moves by multipv the next task is classify the moves to get candidate move appropriate at the level the engine is supposed to play. Thanks Ferdy and for the link. Thanks also Sesse for the link and feedback. Interesting stuff and far more involved that I thought it would be :)