Search found 1727 matches

by Gerd Isenberg
Fri Jan 15, 2021 9:51 pm
Forum: Computer Chess Club: Programming and Technical Discussions
Topic: How to efficiently generate individual moves from bitmasks of pieces?
Replies: 5
Views: 137

Re: How to efficiently generate individual moves from bitmasks of pieces?

I have been trying to build a chess engine for a variant of chess called Monster Chess, in which white starts with just a king and 4 pawns but has two moves for every move black has - I posted about that here, if you are interested ( link ). In any case I have been trying to implement simple bitboa...
by Gerd Isenberg
Tue Jan 12, 2021 8:20 pm
Forum: Computer Chess Club: Programming and Technical Discussions
Topic: Why using the game result instead of evaluation scores
Replies: 12
Views: 388

Re: Why using the game result instead of evaluation scores

I got that a little bit later. One needs to scale the engine specific score to a win percentage sigmoid for a 0-1 or -1,0,1 range.
Similar to TD(λ) one may even try to interpolate the final result into that score. I don't know whether this was tried before.
by Gerd Isenberg
Tue Jan 12, 2021 7:46 pm
Forum: Computer Chess Club: Programming and Technical Discussions
Topic: Why using the game result instead of evaluation scores
Replies: 12
Views: 388

Re: Why using the game result instead of evaluation scores

Hello everybody, currently some threads for tuning are ongoing. Here is another question i am interested in. When creating the training data it wouldn't be a big deal to use a trainer (any engine) that computes a score for each of the positions and stores them. Instead of computing the error with t...
by Gerd Isenberg
Thu Jan 07, 2021 4:24 pm
Forum: Computer Chess Club: General Topics
Topic: Translation of Yu Nasu's NNUE paper
Replies: 6
Views: 935

Re: Translation of Yu Nasu's NNUE paper

I was very interested in how the NNUE is actually designed an implemented. Yu Nasu originally introduced it in his 2018 paper "Efficiently Updatable Neural-Network-basedEvaluation Functions for Computer Shogi" (available here https://www.apply.computer-shogi.org/wcsc28/appeal/the_end_of_genesis_T.N...
by Gerd Isenberg
Thu Dec 31, 2020 7:12 pm
Forum: Computer Chess Club: General Topics
Topic: Dispelling the Myth of NNUE with LazySMP: An Analysis
Replies: 39
Views: 1884

Re: Dispelling the Myth of NNUE with LazySMP: An Analysis

While doing a lot of branchless avx2 computation versus all the conditional GP stuff of a classic eval - I would expect a decrease from HT improvement. Afaik HT profits most if threads have stalls due to branch misspredictions, dependency or cache misses, since the other hyper thread can keep the pi...
by Gerd Isenberg
Sun Dec 20, 2020 9:25 am
Forum: Computer Chess Club: General Topics
Topic: Can't access CPW website from Ukraine
Replies: 2
Views: 692

Re: Can't access CPW website from Ukraine

Hi Maksim,
i will forward your message to the site admin.
Regards,
Gerd
by Gerd Isenberg
Sat Nov 21, 2020 9:14 am
Forum: Computer Chess Club: General Topics
Topic: Chess opening database with names
Replies: 16
Views: 2042

Re: Chess opening database with names

Hi Gerd, the best what I found ... An opening book (literature) about B01. The author of the book (a Grandmaster) wrote to line x ... balanced, a bit better for white. And wrote to the same line (after move transposition) in the same book a bit later ... clear better position for white. But I like ...
by Gerd Isenberg
Fri Nov 20, 2020 8:07 pm
Forum: Computer Chess Club: General Topics
Topic: Chess opening database with names
Replies: 16
Views: 2042

Re: Chess opening database with names

Note that many such collections are inaccurate. I see games posted that have ECO classifications not even close to right. But even with the best kind of database, the classification system has some ambiguities. For example, the Semi-Tarrasch (D41/D42) has a lot of lines that overlap the Caro-Kann (...
by Gerd Isenberg
Wed Nov 18, 2020 4:58 pm
Forum: Computer Chess Club: General Topics
Topic: CPU Vector Unit, the new jam for NNs...
Replies: 8
Views: 1774

Re: CPU Vector Unit, the new jam for NNs...

What do the latest ARM CPU with AI acceleration offer in comparison? ARM has Advanced SIMD aka NEON , a combined 64- and 128-bit SIMD instruction set. But NEON is old. I heard that AI acceleration in recent ARM flagships especially in Huawei's Kirin are being showcased with far leading benchmarks i...
by Gerd Isenberg
Wed Nov 18, 2020 2:34 pm
Forum: Computer Chess Club: General Topics
Topic: CPU Vector Unit, the new jam for NNs...
Replies: 8
Views: 1774

Re: CPU Vector Unit, the new jam for NNs...

Jhoravi wrote:
Wed Nov 18, 2020 1:40 pm
What do the latest ARM CPU with AI acceleration offer in comparison?
ARM has Advanced SIMD aka NEON, a combined 64- and 128-bit SIMD instruction set.