Search found 507 matches

by D Sceviour
Wed Mar 11, 2020 2:20 am
Forum: Computer Chess Club: Programming and Technical Discussions
Topic: Removing Large Arrays
Replies: 36
Views: 5172

Re: Removing Large Arrays

Often, I test the speed of different function methods. The results indicate the fixed array PushToEdges[] seems to be faster on average: 10 second test ... PushToEdges 16519006400 push_to_edge 16610155136 The optimizer can remove what we are testing because the results are unused. (warning: variabl...
by D Sceviour
Tue Mar 10, 2020 9:00 pm
Forum: Computer Chess Club: Programming and Technical Discussions
Topic: Removing Large Arrays
Replies: 36
Views: 5172

Re: Removing Large Arrays

Often, I test the speed of different function methods. The results indicate the fixed array PushToEdges[] seems to be faster on average: 10 second test ... PushToEdges 16519006400 push_to_edge 16610155136 In a real chess program with many instances fighting for the caches, things may turn out diffe...
by D Sceviour
Tue Mar 10, 2020 7:55 pm
Forum: Computer Chess Club: Programming and Technical Discussions
Topic: Removing Large Arrays
Replies: 36
Views: 5172

Re: Removing Large Arrays

Often, I test the speed of different function methods. The results indicate the fixed array PushToEdges[] seems to be faster on average: 10 second test ... PushToEdges 16519006400 push_to_edge 16610155136 nonsense, the compiler completely eliminates the inner loops, so it boils down to nothing (bus...
by D Sceviour
Tue Mar 10, 2020 7:30 pm
Forum: Computer Chess Club: Programming and Technical Discussions
Topic: Removing Large Arrays
Replies: 36
Views: 5172

Re: Removing Large Arrays

Often, I test the speed of different function methods. The results indicate the fixed array PushToEdges[] seems to be faster on average: 10 second test ... PushToEdges 16519006400 push_to_edge 16610155136 // gcc -O3 -Wall -m64 test.c -o test -lm -s // speed test method for different functions #inclu...
by D Sceviour
Wed Feb 19, 2020 4:37 pm
Forum: Computer Chess Club: Programming and Technical Discussions
Topic: SIMD methods in TT probing and replacement
Replies: 15
Views: 1721

Re: SIMD methods in TT probing and replacement

It is increasingly difficult to cram everything into a long integer. How about a bigger hash entry? The first entry is an unsigned 64-bit hash key. This cannot be shortened mainly because this is about the minimum required to create a unique key for a standard 8x8 board in chess, and to minimize the...
by D Sceviour
Mon Feb 17, 2020 4:43 pm
Forum: Computer Chess Club: Programming and Technical Discussions
Topic: Zobrist key independence
Replies: 38
Views: 4074

Re: Zobrist key independence

I assume this was in response to the post on Hamming Distance. If the hamming distance of the zobrist keys are zero, then there is a likelihood of hash key failure, especially with lock-less hashing. As a curiosity, I tested the polyglot keys and found: minimum hamming distance = 18 maximum hamming ...
by D Sceviour
Sun Feb 16, 2020 9:49 pm
Forum: Computer Chess Club: Programming and Technical Discussions
Topic: hash collisions
Replies: 370
Views: 28483

Re: hash collisions

It has not been suggested here yet, but a consideration to look at for hash collision testing is the Hamming Distance https://www.chessprogramming.org/Population_Count#HammingDistance . Some programs like Crafty and Fruit use preset hash keys, but other engines are generating random numbers for ever...
by D Sceviour
Fri Feb 14, 2020 3:54 pm
Forum: Computer Chess Club: Programming and Technical Discussions
Topic: Idea in move ordering ...
Replies: 7
Views: 1664

Re: Idea in move ordering ...

The idea is called a piece-square table, often seen as PST() or PSQT(). Programmers prefer to use weights rather than percentages, because weights can easily be configured for final position evaluations. Percentages could be calculated from most piece-square tables by taking the maximum weight as 10...
by D Sceviour
Sat Feb 01, 2020 4:00 am
Forum: Computer Chess Club: Programming and Technical Discussions
Topic: hash collisions
Replies: 370
Views: 28483

Re: hash collisions

I recently got a note from TCEC team and they were basically saying that since they are going to be running on 176 cores for the next season, they are concerned that engines be able to handle this well. They mention this testing technique suggested by the Demmolito author: I found a crashing bug li...