Search found 158 matches

by abulmo2
Thu Jan 16, 2020 3:49 pm
Forum: Computer Chess Club: Programming and Technical Discussions
Topic: Exploring an unusual evaluation function
Replies: 19
Views: 614

Re: Exploring an unusual evaluation function

n terms of style of play, restricting the evaluation of a position to the number of moves and no more, results in some disastrous sacrifices. For example a rook with no moves immediately available is considered effectively worthless. I anticipated this but thought that maybe it’s potential might be...
by abulmo2
Thu Jan 16, 2020 12:24 am
Forum: Computer Chess Club: General Topics
Topic: New engine releases 2020
Replies: 51
Views: 5268

Re: New engine releases 2020

I just release Amoeba 3.1. It fixes a few bugs and probably adds a bunch of others. It should be about 30 Elo stronger than the previous version
https://github.com/abulmo/amoeba/releases/tag/v3.1
Have fun with it.
by abulmo2
Tue Jan 14, 2020 9:50 pm
Forum: Computer Chess Club: Programming and Technical Discussions
Topic: Board adaptive / tuning evaluation function - no NN/AI
Replies: 14
Views: 648

Re: Board adaptive / tuning evaluation function - no NN/AI

It sounds like the oracle approach. It was successfull in the 1990s, but was abandonned in favor of tapered evaluation.
by abulmo2
Sun Dec 29, 2019 7:16 pm
Forum: Computer Chess Club: Programming and Technical Discussions
Topic: Microsoft Visual C vs Gcc on Windows
Replies: 4
Views: 429

Re: Microsoft Visual C vs Gcc on Windows

PGO can also produce faster binary with gcc.
by abulmo2
Fri Dec 20, 2019 3:19 pm
Forum: Computer Chess Club: General Topics
Topic: New engine releases 2019
Replies: 460
Views: 114955

Re: New engine releases 2019

Damir wrote:
Thu Dec 19, 2019 7:58 pm
I suppose nobody here is dumb enough to use it..... :lol: :lol:
Dumb, but fast :-)
If you want to see high nodes per second on a single score, you should give it a try.
by abulmo2
Fri Dec 20, 2019 9:26 am
Forum: Computer Chess Club: General Topics
Topic: A complete 1200 lines of code engine (Dumb 1.4)
Replies: 6
Views: 497

Re: A complete 1200 lines of code engine (Dumb 1.4)

A pity that there seems to be no, or no reliable package for compiling D in Msys2, which I use now exclusively for compiling C and C++ (with GCC and G++). My previous investigations showed that GDC asked for trouble and isn't recommended and may be even dead since years. GDC is not dead at all. It ...
by abulmo2
Fri Dec 20, 2019 8:40 am
Forum: Computer Chess Club: General Topics
Topic: A complete 1200 lines of code engine (Dumb 1.4)
Replies: 6
Views: 497

Re: A complete 1200 lines of code engine (Dumb 1.4)

Richard could you please add a non popcount pre-SSE4.2 build too? I replaced the executable in the windows zip file, targeting core2 CPU. Dumb does not use popcount, and the raw speed does not seem affected (tested on an old core i7). Hope it works on your CPU. (A pity I have never compiled D befor...
by abulmo2
Wed Dec 18, 2019 11:27 pm
Forum: Computer Chess Club: General Topics
Topic: A complete 1200 lines of code engine (Dumb 1.4)
Replies: 6
Views: 497

A complete 1200 lines of code engine (Dumb 1.4)

I have updated Dumb to the version 1.4. The code is slightly more compact (1198* lines of code), and the strength has significantly improved. I guess it should be between 2300 - 2400 elo on the CCRL 40/4 list. As we are approaching the year 2020, with the end of windows 7, I only provide 64-bit comp...
by abulmo2
Mon Dec 16, 2019 8:10 pm
Forum: Computer Chess Club: General Topics
Topic: Evaluation challenge II
Replies: 6
Views: 469

Re: Evaluation challenge II

Much harder for Amoeba 3.0. With 1 threads, it found f2f3 at depth 26 (and 35s) then sticked to it. With 4 threads it found it at depth 25 (and 22s). The development version (future Amoeba 3.1) found it at depth 22 in 6s (4 threads also). My weaker program Dumb 1.3 found it at depth 14 (but 60s) and...
by abulmo2
Mon Dec 16, 2019 1:40 pm
Forum: Computer Chess Club: General Topics
Topic: Evaluation challenge
Replies: 26
Views: 1285

Re: Evaluation challenge

Amoeba 3.0 running on 1 thread with default settings, finds it at depth 7 and stay stick to it for several minutes. info depth 7 seldepth 19 score cp -60 nps 1045263 time 13 nodes 14135 hashfull 2 pv d1d3 c4d3 g2f1 b8c8 f1d3 a7a6 h1g1 ... info depth 28 seldepth 50 score cp 26 nps 1771418 time 245257...