issues
search
rn5f107s2
/
Molybdenum
Molybdenum is a chess engine written in C++.
GNU General Public License v3.0
7
stars
2
forks
source link
issues
Newest
Newest
Most commented
Recently updated
Oldest
Least commented
Least recently updated
NNUE file
#97
gabrielp10031978
opened
3 weeks ago
3
Github workflow will store the windows exe
#96
aytchell
closed
3 weeks ago
0
Compile github workflow on windows
#95
aytchell
closed
3 weeks ago
0
Add github workflow which compiles stuff
#94
aytchell
closed
3 weeks ago
1
Implement lazySMP
#93
rn5f107s2
closed
2 months ago
0
Refactor UCI handling
#92
rn5f107s2
closed
3 months ago
0
Make the net as well as histories non global
#91
rn5f107s2
closed
3 months ago
0
Fixed construction of thread
#90
aytchell
closed
3 months ago
0
Extend research if all prior moves were bad
#89
rn5f107s2
closed
3 months ago
1
Move eval divisor to where eval scale is applied
#88
rn5f107s2
closed
3 months ago
0
Dont evaluate before ttCutoffs
#87
rn5f107s2
closed
3 months ago
0
Fix quantization of output bias
#86
rn5f107s2
closed
3 months ago
0
Research PVNode LMR fail highs with full depth but zero window first
#85
rn5f107s2
closed
3 months ago
0
New net trained using wdl 0.5
#84
rn5f107s2
closed
3 months ago
0
Prefetch the TTEntry after a nullmove
#83
rn5f107s2
closed
3 months ago
0
Update killers in a different way at root
#82
rn5f107s2
closed
3 months ago
0
Dont singular extend all moves in a node if singularsearch failed low
#81
rn5f107s2
closed
4 months ago
0
Net with additional ~100m fens
#80
rn5f107s2
closed
4 months ago
0
Add another 13m fens
#79
rn5f107s2
closed
5 months ago
0
Add another ~75m Positions
#78
rn5f107s2
closed
5 months ago
0
Increase the Hiddenlayer to 256
#77
rn5f107s2
closed
5 months ago
0
New net using the SCReLU activation function
#76
rn5f107s2
closed
5 months ago
0
Please add ARMv9.2, SVE2, Dot Product and Neon support
#75
Chess321
closed
4 months ago
2
Additional ~20m fens
#74
rn5f107s2
closed
5 months ago
0
Remove redundant accumulator update
#73
rn5f107s2
closed
5 months ago
0
New net with an additional ~80m fens
#72
rn5f107s2
closed
5 months ago
0
New net using an additional 40m fens
#71
rn5f107s2
closed
5 months ago
0
Default make target should build Datagen
#70
aytchell
closed
5 months ago
0
New net with increased hidden layer size
#69
rn5f107s2
closed
5 months ago
0
MCP more when not improving
#68
rn5f107s2
closed
6 months ago
0
Dont singular extend at root
#67
rn5f107s2
closed
6 months ago
0
Adjust Se margin based on root movecount
#66
rn5f107s2
closed
6 months ago
0
Adjust returned score based on ttScore
#65
rn5f107s2
closed
6 months ago
0
Tuned at 1200 Iterations 40+0.4, excludung some parametetrs
#64
rn5f107s2
closed
6 months ago
0
Also apply fractional depth to fp
#63
rn5f107s2
closed
6 months ago
0
Subtract quarter reductions of previous ply from mcp depth
#62
rn5f107s2
closed
6 months ago
0
Deduct fractional depths from history pruning margin
#61
rn5f107s2
closed
6 months ago
0
Tune timemanagment at 20+0.2 for 1.5k Iterations
#60
rn5f107s2
closed
6 months ago
0
If our opponents move worsened their static eval it was probably a ba…
#59
rn5f107s2
closed
7 months ago
0
Added make target "datagen"
#58
aytchell
closed
7 months ago
0
Fix oversight that only applied tuned hist limits to bonuses and kept…
#57
rn5f107s2
closed
7 months ago
0
Store TTBound and depth as 8 bit integer saving a few bytes of the tt…
#56
rn5f107s2
closed
7 months ago
0
If the new entry has the same key as the old one, prefer the one with…
#55
rn5f107s2
closed
7 months ago
0
If the ttMove is a cpature use it for ordering in qs
#54
rn5f107s2
closed
7 months ago
0
Probe the TT in qsearch and cut-off arcodingly
#53
rn5f107s2
closed
7 months ago
0
If a move failed singular search high, sort it right after the ttMove
#52
rn5f107s2
closed
7 months ago
0
Singular Extensions,
#51
rn5f107s2
closed
7 months ago
0
Refactor movepicker
#50
rn5f107s2
closed
7 months ago
0
Fix compiling errors on windows
#49
rn5f107s2
closed
8 months ago
0
use clang++
#48
aytchell
closed
8 months ago
0
Next