FuryBench Testing Framework

home toIntegralPlentyChessObsidianCaissaQuanticadeSerendipitySpaghetSchoenemannClarityIlluminaPawnocchioVineAnura

multilayer_packus_precision_loss_simul vs multilayer_better diff

ELO    -3.37 +- 2.78 (95%)
CONF   N=20000 Threads=1 Hash=16MB
GAMES  N: 20226 W: 4752 L: 4948 D: 10526
PENTA  [273, 2436, 4866, 2290, 248]
AuthorSwedishChef
Creation1761058660
Dev Sourcehttps://github.com/aronpetko/vine
Dev Sha0e433e7e30345f197c740d6db8380adbec62ce08
Dev Branchmultilayer_packus_precision_loss_simul
Dev Bench696954
Dev Networkcombined_v48_p17_test1_multilayer_notranspose2
Dev OptionsThreads=1 Hash=16 Minimal=true
Dev TimeN=20000
Base Sourcehttps://github.com/aronpetko/vine
Base Shae3101ae88109cc28a3e09df0c742686a9d4f0245
Base Branchmultilayer_better
Base Bench695473
Base Networkcombined_v48_p17_test1_multilayer_notranspose2
Base OptionsThreads=1 Hash=16 Minimal=true
Base TimeN=20000
Opening BookUHO_Lichess_4852_v1.epd
Upload PGNsFALSE
Syzygy WDLDISABLED
Syzygy ADJ.OPTIONAL
Win ADJ.None
Draw ADJ.None
ScalingBASE 1000000 NPS
#UserGamesPentaPairsTimeCrash