LCZero Benchmarks
 Share
The version of the browser you are using is no longer supported. Please upgrade to a supported browser.Dismiss

 
%
123
 
 
 
 
 
 
 
 
 
ABCDEFGHIJKLMNOPQRSTUVWXYZ
1
ThreadsEngine version/typeSpeed npsNeural Net NameRemark
2
012lc0 v0.22.020x2569078611248
3
3 x RTX 2080 TI2lc0 v0.23.1+git.6837b8320x2568081442850-t 2 --backend=demux "--backend-opts=(backend=cudnn-fp16,gpu=0),(backend=cudnn-fp16,gpu=1),(backend=cudnn-fp16,gpu=2)" --nncache=2000000 --minibatch-size=1024; go nodes 5000000
4
2x RTX TITAN4lc0 v0.20 dev (with PR 619)20x25680000--threads=4 --backend=roundrobin --nncache=10000000 --cpuct=3.0 --minibatch-size=256 --max-collision-events=64 --max-prefetch=64 --backend-opts=(backend=cudnn-fp16,gpu=0),(backend=cudnn-fp16,gpu=1) go infinite; NPS checked after 100 seconds (peak was over 100k, then it starts dropping)
5
4xV1004lc0 cuda92 cudnn714 ubuntu20x2567870010040./lc0 --backend=multiplexing --backend-opts="x(backend=cudnnhalf,gpu=0,max_batch=512),y(backend=cudnnhalf,gpu=1,max_batch=512),yy(backend=cudnnhalf,gpu=2,max_batch=512),yyy(backend=cudnnhalf,gpu=3,max_batch=512)" --no-smart-pruning --minibatch-size=1024 --threads=4
6
RTX 2070 Super & 20704lc0 v0.20 dev (with PR 619)20x2567840132332same as above
7
RTX 2070 Super & 207012lc0 v0.22.020x2567605242425ba
8
2lc0-cudnn Batchsize=256 node-collisions=3220x25666723kb1-256x20-2100000
9
GTX 1070 @ stock2lc0 v0.20.320x2565415032930-t 2 --backend=cudnn-fp16 --minibatch-size=1024 --nncache=20000000; go nodes 5000000
10
mkl3lc0 v0.20.420x2565383811248--threads=3 --backend=roundrobin --nncache=10000000 --cpuct=3.0 --minibatch-size=256 --max-collision-events=64 --max-prefetch=64 --backend-opts=(backend=cudnn-fp16,gpu=0) go infinite; NPS checked at 100 seconds
11
2 x RTX 20604lc0 v0.22.020x2565241040685-t 4 -backend=demux -nncache=1000000 -minibatch-size=512 -max-prefetch=32 -backend-opts=(backend=cudnn-fp16,gpu=0),(backend=cudnn-fp16,gpu=1) go nodes 5000000
12
2 x RTX 20604lc0 v0.22.020x32052340T40B.2-106-t 4 -backend=demux -nncache=1000000 -minibatch-size=512 -max-prefetch=32 -backend-opts=(backend=cudnn-fp16,gpu=0),(backend=cudnn-fp16,gpu=1) go nodes 5000000
13
RTX Titan3lc0 v0.20.1-rc120x2565055832392--minibatch-size=512 -t 3 --backend=cudnn-fp16 --nncache=10000000; go infinite; note down NPS after 1 min
14
RTX 2080 Ti @ 338W (~1785 MHz)2
lc0 v0.20.2 (linux: fedora 29, 415.27, cuda 10.0, cudnn 7.4.2.24)
20x2565045632392-t 2 --backend=cudnn-fp16 --minibatch-size=1024 --nncache=20000000; go nodes 5000000
15
RTX Titan3
lc0 v0.20.2 (linux: fedora 29, 415.27, cuda 10.0, cudnn 7.4.2.24)
20x2563939032930--minibatch-size=512 -t 3 --backend=cudnn-fp16 --nncache=2000000; go nodes 5000000
16
RTX 2070 (slight OC, +66Mhz core)lc0-v0.19.1.1-windows-cuda.zip20x2563172332085--minibatch-size=1024 -t 2 --backend=multiplexing --backend-opts="x(backend=""cudnn-fp16"",gpu=0)" ; go nodes 1000000
17
TITAN V2lc0 v0.20.1-rc220x2563100410048--minibatch-size=512 -t 2 --backend=cudnn-fp16 --nncache=2000000; go nodes 1000000
18
RTX2080(laptop)2lc0 v0.22.020x25624142T40B.2-106lc0.exe -b cudnn-fp16 -w T40B.4-160;go nodes 100000
19
RTX 20604lc0-v0.18.1-windows-cuda10.0-cudnn7.3-for-2080.zip20x2562141311250.\lc0 --weights=weights_run1_11248.pb.gz --threads=4 --minibatch-size=256 --allowed-node-collisions=256 --cpuct=2.8 --nncache=10000000 --backend=multiplexing --backend-opts="(backend=cudnn,gpu=0),(backend=cudnn,gpu=1)"
20
GeForce GTX 780 Ti4lc0-v0.26.020x256110001541000Default Settings, backend = cudnn-fp16, NNCacheSize = 20000000, MiniBatchSize = 1024, MaxCollisionsEvents = 1024, Threads = 4
21
RX 5700 XT2lc0-v0.25.1-windows10-gpu-dx1220x2561098142850./lc0 benchmark
22
GTX 1080 TiLC0 V17.2 dev20x256920810954GPU load 98-99%, GPU Temp <= 82°C, Fanspeed 95%, go movetime 130 000
23
GTX 1650 SUPER2lc0-v0.24.1-windows-gpu-nvidia-cuda20x256877342850./lc0 benchmark --threads=2 --backend=cudnn-fp16 --minibatch-size=512 --movetime=50000
24
GTX 10802?800011250
25
2x GTX 1060 (6GB)1lc0-win-20180526 (cuda 9.2)20x2567596kb1-256x20-2000000-t 12 --backend=multiplexing "--backend-opts=(backend=cudnn-fp16,gpu=0),(backend=cudnn-fp16,gpu=1)" --nncache=2000000 --minibatch-size=1024; go nodes 5000000
26
GTX 1070Ti2lc0.20.1 (cuda)20x256649632965--nncache=8000000 --max-collision-events=256 --minibatch-size=256 --backend=multiplexing --cpuct=3.1
27
GTX 1070Ti2Lc0 v0.1720x256565711149--futile-search-aversion=0 (the equivalent of --no-smart-pruning) and otherwise default settings
28
GeForce GTX 1060 (3GB)lc0 v0.24.1 cuda 10.0.0 cudnn 7.4.220x2563365LS 14.3lc0 benchmark (v0.24.1)
29
GTX 1060 (6GB)2lc0 v0.24.1 (cuda 10.2)20x256335920x256SE-jj-9-53420000
30
GTX980m 10%underclock1lc0-win-20180522 (cuda 9.2)20x2562258kb1-256x20-2000000
31
GTX 9602lc0 v0.18.1 windows cuda20x256212311248
32
GTX980m 10%underclock220x2561855kb1-256x20-2000000--no-smart-pruning (value adjusted from 1093 to 1855 after rerun)
33
GTX 750 Ti @ 1350/1350 MHz1lc0 v0.18.1 windows cuda20x256165811258GPU Load 98%, GPU Temp 59°C, Fan Speed 39%, unknown command line flag --no-smart-pruning; -nncache 300000; go nodes 725000
34
AMD R9 Fury X (core + mem 20% overclock)
4v0.21.2-rc3 (Manjaro, OpenCL 2.1 AMD-APP (2841.4))20x256151842512result at 130.000 nodes, peaks at 550.000 nodes at 1880nps, then starts decreasing
35
GTX 9501v0.16.0 custom build for Debian (cuda 9.1.85)20x256150110687--no-smart-pruning --minibatch-size=1024 --threads=2
36
GTX 750 Ti @ stock2lc0 v0.18.1 windows cudaWhat is the "0"131411258GPU Load 100%, GPU Temp 50°C, Fan Speed 35%, unknown command line flag --no-smart-pruning
37
Nvidia Quadro K22002LC0 Ver 20.1 rc1 (Windows)20x256130432194
38
AMD RX480 (core + mem 20% overclock)4 v0.21.2-rc3 (Manjaro, OpenCL 2.1 AMD-APP (2841.4))20x256113342512result at 130.000 nodes, peaks at 540.000 nodes at 1385nps, then starts decreasing
39
MX1502LC0 Ver 17 RC2 (Windows)20x25680511045--minibatch-size=256 --backend=cudnn
40
Nvidia Geforce 840M4v0.21.2-rc3 (Manjaro, cuda 10.1.0, cudnn 7.5.1)20x25643342512
41
li16v0.21.2 OpenBlas (Windows)20x25628842699
42
GTX 4702lczero v0.10 OpenCL (Windows)20x25613010021
43
AMD Ryzen 3 1200 @stock4lczero v0.10 OpenBlas (Windows)20x2563510021tested using bumblebee + optirun, peaks at 570.000 nodes at 539nps, then starts decreasing, stays around 527-528nps.
44
TITAN V3lc0-win-20180708-cuda92-cudnn71420x25611248
45
Intel Core i7-4900MQ 4x 2,80GHz 500GB 16GB Quadro K2100M RW
LC0 Ver 20.1 rc1 (Windows)20x25632805This was performed at the starting position. Whole game average was 63k (see latest paper and talkchess clarifications)
46
(for cudnn client start with --no-smart-pruning flag and do "go nodes 130000" and wait until it finishes). For faster GPUs let it run for 1m or even 5m nodes.
×x
47
4x TPU20x256?--threads=4 --backend=roundrobin --nncache=10000000 --cpuct=3.0 --minibatch-size=256 --max-collision-events=64 --max-prefetch=64 --backend-opts=(backend=cudnn-fp16,gpu=0),(backend=cudnn-fp16,gpu=1) go infinite; NPS checked after 100 seconds (peak was over 100k, then it starts dropping)gp
48
CPU
49
GeForce GTX 780 Ti
50
Have clean up for the last time! after that i will close it!
51
LCZero Benchmark Nodes/sec. GPU & CPU
52
Please put your own bench scores here in sorted NPS order if you can. If you don't know what engine type, gpu is opencl and cpu is openblas + Networks ID !
53
RTX 2080 Ti @ 1290 MHz (~169W)2-t 2 --backend=cudnn-fp16 --minibatch-size=1024 --nncache=20000000; go nodes 5000000
54
Run go infinite from start position and abort after depth 30 and report NPS output.
55
TITAN XP220x25610751This benchmark is done in Arena Chess. I suspect on Cutechess-cli Leela will achieve an even higher NPS.
56
2lc0 v0.20.1-rc120x25646446--backend=cudnn-fp16 --nncache=10000000; go infinite; note down NPS after 1 min
57
2lc0 v0.18.13143311250
58
6lc0-v0.19.020x2562932931748lc0 -t 6 -w weights_31748.txt --backend=multiplexing "--backend-opts="a(backend=cudnn-fp16,gpu=0,minibatch-size=512,nncache=2000000),b(backend=cudnn,gpu=1)"
59
2lc0-v0.18.1-windows-cuda10.0-cudnn7.3-for-2080.zip2613511250?--futile-search-aversion=0 --minibatch-size=1024 -t 2 --backend-opts="x(backend=""cudnn-fp16"",gpu=0)"
60
2lc0-v0.20.2?2179732742--minibatch-size=512 -t 2 --backend=cudnn-fp16 --nncache=2000000; go nodes 1000000
61
4lc0-v0.18.120x2562141311248
62
4LC0 Ver 17 RC2 (Windows)700510970-t 4 --minibatch-size=512 --backend=multiplexing --backend-opts=(backend=cudnn,gpu=0,max_batch=1024),(backend=cudnn,gpu=1,max_batch=1024)
63
2LC0 Ver 17 RC2 (Windows)20x256359810970--threads=1 --fpu-reduction=0.2 --cpuct=1.2 --slowmover=1.5 --move-overhead=10 --no-smart-pruning
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
Loading...