LeelaChessZero

LeelaChessZero (lc0 / lczero) is a chess engine automated vian neural networks. This test profile can be used for OpenCL, CUDA + cuDNN, and BLAS (CPU-based) benchmarking.


LeelaChessZero 0.30

Backend: BLAS

OpenBenchmarking.org metrics for this test profile configuration based on 162 public results since 16 December 2023 with the latest data as of 14 June 2024.

Below is an overview of the generalized performance for components where there is sufficient statistically significant data based upon user-uploaded results. It is important to keep in mind particularly in the Linux/open-source space there can be vastly different OS configurations, with this overview intended to offer just general guidance as to the performance expectations.

Component
Details
Percentile Rank
# Compatible Public Results
Nodes Per Second (Average)
Zen 4 [96 Cores / 192 Threads]
99th
4
864 +/- 23
Zen 4 [192 Cores / 384 Threads]
99th
4
862 +/- 10
Zen 4 [64 Cores / 128 Threads]
93rd
5
524 +/- 21
Zen 4 [64 Cores / 128 Threads]
87th
5
331 +/- 21
Zen 4 [16 Cores / 32 Threads]
81st
5
209 +/- 4
Zen 3 [24 Cores / 48 Threads]
80th
4
208 +/- 24
Zen 4 [6 Cores / 12 Threads]
79th
4
184
Mid-Tier
75th
< 180
Zen 4 [6 Cores / 12 Threads]
72nd
4
161 +/- 1
Zen 2 [8 Cores / 16 Threads]
68th
4
132 +/- 2
Zen 2 [64 Cores / 128 Threads]
64th
5
115 +/- 4
Cascade Lake [18 Cores / 36 Threads]
60th
3
104 +/- 4
Zen 4 [8 Cores / 16 Threads]
58th
3
94
Tiger Lake [4 Cores / 8 Threads]
52nd
6
70 +/- 6
Emerald Rapids [64 Cores / 128 Threads]
51st
3
67 +/- 1
Median
50th
67
Tiger Lake [4 Cores / 8 Threads]
47th
4
62 +/- 1
Zen 2 [8 Cores / 8 Threads]
43rd
3
60 +/- 1
Raptor Lake [24 Cores / 32 Threads]
38th
6
49 +/- 3
Zen 3 [8 Cores / 16 Threads]
30th
3
42 +/- 2
Cascade Lake [16 Cores / 32 Threads]
27th
3
37 +/- 1
Kaby/Coffee/Whiskey Lake [2 Cores / 8 Threads]
26th
3
34
Low-Tier
25th
< 34
Zen 3 [8 Cores / 16 Threads]
21st
3
31 +/- 2
Zen 3 [8 Cores / 16 Threads]
21st
4
30
Raptor Lake [4 Cores / 8 Threads]
17th
3
24
Raptor Lake [14 Cores / 20 Threads]
14th
3
19