oneDNN

This is a test of the Intel oneDNN as an Intel-optimized library for Deep Neural Networks and making use of its built-in benchdnn functionality. The result is the total perf time reported. Intel oneDNN was formerly known as DNNL (Deep Neural Network Library) and MKL-DNN before being rebranded as part of the Intel oneAPI toolkit.


oneDNN 3.4

Harness: Convolution Batch Shapes Auto - Engine: CPU

OpenBenchmarking.org metrics for this test profile configuration based on 258 public results since 1 March 2024 with the latest data as of 14 October 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
ms (Average)
Emerald Rapids [128 Cores / 256 Threads]
96th
10
0.3
Granite Rapids [128 Cores / 256 Threads]
96th
3
0.3
Sapphire Rapids [120 Cores / 240 Threads]
90th
5
0.4
Granite Rapids [256 Cores / 512 Threads]
87th
4
0.5
Zen 4 [192 Cores / 384 Threads]
85th
3
0.5
Zen 4 [192 Cores / 384 Threads]
85th
6
0.5
Zen 4 [64 Cores / 128 Threads]
82nd
4
0.6
Emerald Rapids [64 Cores / 128 Threads]
80th
6
0.6
Zen 4 [256 Cores / 512 Threads]
77th
5
0.6
Mid-Tier
75th
> 0.6
Sapphire Rapids [60 Cores / 120 Threads]
73rd
5
0.7
Zen 4 [96 Cores / 192 Threads]
72nd
3
0.7
Zen 4 [96 Cores / 192 Threads]
72nd
3
0.7
Zen 4 [64 Cores / 128 Threads]
65th
3
1.2
Zen 3 [24 Cores / 48 Threads]
63rd
4
2.7
Zen 4 [12 Cores / 24 Threads]
60th
4
4.0
Zen 4 [16 Cores / 32 Threads]
59th
4
4.2
Zen 2 [32 Cores / 64 Threads]
52nd
3
5.6 +/- 0.2
Zen 4 [16 Cores / 32 Threads]
51st
5
5.9 +/- 0.7
Median
50th
6.1
Zen 4 [12 Cores / 24 Threads]
50th
3
6.3 +/- 0.1
Raptor Lake [24 Cores / 32 Threads]
50th
5
6.3 +/- 0.7
Raptor Lake [8 Cores / 16 Threads]
48th
8
6.4
Zen 4 [12 Cores / 24 Threads]
44th
3
6.5
Zen 4 [16 Cores / 32 Threads]
44th
5
6.5 +/- 0.1
Sapphire Rapids [96 Cores / 192 Threads]
38th
3
7.7
Zen 4 [8 Cores / 16 Threads]
38th
3
7.8
Zen 4 [6 Cores / 12 Threads]
35th
3
7.9
Zen 4 [8 Cores / 16 Threads]
34th
4
8.2
Sapphire Rapids [112 Cores / 224 Threads]
33rd
8
8.2 +/- 1.0
Alder Lake [14 Cores / 20 Threads]
33rd
4
8.2
Zen 4 [6 Cores / 12 Threads]
33rd
3
8.2
Zen 4 [6 Cores / 12 Threads]
30th
4
8.3
Sapphire Rapids [56 Cores / 112 Threads]
28th
3
8.8 +/- 0.1
Tiger Lake [4 Cores / 8 Threads]
28th
4
8.8
Low-Tier
25th
> 9.0
Zen 4 [4 Cores / 8 Threads]
24th
4
9.1
Sapphire Rapids [48 Cores / 96 Threads]
22nd
3
9.2
Tiger Lake [4 Cores / 8 Threads]
20th
3
9.6 +/- 0.5
Cascade Lake [18 Cores / 36 Threads]
19th
3
9.9
Haswell [16 Cores / 16 Threads]
16th
3
10.6
Ice Lake [4 Cores / 8 Threads]
13th
4
13.2 +/- 0.6
Zen 3 [8 Cores / 16 Threads]
13th
4
13.7 +/- 0.2
Zen 4 [8 Cores / 16 Threads]
10th
4
15.1 +/- 0.1
Zen 4 [8 Cores / 16 Threads]
8th
4
16.1 +/- 0.1
Zen [32 Cores / 64 Threads]
7th
3
16.8 +/- 2.0
Zen 2 [6 Cores / 12 Threads]
6th
3
22.6 +/- 0.1
Zen 3 [8 Cores / 16 Threads]
4th
4
25.2 +/- 0.3
Zen 3 [8 Cores / 16 Threads]
3rd
3
30.8 +/- 0.2
Zen 2 [8 Cores / 8 Threads]
2nd
3
32.3 +/- 0.4