Geekbench ML Score

9878
Core ML Neural Engine Inference Score
Geekbench ML 0.6.0 for iOS AArch64

Result Information

Upload Date July 18 2024 09:09 PM
Views 1

System Information

System Information
Operating System iOS 18.0
Model iPad16,6
Model IDiPad16,6
Motherboard J721AP
CPU Information
Name ARM
Topology 1 Processor, 10 Cores
Identifier ARM
Base Frequency 4.40 GHz
Cluster 14 Cores
Cluster 26 Cores
L1 Instruction Cache128 KB x 1
L1 Data Cache64.0 KB x 1
L2 Cache4.00 MB x 1
Memory Information
Size 15.07 GB
Inference Information
Framework Core ML
Backend Neural Engine
Device Default

Inference Performance

Workload Accuracy Score
Image Classification (F32)
100% 3966
742.2 IPS
 
Image Classification (F16)
100% 9344
1.75 KIPS
 
Image Classification (I8)
99% 10521
1.97 KIPS
 
Image Segmentation (F32)
100% 1998
33.4 IPS
 
Image Segmentation (F16)
100% 2448
40.9 IPS
 
Image Segmentation (I8)
100% 2386
39.9 IPS
 
Pose Estimation (F32)
100% 26575
32.2 IPS
 
Pose Estimation (F16)
100% 102471
124.1 IPS
 
Pose Estimation (I8)
100% 105872
128.2 IPS
 
Object Detection (F32)
100% 3238
241.8 IPS
 
Object Detection (F16)
100% 6150
459.1 IPS
 
Object Detection (I8)
97% 3342
249.5 IPS
 
Face Detection (F32)
100% 11201
133.2 IPS
 
Face Detection (F16)
97% 16119
191.7 IPS
 
Face Detection (I8)
96% 15522
184.6 IPS
 
Depth Estimation (F32)
100% 19510
151.3 IPS
 
Depth Estimation (F16)
100% 49169
381.3 IPS
 
Depth Estimation (I8)
37% 45717
354.5 IPS
 
Style Transfer (F32)
100% 53851
70.8 IPS
 
Style Transfer (F16)
100% 146493
192.7 IPS
 
Style Transfer (I8)
100% 146967
193.3 IPS
 
Image Super-Resolution (F32)
100% 8018
286.3 IPS
 
Image Super-Resolution (F16)
100% 22993
821.1 IPS
 
Image Super-Resolution (I8)
100% 22557
805.6 IPS
 
Text Classification (F32)
100% 1364
1.96 KIPS
 
Text Classification (F16)
100% 4694
6.75 KIPS
 
Text Classification (I8)
96% 904
1.30 KIPS
 
Machine Translation (F32)
100% 1059
19.5 IPS
 
Machine Translation (F16)
100% 2272
41.8 IPS
 
Machine Translation (I8)
99% 961
17.7 IPS
 

Compare

Set Baseline Geekbench 6 now available