Geekbench ML Score

639
Core ML Neural Engine Inference Score
Geekbench ML 0.6.0 for iOS AArch64

Result Information

User decimuschan
Upload Date August 22 2024 06:12 AM
Views 3

System Information

System Information
Operating System iOS 16.7.8
Model iPhone 8 Plus
Model IDiPhone10,5
Motherboard D211AP
CPU Information
Name Apple A11 Bionic
Topology 1 Processor, 6 Cores
Identifier ARM
Base Frequency 2.39 GHz
Cluster 12 Cores
Cluster 24 Cores
L1 Instruction Cache48.0 KB x 1
L1 Data Cache32.0 KB x 1
L2 Cache8.00 MB x 1
Memory Information
Size 2.90 GB
Inference Information
Framework Core ML
Backend Neural Engine
Device Default

Inference Performance

Workload Accuracy Score
Image Classification (F32)
100% 178
33.3 IPS
 
Image Classification (F16)
100% 143
26.8 IPS
 
Image Classification (I8)
99% 144
26.9 IPS
 
Image Segmentation (F32)
100% 188
3.13 IPS
 
Image Segmentation (F16)
100% 200
3.34 IPS
 
Image Segmentation (I8)
100% 199
3.32 IPS
 
Pose Estimation (F32)
100% 2255
2.73 IPS
 
Pose Estimation (F16)
100% 2746
3.32 IPS
 
Pose Estimation (I8)
100% 2524
3.06 IPS
 
Object Detection (F32)
100% 347
25.9 IPS
 
Object Detection (F16)
100% 346
25.8 IPS
 
Object Detection (I8)
97% 303
22.6 IPS
 
Face Detection (F32)
100% 932
11.1 IPS
 
Face Detection (F16)
100% 964
11.5 IPS
 
Face Detection (I8)
99% 985
11.7 IPS
 
Depth Estimation (F32)
100% 2509
19.5 IPS
 
Depth Estimation (F16)
100% 2794
21.7 IPS
 
Depth Estimation (I8)
99% 2335
18.1 IPS
 
Style Transfer (F32)
100% 4715
6.20 IPS
 
Style Transfer (F16)
100% 4924
6.48 IPS
 
Style Transfer (I8)
100% 4921
6.47 IPS
 
Image Super-Resolution (F32)
100% 1125
40.2 IPS
 
Image Super-Resolution (F16)
100% 997
35.6 IPS
 
Image Super-Resolution (I8)
100% 1148
41.0 IPS
 
Text Classification (F32)
100% 50
71.3 IPS
 
Text Classification (F16)
100% 201
288.6 IPS
 
Text Classification (I8)
96% 197
282.5 IPS
 
Machine Translation (F32)
100% 313
5.76 IPS
 
Machine Translation (F16)
100% 246
4.52 IPS
 
Machine Translation (I8)
99% 296
5.45 IPS
 

Compare

Set Baseline Geekbench 6 now available