Geekbench ML Score

4475
Core ML GPU Inference Score
Geekbench ML 0.6.0 for iOS AArch64

Result Information

Upload Date August 15 2024 02:28 PM
Views 1

System Information

System Information
Operating System iOS 18.0
Model iPhone 15 Pro Max
Model IDiPhone16,2
Motherboard D84AP
CPU Information
Name Apple A17 Pro
Topology 1 Processor, 6 Cores
Identifier ARM
Base Frequency 3.78 GHz
Cluster 12 Cores
Cluster 24 Cores
L1 Instruction Cache128 KB x 1
L1 Data Cache64.0 KB x 1
L2 Cache4.00 MB x 1
Memory Information
Size 7.48 GB
Inference Information
Framework Core ML
Backend GPU
Device Default

Inference Performance

Workload Accuracy Score
Image Classification (F32)
100% 1615
302.2 IPS
 
Image Classification (F16)
100% 1489
278.6 IPS
 
Image Classification (I8)
99% 1715
321.0 IPS
 
Image Segmentation (F32)
100% 1359
22.7 IPS
 
Image Segmentation (F16)
100% 1552
25.9 IPS
 
Image Segmentation (I8)
100% 1569
26.2 IPS
 
Pose Estimation (F32)
100% 14667
17.8 IPS
 
Pose Estimation (F16)
100% 15526
18.8 IPS
 
Pose Estimation (I8)
100% 15095
18.3 IPS
 
Object Detection (F32)
100% 1459
109.0 IPS
 
Object Detection (F16)
100% 1464
109.3 IPS
 
Object Detection (I8)
97% 1880
140.4 IPS
 
Face Detection (F32)
100% 6203
73.8 IPS
 
Face Detection (F16)
98% 8530
101.4 IPS
 
Face Detection (I8)
97% 8676
103.2 IPS
 
Depth Estimation (F32)
100% 12231
94.9 IPS
 
Depth Estimation (F16)
100% 13568
105.2 IPS
 
Depth Estimation (I8)
99% 12600
97.7 IPS
 
Style Transfer (F32)
100% 29047
38.2 IPS
 
Style Transfer (F16)
100% 35139
46.2 IPS
 
Style Transfer (I8)
100% 35760
47.0 IPS
 
Image Super-Resolution (F32)
100% 7035
251.2 IPS
 
Image Super-Resolution (F16)
100% 6382
227.9 IPS
 
Image Super-Resolution (I8)
100% 6678
238.5 IPS
 
Text Classification (F32)
100% 2846
4.09 KIPS
 
Text Classification (F16)
100% 4821
6.93 KIPS
 
Text Classification (I8)
96% 4823
6.93 KIPS
 
Machine Translation (F32)
100% 630
11.6 IPS
 
Machine Translation (F16)
100% 630
11.6 IPS
 
Machine Translation (I8)
99% 645
11.9 IPS
 

Compare

Set Baseline Geekbench 6 now available