Geekbench ML Score

384
TensorFlow Lite CPU Inference Score
Geekbench ML 0.5.0 for Android AArch64

Result Information

Upload Date October 23 2021 03:26 AM
Views 298

System Information

System Information
Operating System Android 11
Model Samsung Galaxy S21 5G
Model IDsamsung SM-G991N
Motherboard exynos2100
Governorenergy_step
CPU Information
Name ARM ARMv8
Topology 1 Processor, 8 Cores
Identifier ARM implementer 65 architecture 8 variant 1 part 3396 revision 0
Base Frequency 2.21 GHz
Cluster 14 Cores @ 2.21 GHz
Cluster 23 Cores @ 2.81 GHz
Cluster 31 Core @ 2.91 GHz
Memory Information
Size 7.02 GB
Inference Information
Framework TensorFlow Lite
Backend CPU

Inference Performance

Workload Accuracy Score
Image Classification (F32)
100% 357
38.3 IPS
 
Image Classification (F16)
100% 339
36.4 IPS
 
Image Classification (I8)
81% 162
17.4 IPS
 
Image Segmentation (F32)
100% 541
3.34 IPS
 
Image Segmentation (F16)
100% 545
3.37 IPS
 
Image Segmentation (I8)
89% 893
5.52 IPS
 
Pose Estimation (F32)
100% 269
1.09 IPS
 
Pose Estimation (F16)
100% 270
1.09 IPS
 
Pose Estimation (I8)
80% 224
0.90 IPS
 
Object Detection (F32)
100% 339
11.9 IPS
 
Object Detection (F16)
100% 341
12.0 IPS
 
Object Detection (I8)
52% 261
9.15 IPS
 
Face Detection (F32)
100% 548
6.30 IPS
 
Face Detection (F16)
100% 552
6.34 IPS
 
Face Detection (I8)
75% 556
6.39 IPS
 
Text Classification (F32)
100% 321
126.9 IPS
 
Text Classification (F16)
100% 330
130.4 IPS
 
Machine Translation (F32)
100% 543
83.6 IPS
 
Machine Translation (F16)
100% 541
83.3 IPS
 

Compare

Set Baseline

Share

Geekbench ML now available for download