Geekbench ML Score

955
TensorFlow Lite GPU Inference Score
Geekbench ML 0.5.0 for Android AArch64

Result Information

Upload Date January 19 2022 09:01 AM
Views 1

System Information

System Information
Operating System Android 12
Model samsung SM-G970N
Model IDsamsung SM-G970N
Motherboard exynos9820
Governorschedutil
CPU Information
Name ARMv8
Topology 1 Processor, 8 Cores
Identifier ARM implementer 83 architecture 8 variant 1 part 3 revision 0
Base Frequency 1.95 GHz
Cluster 14 Cores @ 1.95 GHz
Cluster 22 Cores @ 2.31 GHz
Cluster 32 Cores @ 2.73 GHz
Memory Information
Size 5.29 GB
Inference Information
Framework TensorFlow Lite
Backend GPU

Inference Performance

Workload Accuracy Score
Image Classification (F32)
100% 603
64.6 IPS
 
Image Classification (F16)
100% 861
92.2 IPS
 
Image Classification (I8)
80% 727
77.9 IPS
 
Image Segmentation (F32)
100% 1703
10.5 IPS
 
Image Segmentation (F16)
100% 3045
18.8 IPS
 
Image Segmentation (I8)
89% 3421
21.1 IPS
 
Pose Estimation (F32)
100% 432
1.75 IPS
 
Pose Estimation (F16)
100% 1969
7.97 IPS
 
Pose Estimation (I8)
80% 1866
7.55 IPS
 
Object Detection (F32)
100% 369
12.9 IPS
 
Object Detection (F16)
100% 552
19.4 IPS
 
Object Detection (I8)
54% 504
17.7 IPS
 
Face Detection (F32)
100% 1261
14.5 IPS
 
Face Detection (F16)
97% 2065
23.7 IPS
 
Face Detection (I8)
75% 1668
19.2 IPS
 
Text Classification (F32)
100% 496
196.0 IPS
 
Text Classification (F16)
98% 611
241.7 IPS
 
Machine Translation (F32)
100% 544
83.8 IPS
 
Machine Translation (F16)
100% 534
82.3 IPS
 

Compare

Set Baseline

Share

Geekbench ML now available for download