samsung SM-G990B2 samsung SM-G990B2 Difference
Single Precision 226 226 100.0%
samsung SM-G990B2
 
samsung SM-G990B2
 
Half Precision 224 224 100.0%
samsung SM-G990B2
 
samsung SM-G990B2
 
Quantized 12620 12620 100.0%
samsung SM-G990B2
 
samsung SM-G990B2
 
Geekbench AI 1.1.0 Geekbench AI 1.1.0

System Information

samsung SM-G990B2 samsung SM-G990B2
AI Framework TensorFlow Lite TensorFlow Lite
AI Backend QNN QNN
AI Device ARM SM8350 ARM SM8350
Operating System Android 14 Android 14
Model samsung SM-G990B2 samsung SM-G990B2
Processor ARM SM8350 @ 1.80 GHz
1 Processor, 8 Cores
ARM SM8350 @ 1.80 GHz
1 Processor, 8 Cores
Processor ID ARM implementer 65 architecture 8 variant 1 part 3396 revision 0 ARM implementer 65 architecture 8 variant 1 part 3396 revision 0
Motherboard lahaina lahaina
Memory 5.24 GB 5.24 GB

AI Performance

samsung SM-G990B2 samsung SM-G990B2 Difference
Image Classification (SP) 155 155 100.0%
samsung SM-G990B2
 
samsung SM-G990B2
 
Image Classification (HP) 158 158 100.0%
samsung SM-G990B2
 
samsung SM-G990B2
 
Image Classification (Q) 12862 12862 100.0%
samsung SM-G990B2
 
samsung SM-G990B2
 
Image Segmentation (SP) 204 204 100.0%
samsung SM-G990B2
 
samsung SM-G990B2
 
Image Segmentation (HP) 171 171 100.0%
samsung SM-G990B2
 
samsung SM-G990B2
 
Image Segmentation (Q) 12378 12378 100.0%
samsung SM-G990B2
 
samsung SM-G990B2
 
Pose Estimation (SP) 227 227 100.0%
samsung SM-G990B2
 
samsung SM-G990B2
 
Pose Estimation (HP) 235 235 100.0%
samsung SM-G990B2
 
samsung SM-G990B2
 
Pose Estimation (Q) 91517 91517 100.0%
samsung SM-G990B2
 
samsung SM-G990B2
 
Object Detection (SP) 126 126 100.0%
samsung SM-G990B2
 
samsung SM-G990B2
 
Object Detection (HP) 125 125 100.0%
samsung SM-G990B2
 
samsung SM-G990B2
 
Object Detection (Q) 5473 5473 100.0%
samsung SM-G990B2
 
samsung SM-G990B2
 
Face Detection (SP) 320 320 100.0%
samsung SM-G990B2
 
samsung SM-G990B2
 
Face Detection (HP) 325 325 100.0%
samsung SM-G990B2
 
samsung SM-G990B2
 
Face Detection (Q) 27061 27061 100.0%
samsung SM-G990B2
 
samsung SM-G990B2
 
Depth Estimation (SP) 275 275 100.0%
samsung SM-G990B2
 
samsung SM-G990B2
 
Depth Estimation (HP) 299 299 100.0%
samsung SM-G990B2
 
samsung SM-G990B2
 
Depth Estimation (Q) 33182 33182 100.0%
samsung SM-G990B2
 
samsung SM-G990B2
 
Style Transfer (SP) 492 492 100.0%
samsung SM-G990B2
 
samsung SM-G990B2
 
Style Transfer (HP) 477 477 100.0%
samsung SM-G990B2
 
samsung SM-G990B2
 
Style Transfer (Q) 101271 101271 100.0%
samsung SM-G990B2
 
samsung SM-G990B2
 
Image Super-Resolution (SP) 159 159 100.0%
samsung SM-G990B2
 
samsung SM-G990B2
 
Image Super-Resolution (HP) 142 142 100.0%
samsung SM-G990B2
 
samsung SM-G990B2
 
Image Super-Resolution (Q) 30824 30824 100.0%
samsung SM-G990B2
 
samsung SM-G990B2
 
Text Classification (SP) 193 193 100.0%
samsung SM-G990B2
 
samsung SM-G990B2
 
Text Classification (HP) 193 193 100.0%
samsung SM-G990B2
 
samsung SM-G990B2
 
Text Classification (Q) 2153 2153 100.0%
samsung SM-G990B2
 
samsung SM-G990B2
 
Machine Translation (SP) 302 302 100.0%
samsung SM-G990B2
 
samsung SM-G990B2
 
Machine Translation (HP) 318 318 100.0%
samsung SM-G990B2
 
samsung SM-G990B2
 
Machine Translation (Q) 213 213 100.0%
samsung SM-G990B2
 
samsung SM-G990B2