Geekbench AI Score

267
Single Precision Score
270
Half Precision Score
660
Quantized Score
Geekbench AI 1.0.0 for Android AArch64

Result Information

Upload Date August 21 2024 10:44 PM
Views 3

System Information

AI Information
Framework TensorFlow Lite
Backend NNAPI
Device HiSilicon Kirin 9000
System Information
Operating System Android 12
Model Huawei Mate 40 Pro
Model IDHUAWEI NOH-NX9
Motherboard NOH
Governorschedutil
CPU Information
Name ARM ARMv8
Topology 1 Processor, 8 Cores
Identifier ARM implementer 65 architecture 8 variant 1 part 3341 revision 0
Base Frequency 2.04 GHz
Cluster 14 Cores @ 2.04 GHz
Cluster 23 Cores @ 2.54 GHz
Cluster 31 Core @ 3.13 GHz
Memory Information
Size 7.25 GB

AI Performance

Workload Accuracy Score
Image Classification (SP)
100% 165
30.8 IPS
 
Image Classification (HP)
100% 166
30.9 IPS
 
Image Classification (Q)
97% 431
80.4 IPS
 
Image Segmentation (SP)
100% 220
3.57 IPS
 
Image Segmentation (HP)
100% 222
3.60 IPS
 
Image Segmentation (Q)
98% 449
7.31 IPS
 
Pose Estimation (SP)
100% 269
0.31 IPS
 
Pose Estimation (HP)
100% 268
0.31 IPS
 
Pose Estimation (Q)
98% 1112
1.30 IPS
 
Object Detection (SP)
100% 165
13.1 IPS
 
Object Detection (HP)
100% 166
13.2 IPS
 
Object Detection (Q)
89% 476
38.1 IPS
 
Face Detection (SP)
100% 411
4.88 IPS
 
Face Detection (HP)
100% 417
4.95 IPS
 
Face Detection (Q)
97% 944
11.3 IPS
 
Depth Estimation (SP)
100% 334
2.57 IPS
 
Depth Estimation (HP)
100% 335
2.58 IPS
 
Depth Estimation (Q)
73% 951
7.78 IPS
 
Style Transfer (SP)
100% 544
0.70 IPS
 
Style Transfer (HP)
100% 548
0.71 IPS
 
Style Transfer (Q)
98% 1728
2.23 IPS
 
Image Super-Resolution (SP)
100% 172
6.35 IPS
 
Image Super-Resolution (HP)
100% 173
6.38 IPS
 
Image Super-Resolution (Q)
97% 659
24.4 IPS
 
Text Classification (SP)
100% 235
314.2 IPS
 
Text Classification (HP)
100% 244
325.2 IPS
 
Text Classification (Q)
91% 426
573.1 IPS
 
Machine Translation (SP)
100% 387
6.67 IPS
 
Machine Translation (HP)
100% 391
6.74 IPS
 
Machine Translation (Q)
57% 356
8.71 IPS
 

Compare

Set Baseline Geekbench 6 now available