Geekbench AI Score

1129
Single Precision Score
1136
Half Precision Score
1889
Quantized Score
Geekbench AI 1.2.0 for Android AArch64

Result Information

Upload Date May 23 2025 01:59 AM
Views 3

System Information

AI Information
Framework TensorFlow Lite
Backend CPU
Device ARMv8
System Information
Operating System Android 12
Model HUAWEI PCE-W29
Model IDHUAWEI PCE-W29
Motherboard PCEL
CPU Information
Name ARMv8
Topology 1 Processor, 8 Cores
Identifier ARM implementer 72 architecture 8 variant 2 part 3330 revision 2
Base Frequency 1.53 GHz
Cluster 14 Cores @ 1.53 GHz
Cluster 23 Cores @ 2.15 GHz
Cluster 31 Core @ 2.49 GHz
Memory Information
Size 11.54 GB

AI Performance

Workload Accuracy Score
Image Classification (SP)
100% 481
89.4 IPS
 
Image Classification (HP)
100% 504
93.7 IPS
 
Image Classification (Q)
100% 1345
250.1 IPS
 
Image Segmentation (SP)
100% 611
9.91 IPS
 
Image Segmentation (HP)
100% 613
9.93 IPS
 
Image Segmentation (Q)
98% 1576
25.6 IPS
 
Pose Estimation (SP)
100% 2451
2.86 IPS
 
Pose Estimation (HP)
100% 2434
2.84 IPS
 
Pose Estimation (Q)
84% 3210
3.81 IPS
 
Object Detection (SP)
100% 479
38.0 IPS
 
Object Detection (HP)
100% 477
37.8 IPS
 
Object Detection (Q)
83% 1330
107.5 IPS
 
Face Detection (SP)
100% 1580
18.8 IPS
 
Face Detection (HP)
100% 1588
18.9 IPS
 
Face Detection (Q)
95% 3284
39.2 IPS
 
Depth Estimation (SP)
100% 2199
16.9 IPS
 
Depth Estimation (HP)
99% 2204
17.0 IPS
 
Depth Estimation (Q)
64% 3117
28.5 IPS
 
Style Transfer (SP)
89% 4122
5.35 IPS
 
Style Transfer (HP)
89% 4211
5.47 IPS
 
Style Transfer (Q)
98% 7437
9.59 IPS
 
Image Super-Resolution (SP)
100% 1293
47.7 IPS
 
Image Super-Resolution (HP)
100% 1295
47.8 IPS
 
Image Super-Resolution (Q)
97% 1782
66.0 IPS
 
Text Classification (SP)
100% 569
759.0 IPS
 
Text Classification (HP)
100% 565
754.4 IPS
 
Text Classification (Q)
88% 890
1.20 KIPS
 
Machine Translation (SP)
100% 931
16.0 IPS
 
Machine Translation (HP)
100% 927
16.0 IPS
 
Machine Translation (Q)
50% 530
18.8 IPS
 

Compare

Set Baseline Geekbench 6 now available