Geekbench AI Score

964
Single Precision Score
1126
Half Precision Score
785
Quantized Score
Geekbench AI 1.2.0 for Windows AVX2

Result Information

User bormolino
Upload Date January 11 2025 07:47 PM
Views 68

System Information

AI Information
Framework ONNX
Backend DirectML
Device Intel(R) UHD Graphics 620
System Information
Operating System Microsoft Windows 10 Pro (64-bit)
Model LENOVO 20RH001FGE
Motherboard LENOVO 20RH001FGE
Power PlanAusbalanciert
CPU Information
Name Intel Core i7-8565U
Topology 1 Processor, 4 Cores, 8 Threads
Identifier GenuineIntel Family 6 Model 142 Stepping 12
Base Frequency 1.99 GHz
Cluster 14 Cores
Memory Information
Size 16.00 GB

AI Performance

Workload Accuracy Score
Image Classification (SP)
100% 411
76.5 IPS
 
Image Classification (HP)
100% 472
87.7 IPS
 
Image Classification (Q)
100% 424
78.9 IPS
 
Image Segmentation (SP)
100% 596
9.67 IPS
 
Image Segmentation (HP)
100% 702
11.4 IPS
 
Image Segmentation (Q)
98% 503
8.18 IPS
 
Pose Estimation (SP)
100% 2349
2.74 IPS
 
Pose Estimation (HP)
100% 3741
4.36 IPS
 
Pose Estimation (Q)
97% 2133
2.50 IPS
 
Object Detection (SP)
100% 496
39.4 IPS
 
Object Detection (HP)
98% 497
39.5 IPS
 
Object Detection (Q)
86% 451
36.2 IPS
 
Face Detection (SP)
100% 878
10.4 IPS
 
Face Detection (HP)
100% 1106
13.1 IPS
 
Face Detection (Q)
97% 763
9.10 IPS
 
Depth Estimation (SP)
100% 1744
13.4 IPS
 
Depth Estimation (HP)
98% 2192
16.9 IPS
 
Depth Estimation (Q)
77% 1491
11.9 IPS
 
Style Transfer (SP)
100% 6034
7.76 IPS
 
Style Transfer (HP)
100% 9039
11.6 IPS
 
Style Transfer (Q)
98% 5044
6.50 IPS
 
Image Super-Resolution (SP)
100% 1159
42.8 IPS
 
Image Super-Resolution (HP)
100% 1623
59.9 IPS
 
Image Super-Resolution (Q)
99% 912
33.8 IPS
 
Text Classification (SP)
100% 485
647.6 IPS
 
Text Classification (HP)
100% 309
412.0 IPS
 
Text Classification (Q)
97% 302
405.0 IPS
 
Machine Translation (SP)
100% 469
8.07 IPS
 
Machine Translation (HP)
100% 485
8.35 IPS
 
Machine Translation (Q)
70% 277
5.20 IPS
 

Compare

Set Baseline Geekbench 6 now available