Geekbench AI Score

1748
Single Precision Score
716
Half Precision Score
1704
Quantized Score
Geekbench AI 1.7.0 for Windows x86 (64-bit)

Result Information

User blade9722
Upload Date May 02 2026 05:47 AM
Views 11

System Information

AI Information
Framework ONNX
Backend CPU
Device Intel Xeon E5-2697 v2
System Information
Operating System Microsoft Windows 11 Pro (64-bit)
Model JGINYUE
Motherboard JGINYUE X79M-G ARGB
Power PlanBilanciato
CPU Information
Name Intel Xeon E5-2697 v2
Topology 1 Processor, 12 Cores, 24 Threads
Identifier GenuineIntel Family 6 Model 62 Stepping 4
Base Frequency 2.70 GHz
Cluster 112 Cores
Memory Information
Size 16.00 GB

AI Performance

Workload Accuracy Score
Image Classification (SP)
100% 1533
285.0 IPS
 
Image Classification (HP)
100% 255
47.4 IPS
 
Image Classification (Q)
96% 1291
241.1 IPS
 
Image Segmentation (SP)
100% 696
11.3 IPS
 
Image Segmentation (HP)
100% 395
6.41 IPS
 
Image Segmentation (Q)
97% 1884
30.6 IPS
 
Pose Estimation (SP)
100% 2978
3.48 IPS
 
Pose Estimation (HP)
100% 2249
2.62 IPS
 
Pose Estimation (Q)
86% 2941
3.48 IPS
 
Object Detection (SP)
100% 1317
104.5 IPS
 
Object Detection (HP)
100% 322
25.6 IPS
 
Object Detection (Q)
58% 1011
110.4 IPS
 
Face Detection (SP)
100% 2747
32.6 IPS
 
Face Detection (HP)
100% 716
8.51 IPS
 
Face Detection (Q)
96% 2423
28.9 IPS
 
Depth Estimation (SP)
100% 3824
29.5 IPS
 
Depth Estimation (HP)
99% 974
7.51 IPS
 
Depth Estimation (Q)
78% 2784
22.2 IPS
 
Style Transfer (SP)
100% 7218
9.28 IPS
 
Style Transfer (HP)
100% 5987
7.70 IPS
 
Style Transfer (Q)
87% 5057
6.58 IPS
 
Image Super-Resolution (SP)
100% 1312
48.5 IPS
 
Image Super-Resolution (HP)
100% 1052
38.9 IPS
 
Image Super-Resolution (Q)
97% 1331
49.3 IPS
 
Text Classification (SP)
100% 662
883.1 IPS
 
Text Classification (HP)
100% 308
411.5 IPS
 
Text Classification (Q)
97% 543
727.6 IPS
 
Machine Translation (SP)
100% 970
16.7 IPS
 
Machine Translation (HP)
100% 361
6.22 IPS
 
Machine Translation (Q)
68% 1161
22.4 IPS
 

Compare

Set Baseline Geekbench AI now available