Allwinner A523 PRO vs Allwinner A523 PRO
| Allwinner A523 PRO | Allwinner A523 PRO | Difference | |
|---|---|---|---|
| Single Precision | 226 | 226 | 100.0% |
| Allwinner A523 PRO | |||
| Allwinner A523 PRO | |||
| Half Precision | 230 | 230 | 100.0% |
| Allwinner A523 PRO | |||
| Allwinner A523 PRO | |||
| Quantized | 396 | 396 | 100.0% |
| Allwinner A523 PRO | |||
| Allwinner A523 PRO | |||
| Geekbench AI 1.2.0 | Geekbench AI 1.2.0 | ||
System Information
| Allwinner A523 PRO | Allwinner A523 PRO | |
|---|---|---|
| AI Framework | TensorFlow Lite | TensorFlow Lite |
| AI Backend | CPU | CPU |
| AI Device | ARM ARMv8 | ARM ARMv8 |
| Operating System | Android 13 | Android 13 |
| Model | Allwinner A523 PRO | Allwinner A523 PRO |
| Processor |
ARM ARMv8 @ 1.42 GHz
1 Processor, 8 Cores |
ARM ARMv8 @ 1.42 GHz
1 Processor, 8 Cores |
| Processor ID | ARM implementer 65 architecture 8 variant 2 part 3333 revision 0 | ARM implementer 65 architecture 8 variant 2 part 3333 revision 0 |
| Motherboard | exdroid | exdroid |
| Memory | 3.67 GB | 3.67 GB |
AI Performance
| Allwinner A523 PRO | Allwinner A523 PRO | Difference | |
|---|---|---|---|
| Image Classification (SP) | 122 | 122 | 100.0% |
| Allwinner A523 PRO | |||
| Allwinner A523 PRO | |||
| Image Classification (HP) | 128 | 128 | 100.0% |
| Allwinner A523 PRO | |||
| Allwinner A523 PRO | |||
| Image Classification (Q) | 270 | 270 | 100.0% |
| Allwinner A523 PRO | |||
| Allwinner A523 PRO | |||
| Image Segmentation (SP) | 166 | 166 | 100.0% |
| Allwinner A523 PRO | |||
| Allwinner A523 PRO | |||
| Image Segmentation (HP) | 167 | 167 | 100.0% |
| Allwinner A523 PRO | |||
| Allwinner A523 PRO | |||
| Image Segmentation (Q) | 435 | 435 | 100.0% |
| Allwinner A523 PRO | |||
| Allwinner A523 PRO | |||
| Pose Estimation (SP) | 590 | 590 | 100.0% |
| Allwinner A523 PRO | |||
| Allwinner A523 PRO | |||
| Pose Estimation (HP) | 569 | 569 | 100.0% |
| Allwinner A523 PRO | |||
| Allwinner A523 PRO | |||
| Pose Estimation (Q) | 943 | 943 | 100.0% |
| Allwinner A523 PRO | |||
| Allwinner A523 PRO | |||
| Object Detection (SP) | 114 | 114 | 100.0% |
| Allwinner A523 PRO | |||
| Allwinner A523 PRO | |||
| Object Detection (HP) | 115 | 115 | 100.0% |
| Allwinner A523 PRO | |||
| Allwinner A523 PRO | |||
| Object Detection (Q) | 279 | 279 | 100.0% |
| Allwinner A523 PRO | |||
| Allwinner A523 PRO | |||
| Face Detection (SP) | 318 | 318 | 100.0% |
| Allwinner A523 PRO | |||
| Allwinner A523 PRO | |||
| Face Detection (HP) | 320 | 320 | 100.0% |
| Allwinner A523 PRO | |||
| Allwinner A523 PRO | |||
| Face Detection (Q) | 692 | 692 | 100.0% |
| Allwinner A523 PRO | |||
| Allwinner A523 PRO | |||
| Depth Estimation (SP) | 316 | 316 | 100.0% |
| Allwinner A523 PRO | |||
| Allwinner A523 PRO | |||
| Depth Estimation (HP) | 324 | 324 | 100.0% |
| Allwinner A523 PRO | |||
| Allwinner A523 PRO | |||
| Depth Estimation (Q) | 760 | 760 | 100.0% |
| Allwinner A523 PRO | |||
| Allwinner A523 PRO | |||
| Style Transfer (SP) | 992 | 992 | 100.0% |
| Allwinner A523 PRO | |||
| Allwinner A523 PRO | |||
| Style Transfer (HP) | 970 | 970 | 100.0% |
| Allwinner A523 PRO | |||
| Allwinner A523 PRO | |||
| Style Transfer (Q) | 1814 | 1814 | 100.0% |
| Allwinner A523 PRO | |||
| Allwinner A523 PRO | |||
| Image Super-Resolution (SP) | 185 | 185 | 100.0% |
| Allwinner A523 PRO | |||
| Allwinner A523 PRO | |||
| Image Super-Resolution (HP) | 191 | 191 | 100.0% |
| Allwinner A523 PRO | |||
| Allwinner A523 PRO | |||
| Image Super-Resolution (Q) | 462 | 462 | 100.0% |
| Allwinner A523 PRO | |||
| Allwinner A523 PRO | |||
| Text Classification (SP) | 87 | 87 | 100.0% |
| Allwinner A523 PRO | |||
| Allwinner A523 PRO | |||
| Text Classification (HP) | 95 | 95 | 100.0% |
| Allwinner A523 PRO | |||
| Allwinner A523 PRO | |||
| Text Classification (Q) | 107 | 107 | 100.0% |
| Allwinner A523 PRO | |||
| Allwinner A523 PRO | |||
| Machine Translation (SP) | 164 | 164 | 100.0% |
| Allwinner A523 PRO | |||
| Allwinner A523 PRO | |||
| Machine Translation (HP) | 165 | 165 | 100.0% |
| Allwinner A523 PRO | |||
| Allwinner A523 PRO | |||
| Machine Translation (Q) | 66 | 66 | 100.0% |
| Allwinner A523 PRO | |||
| Allwinner A523 PRO | |||
