Test results
Performance testing
For performance testing, we took a video of a corridor with a flow of people moving at a speed of 5-10 people/sec (FPS=25). The volume of the database for identification is 249 persons.
Metrics
Metric name | Description |
Maximum | Maximum value of parameters (CPU load, GPU load, amount of RAM, amount of VRAM) for the entire test time |
95th percentile |
|
Median |
|
FPS | Number of frames processed per second |
Number of tracks | Number of tracks created during the test |
Number of tracks with biometric templates | Number of tracks that contain the generated biometric face template |
Number of matches with the face database | Number of matches of faces from videos with faces from the database |
Test 1
Test machine specification
CPU | RAM | Graphics Card |
AMD Ryzen 9 5950X @ 3.4 GHz: 16 cores (32 threads) | 118GB DDR4 | NVIDIA GeForce GTX 1080 Ti |
Test Results
Number of video streams | 1 | 5 | 8 | 10 | |
Maximum | Number of CPU cores | 2 | 8 | 14 | 16 |
GPU load | 85% | 95% | 96% | 97% | |
Amount of RAM | 3.95 GB | 4.94 GB | 6.42 GB | 8.03 GB | |
Amount of VRAM | 2.08 GB | 2.41 GB | 2.41 GB | 2.92 GB | |
95th percentile | Number of CPU cores | 1 | 8 | 14 | 16 |
GPU load | 21% | 73% | 78% | 76% | |
Amount of RAM | 2.34 GB | 4.69 GB | 6.42 GB | 7.66 GB | |
Amount of VRAM | 2.08 GB | 2.41 GB | 2.34 GB | 2.92 GB | |
Median | Number of CPU cores | 1 | 7 | 13 | 16 |
GPU load | 11% | 50% | 49% | 39% | |
Amount of RAM | 2.34 GB | 4.69 GB | 6.30 GB | 7.54 GB | |
Amount of VRAM | 2.08 GB | 2.41 GB | 2.34 GB | 2.92 GB | |
Number of tracks | 2553 | 2499 | 1954 | 1472 | |
FPS | 24.8 | 24.6 | 24.6 | 25.2 | |
Number of tracks with biometric templates | 733 | 733 | 639 | 564 | |
Number of matches with the face database | 239 | 238 | 233 | 226 |
Conclusion
~ 1.5 cores per video stream when using GPU in the most high-load case "Safe City".