15. Performance Report
Performance statistics logging is turned on by setting a launch parameter, exportPerfStats
to 1.
15.1. TDA4VM
Note: “A72 Load (%)” are for dual A72 cores in a scale of 100%. For example, 100% A72 loading means that two A72 cores are fully loaded.
15.1.1. ZED Camera, 15 FPS
Source: live ZED camera, 1280x720 on each of left and right image, at 15 FPS. “zed_capture” ROS node and a demo ROS node are running in the ROS 2 Docker container on the target SK board.
Demo |
FPS |
Total time (ms) |
Preproc time (ms) |
Inference time (ms) |
A72 Load (%) |
DDR Read BW (MB/s) |
DDR Write BW (MB/s) |
DDR Total BW (MB/s) |
C71_1 Load (%) |
C66_1 Load (%) |
C66_2 Load (%) |
MCU2_0 Load (%) |
MCU2_1 Load (%) |
MSC_0 (%) |
MSC_1 (%) |
VISS (%) |
NF (%) |
LDC (%) |
SDE (%) |
DOF (%) |
---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
ti_sde |
14.04 |
71.2117 |
NA |
NA |
35.82 |
943 |
884 |
1827 |
1.0 |
1.0 |
0.0 |
2.0 |
1.0 |
0 |
0 |
0 |
0 |
3.16 |
12.7 |
0 |
ti_sde (w/ pcl) |
15.03 |
66.5456 |
NA |
NA |
56.53 |
1565 |
1503 |
3068 |
0.0 |
17.0 |
1.0 |
4.0 |
3.0 |
6.65 |
0 |
0 |
0 |
4.54 |
21.17 |
0 |
ti_vision_cnn (semseg) |
15.14 |
66.0599 |
5.0000 |
10.0000 |
24.37 |
1026 |
890 |
1916 |
14.0 |
1.0 |
0.0 |
2.0 |
1.0 |
3.21 |
0 |
0 |
0 |
2.14 |
0 |
0 |
ti_vision_cnn (objdet) |
15.11 |
66.1995 |
4.0000 |
5.0146 |
22.55 |
801 |
664 |
1465 |
8.0 |
0.0 |
0.0 |
2.0 |
1.0 |
3.17 |
0 |
0 |
0 |
2.11 |
0 |
0 |
ti_estop |
15.05 |
66.4585 |
5.0195 |
10.0049 |
42.85 |
1465 |
1213 |
2678 |
14.0 |
0.0 |
1.0 |
3.0 |
1.0 |
3.28 |
0 |
0 |
0 |
4.52 |
16.37 |
0 |
15.1.2. C920 USB Camera, 30 FPS
Source: live C920 USB camera, 1280x720 in MJPG mode, at 30 FPS. “gscam2” ROS node and a demo ROS node are running in the ROS 2 Docker container on the target SK board.
Demo |
FPS |
Total time (ms) |
Preproc time (ms) |
Inference time (ms) |
A72 Load (%) |
DDR Read BW (MB/s) |
DDR Write BW (MB/s) |
DDR Total BW (MB/s) |
C71_1 Load (%) |
C66_1 Load (%) |
C66_2 Load (%) |
MCU2_0 Load (%) |
MCU2_1 Load (%) |
MSC_0 (%) |
MSC_1 (%) |
VISS (%) |
NF (%) |
LDC (%) |
SDE (%) |
DOF (%) |
---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
ti_vision_cnn (semseg) |
30.17 |
33.1460 |
5.0668 |
10.0111 |
56.4 |
1569 |
1111 |
2680 |
27.0 |
0.0 |
0.0 |
4.0 |
1.0 |
6.4 |
0 |
0 |
0 |
4.4 |
0 |
0 |
ti_vision_cnn (objdet) |
30.30 |
33.0048 |
4.0538 |
5.0490 |
51.3 |
1189 |
709 |
1898 |
15.0 |
1.0 |
0.0 |
4.0 |
1.0 |
6.43 |
0 |
0 |
0 |
4.21 |
0 |
0 |
ti_vision_cnn (6dpose) |
28.84 |
34.6721 |
3.2079 |
28.5254 |
49.87 |
1868 |
809 |
2677 |
78.0 |
0.0 |
0.0 |
2.0 |
1.0 |
6.16 |
0 |
0 |
0 |
0 |
0 |
0 |
15.2. AM68A
Note: “A72 Load (%)” are for dual A72 cores in a scale of 100%. For example, 100% A72 loading means that two A72 cores are fully loaded.
15.2.1. ZED Camera, 15 FPS
Source: live ZED camera, 1280x720 on each of left and right image, at 15 FPS. “zed_capture” ROS node and a demo ROS node are running in the ROS 2 Docker container on the target SK board.
Demo |
FPS |
Total time (ms) |
Preproc time (ms) |
Inference time (ms) |
A72 Load (%) |
DDR Read BW (MB/s) |
DDR Write BW (MB/s) |
DDR Total BW (MB/s) |
C71_1 Load (%) |
C71_2 Load (%) |
MCU2_0 Load (%) |
MCU2_1 Load (%) |
MSC_0 (%) |
MSC_1 (%) |
VISS (%) |
NF (%) |
LDC (%) |
SDE (%) |
DOF (%) |
---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
ti_sde |
14.99 |
66.6916 |
NA |
NA |
32.16 |
1085 |
978 |
2063 |
0.0 |
0.0 |
1.0 |
1.0 |
0 |
0 |
0 |
0 |
4.27 |
16.39 |
0 |
ti_sde (w/ pcl) |
15.00 |
66.6619 |
NA |
NA |
49.11 |
1652 |
1516 |
3168 |
1.0 |
12.0 |
3.0 |
2.0 |
6.21 |
0 |
0 |
0 |
4.26 |
20.64 |
0 |
ti_vision_cnn (semseg) |
15.13 |
66.1060 |
5.0211 |
13.0000 |
23.92 |
1766 |
1398 |
3164 |
19.0 |
0.0 |
1.0 |
0.0 |
3.14 |
0 |
0 |
0 |
2.9 |
0 |
0 |
ti_vision_cnn (objdet) |
15.08 |
66.2920 |
4.0170 |
7.4393 |
23.13 |
1037 |
785 |
1822 |
12.0 |
1.0 |
1.0 |
1.0 |
3.10 |
0 |
0 |
0 |
2.7 |
0 |
0 |
ti_estop |
15.09 |
66.2777 |
5.0098 |
13.0000 |
40.67 |
2233 |
1728 |
3961 |
19.0 |
1.0 |
2.0 |
1.0 |
3.12 |
0 |
0 |
0 |
4.29 |
16.13 |
0 |
15.2.2. C920 USB Camera, 30 FPS
Source: live C920 USB camera, 1280x720 in MJPG mode, at 30 FPS. “gscam2” ROS node and a demo ROS node are running in the ROS 2 Docker container on the target SK board.
Demo |
FPS |
Total time (ms) |
Preproc time (ms) |
Inference time (ms) |
A72 Load (%) |
DDR Read BW (MB/s) |
DDR Write BW (MB/s) |
DDR Total BW (MB/s) |
C71_1 Load (%) |
C71_2 Load (%) |
MCU2_0 Load (%) |
MCU2_1 Load (%) |
MSC_0 (%) |
MSC_1 (%) |
VISS (%) |
NF (%) |
LDC (%) |
SDE (%) |
DOF (%) |
---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
ti_vision_cnn (semseg) |
29.54 |
33.8578 |
5.2307 |
13.0228 |
56.72 |
3060 |
2138 |
5198 |
37.0 |
1.0 |
2.0 |
1.0 |
6.11 |
0 |
0 |
0 |
4.2 |
0 |
0 |
ti_vision_cnn (objdet) |
30.18 |
33.1354 |
4.0749 |
7.0012 |
54.40 |
1595 |
920 |
2515 |
19.0 |
0.0 |
3.0 |
0.0 |
5.97 |
0 |
0 |
0 |
3.99 |
0 |
0 |
ti_vision_cnn (6dpose) |
28.47 |
35.1237 |
3.0829 |
28.6362 |
49.48 |
4733 |
1698 |
6431 |
74.0 |
0.0 |
1.0 |
1.0 |
5.61 |
0 |
0 |
0 |
0 |
0 |
0 |
15.3. AM69A
Note: “A72 Load (%)” are for 8x A72 cores in a scale of 100%. For example, 100% A72 loading means that 8x A72 cores are fully loaded.
15.3.1. ZED Camera, 15 FPS
Source: live ZED camera, 1280x720 on each of left and right image, at 15 FPS. “zed_capture” ROS node and a demo ROS node are running in the ROS 2 Docker container on the target SK board.
Demo |
FPS |
Total time (ms) |
Preproc time (ms) |
Inference time (ms) |
A72 Load (%) |
DDR Read BW (MB/s) |
DDR Write BW (MB/s) |
DDR Total BW (MB/s) |
C71_1 Load (%) |
C71_2 Load (%) |
C71_3 Load (%) |
C71_4 Load (%) |
MCU2_0 Load (%) |
MCU2_1 Load (%) |
MCU3_0 Load (%) |
MCU3_1 Load (%) |
MCU4_0 Load (%) |
MCU4_1 Load (%) |
MSC_0 (%) |
MSC_1 (%) |
VISS (%) |
NF (%) |
LDC (%) |
SDE (%) |
DOF (%) |
---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
ti_sde |
15.12 |
66.1392 |
NA |
NA |
7.22 |
1397 |
930 |
2327 |
0.0 |
0.0 |
1.0 |
1.0 |
2.0 |
1.0 |
1.0 |
1.0 |
1.0 |
1.0 |
0 |
0 |
0 |
0 |
4.37 |
16.17 |
0 |
ti_sde (w/ pcl) |
15.14 |
66.0615 |
NA |
NA |
12.0 |
1915 |
1450 |
3365 |
0.0 |
13.0 |
1.0 |
1.0 |
3.0 |
3.0 |
1.0 |
0.0 |
1.0 |
1.0 |
6.36 |
0 |
0 |
0 |
4.43 |
20.96 |
0 |
ti_vision_cnn (semseg) |
15.15 |
66.0234 |
5.0000 |
9.0029 |
5.13 |
1913 |
1367 |
3280 |
13.0 |
0.0 |
0.0 |
1.0 |
2.0 |
1.0 |
1.0 |
0.0 |
1.0 |
1.0 |
3.13 |
0 |
0 |
0 |
2.13 |
0 |
0 |
ti_vision_cnn (objdet) |
15.15 |
66.0049 |
4.0000 |
7.0000 |
4.50 |
1268 |
728 |
1996 |
10.0 |
0.0 |
0.0 |
0.0 |
1.0 |
1.0 |
1.0 |
1.0 |
1.0 |
1.0 |
3.10 |
0 |
0 |
0 |
2.10 |
0 |
0 |
ti_estop |
15.14 |
66.0312 |
5.0024 |
9.0072 |
9.19 |
2298 |
1687 |
3985 |
13.0 |
0.0 |
0.0 |
0.0 |
3.0 |
1.0 |
1.0 |
1.0 |
1.0 |
1.0 |
3.11 |
0 |
0 |
0 |
4.35 |
16.45 |
0 |
15.3.2. C920 USB Camera, 30 FPS
Source: live C920 USB camera, 1280x720 in MJPG mode, at 30 FPS. “gscam2” ROS node and a demo ROS node are running in the ROS 2 Docker container on the target SK board.
Demo |
FPS |
Total time (ms) |
Preproc time (ms) |
Inference time (ms) |
A72 Load (%) |
DDR Read BW (MB/s) |
DDR Write BW (MB/s) |
DDR Total BW (MB/s) |
C71_1 Load (%) |
C71_2 Load (%) |
C71_3 Load (%) |
C71_4 Load (%) |
MCU2_0 Load (%) |
MCU2_1 Load (%) |
MCU3_0 Load (%) |
MCU3_1 Load (%) |
MCU4_0 Load (%) |
MCU4_1 Load (%) |
MSC_0 (%) |
MSC_1 (%) |
VISS (%) |
NF (%) |
LDC (%) |
SDE (%) |
DOF (%) |
---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
ti_vision_cnn (semseg) |
30.50 |
32.7913 |
4.9929 |
9.0012 |
13.55 |
2990 |
2137 |
5127 |
26.0 |
0.0 |
0.0 |
0.0 |
2.0 |
1.0 |
1.0 |
1.0 |
1.0 |
1.0 |
6.17 |
0 |
0 |
0 |
4.36 |
0 |
0 |
ti_vision_cnn (objdet) |
30.47 |
32.8151 |
3.9918 |
5.0024 |
14.2 |
1717 |
878 |
2595 |
16.0 |
0.0 |
1.0 |
0.0 |
3.0 |
1.0 |
1.0 |
1.0 |
1.0 |
1.0 |
6.21 |
0 |
0 |
0 |
4.38 |
0 |
0 |
ti_vision_cnn (6dpose) |
29.36 |
34.0610 |
3.0026 |
27.9881 |
11.8 |
4850 |
1772 |
6622 |
77.0 |
0.0 |
1.0 |
1.0 |
2.0 |
1.0 |
1.0 |
1.0 |
1.0 |
1.0 |
5.92 |
0 |
0 |
0 |
0 |
0 |
0 |