Jetson Nano - stereoboy/Study GitHub Wiki

References

• User Guide
  • https://developer.nvidia.com/embedded/downloads#?search=Jetson%20Nano%20Developer%20Kit%20User%20Guide

Fan Control

sudo /usr/bin/jetson_clocks
sudo sh -c ‘echo 255 > /sys/devices/pwm-fan/target_pwm’

Monitoring

tegrastats

Benchmark (for Fan)

this benchmarks force to make /sys/devices/pwm-fan/target_pwm as 255 forced

• https://developer.nvidia.com/embedded/jetson-nano-dl-inference-benchmarks

• https://github.com/NVIDIA-AI-IOT/jetson_benchmarks

$ git clone https://github.com/NVIDIA-AI-IOT/jetson_benchmarks.git
Cloning into 'jetson_benchmarks'...
remote: Enumerating objects: 49, done.
remote: Counting objects: 100% (49/49), done.
remote: Compressing objects: 100% (34/34), done.
remote: Total 49 (delta 18), reused 41 (delta 13), pack-reused 0
Unpacking objects: 100% (49/49), done.
$ cd jetson_benchmarks/
$ mkdir models
$ sudo sh install_requirements.sh
...
$ python3 utils/download_models.py --all --csv_file_path ./benchmark_csv/tx2-nano-benchmarks.csv --save_dir ~/work/jetson_benchmarks/models/
inception_v4.prototxt                               100%[===================================================================================================================>] 118.14K   396KB/s    in 0.3s    
vgg19_N2.prototxt                                   100%[===================================================================================================================>]   5.80K  --.-KB/s    in 0s      
super_resolution_bsd500.zip                         100%[===================================================================================================================>] 990.22K  1.13MB/s    in 0.9s    
unet-segmentation.uff                               100%[===================================================================================================================>]   1.88M  2.31MB/s    in 0.8s    
pose_estimation.prototxt                            100%[===================================================================================================================>]  44.90K  --.-KB/s    in 0.1s    
yolov3-tiny-416.zip                                 100%[===================================================================================================================>] 156.92M  4.38MB/s    in 42s     
ResNet50_224x224.prototxt                           100%[===================================================================================================================>]  34.20K  --.-KB/s    in 0.1s    
ssd-mobilenet-v1.zip                                100%[===================================================================================================================>] 171.01M  4.48MB/s    in 40s

wom@wom-nano:~/work/jetson_benchmarks$ sudo python3 benchmark.py --all --csv_file_path ./benchmark_csv/tx2-nano-benchmarks.csv \
                            --model_dir ~/work/jetson_benchmarks/models \
                            --jetson_devkit nano \
                            --gpu_freq 921600000 --power_mode 0 --precision fp16

$ sudo python3 benchmark.py --model_name ssd-mobilenet-v1 --csv_file_path ./benchmark_csv/tx2-nano-benchmarks.csv \
                            --model_dir ~/work/jetson_benchmarks/models \
                            --jetson_devkit nano \
                            --gpu_freq 921600000 --power_mode 0 --precision fp16