Skip to content

CHANDU095/TX2-dnn-power-measurements

 
 

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

24 Commits
 
 
 
 
 
 
 
 

Repository files navigation

TX2-dnn-power-measurements

This project measures power of the TX2 board processing DNN models(take two stream cnn as an example). NVIDIA is providing a command line tool 'nvpmodel' which takes out a lot of the guess work in configuring the CPU and GPU settings to maximize performance and energy usage under different scenarios

Nvpmodel

Nvpmodel introduces five different “modes” on the Jetson TX2. To summarize:

alt text

To call nvpmodel:

$ sudo nvpmodel -m [mode]

where mode is the number of the mode that you want to use. For example:

$ sudo nvpmodel -m 1

You can query which mode is currently being used:

$ sudo nvpmodel -q --verbose

The file /etc/nvpmodel.conf holds the different models. Developers can add their own models to add different modes suitable to their application.

We can develop some code to read out these sensors to perform power measurements on the TX2 board. It is very convenient in general, and even allows to automatize the process for different benchmark programs, etc.

QuickStart

Remember that the Jetson TX2 consists of a GPU along with a CPU cluster.Go to the tx2-power-measurements directory,the tx2_predict.py script with two stream cnn model executes 50 inferences. Then it reports the power measurements.Use the commands below to excute CPU and GPU versions:

  • GPU version

python tx2_predict.py

  • CPU version

CUDA_VISIBLE_DEVICE="" python tx2_predict.py

Make sure all dependencies is installed before running.The reports will show as something like below:

GPU version

alt text

CPU version

alt text

Reference

[1] JetsonHacks url:https://www.jetsonhacks.com/

[2] "NVPModel – NVIDIA Jetson TX2 Development Kit" url:https://www.jetsonhacks.com/2017/03/25/nvpmodel-nvidia-jetson-tx2-development-kit/

About

For the Unet CDLLM-Project

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published

Languages

  • Python 100.0%