DEBIX Model A Installation & Benchmark NPU
Performance measurement with Machine Learning load


DEBIX Model A  an industrial grade single board computer to bring you faster speed and more powerful performance. Based on quad core i.MX 8M Plus with 2.3 TOPS NPU, DEBIX Model A with complete software development and rich I/O ports is ready and capable for direct applications in industry 4.0, IoTs, smart cities and multimedia.    

In this post we will see how to install your Android operating system and measure the performance of your CPU under a Machine Learning load by comparing the results with Raspberry Pi 4.

i.MX 8M Plus

The family CPU i.MX 8M Plus focuses on machine learning and vision, advanced multimedia, and industrial automation with high reliability. It is built to meet the needs of Smart Home, Building, City and Industry 4.0 applications.

Powerful 4x or 2x Arm® Cortex®-A53 processor up 1.8 Ghz with a Neural Processing Unit (NPU) operating at up to 2.3 TOP and Dual Image Signal Processor (ISPs): Resolution up to 12MP and input rate up to 375MPixels/s. 


Installation

  1. Download Android system image 

  2. Burn image with balena Etcher o Win32Image

  3. Start DEBIX

  4. Install Geekbench ML app for Android

Benchmark

NNAPI

The API de Neural Networks (NNAPI) The Android Neural Networks API (NNAPI) is an Android C API designed for running computationally intensive operations for machine learning on Android devices. NNAPI is designed to provide a base layer of functionality for higher-level machine learning frameworks, such as TensorFlow Lite y Caffe2), que crean y preparan redes neuronales. La API está disponible en todos los dispositivos Android 8.1 (API nivel 27) o versiones posteriores.
It provides acceleration for TensorFlow Lite models on Android devices with supported hardware accelerators including:
  • Graphics Processing Unit (GPU)
  • Digital Signal Processor (DSP)
  • Neural Processing Unit (NPU)

Geekbench ML

Geekbench ML uses real-world machine learning tasks to evaluate mobile inference performance. Geekbench ML measures your CPU, GPU, and NPU to determine whether your device is ready for today's and tomorrow's cutting-edge machine learning applications.

Raspberry Pi 4 vs nxp EMB-iMX8MP-02 (DEBIX Model A)
NNAPI

Completed results: https://browser.geekbench.com/ml/v0/inference/compare/207274?baseline=207272

Raspberry Pi 4 vs nxp EMB-iMX8MP-02 (DEBIX Model A)
CPU

Completed results:https://browser.geekbench.com/ml/v0/inference/compare/207264?baseline=207272

 
IoT + 3D Printer
Real-time monitoring a 3D printer