## https://en.wikipedia.org/wiki/ARM_Cortex-M

Infineon [32-bit XMC™ Industrial Microcontroller based on Arm® Cortex®-M](https://www.infineon.com/cms/en/product/microcontroller/32-bit-industrial-microcontroller-based-on-arm-cortex-m/?gclid=EAIaIQobChMI-7r7l8WK8AIVwtmyCh1OpANuEAAYASAAEgKh0PD_BwE&gclsrc=aw.ds)

## [Machine Learning on Arm Cortex-M Microcontrollers](https://knowen-production.s3.amazonaws.com/uploads/attachment/file/5164/Arm%2BML%2Bon%2BCortex-M%2BMicrocontrollers%2B_1_.pdf)

> we present CMSIS-NN, an open-source
> library of optimized software kernels that maximize
> the NN performance on Cortex-M cores with minimal
> memory footprint overhead

![x](https://knowen-production.s3.amazonaws.com/uploads/attachment/file/5163/cortex.PNG 'x')

To quantify the benefits of CMSIS-NN kernels over existing solutions, we also implemented a baseline version using a **1D convolution function (arm_conv from CMSIS-DSP), Caffe-like pooling and ReLU.**
For the CNN application, Table I summarizes the comparison results of the baseline Runtime (ms) functions and the CMSIS-NN kernels. The CMSIS-NN kernels achieve 2.6X to 5.4X improvement in runtime/throughput over the baseline functions. **The energy efficiency improvement is also in line with the throughput improvement.** 


## software

2019 [Efficient Neural Network Deployment for Microcontroller](https://arxiv.org/abs/2007.01348) Hasan Unlu