Compression of Site-Specific Deep Neural Networks for Massive MIMO Precoding

TL;DR

This paper presents a framework for compressing deep neural network precoders in massive MIMO systems, significantly improving energy efficiency while maintaining performance, and provides a benchmark for future energy-efficient DL precoding methods.

Contribution

It introduces a novel compression framework using mixed-precision quantization and neural architecture search for energy-efficient DL-based precoders in mMIMO systems.

Findings

Deep neural network compression achieves up to 35 times higher energy efficiency than WMMSE.

Site-specific conditions significantly influence the energy efficiency of compressed models.

The proposed methods maintain accuracy while reducing energy consumption in massive MIMO precoding.

Abstract

The deployment of deep learning (DL) models for precoding in massive multiple-input multiple-output (mMIMO) systems is often constrained by high computational demands and energy consumption. In this paper, we investigate the compute energy efficiency of mMIMO precoders using DL-based approaches, comparing them to conventional methods such as zero forcing and weighted minimum mean square error (WMMSE). Our energy consumption model accounts for both memory access and calculation energy within DL accelerators. We propose a framework that incorporates mixed-precision quantization-aware training and neural architecture search to reduce energy usage without compromising accuracy. Using a ray-tracing dataset covering various base station sites, we analyze how site-specific conditions affect the energy efficiency of compressed models. Our results show that deep neural network compression generates precoders with up to 35 times higher energy efficiency than WMMSE at equal performance, depending on the scenario and the desired rate. These results establish a foundation and a benchmark for the development of energy-efficient DL-based mMIMO precoders.