# Accelerating Channel Estimation and Demodulation of Uplink OFDM symbols for Large Scale Antenna Systems using GPU

## TL;DR

This paper presents a GPU-accelerated approach for fast channel estimation and demodulation of uplink OFDM symbols in large-scale antenna systems, demonstrating near real-time processing capabilities.

## Contribution

It introduces a combined CPU-GPU processing method for uplink OFDM in large antenna systems, optimizing back-end data processing speed.

## Key findings

- GPU acceleration improves processing speed significantly.
- Performance varies with number of antennas and FFT length.
- Near real-time processing achieved on USRP platform.

## Abstract

Increase in the number of antennas in the front-end increases the volume of data to be processed at the back-end. This establishes a need for acceleration in back-end processing. To solve the issue of high volume data processing at back-end, a GPU is utilized. Acceleration for Least Squares channel estimation and demodulation of uplink OFDM symbols is provided by using a combination of CPU and GPU at the back-end. Single user uplink scenario is implemented in near real-time manner using the USRP platform present in the Large scale antenna systems in ORBIT Testbed. The number of antennas and FFT length are varied to provide different scenarios for comparison. The performance of both CPU and GPU is compared for each process.