Implementation of 3D degridding algorithm on the NVIDIA GPUs using CUDA

TL;DR

This paper presents a GPU-accelerated implementation of a 3D degridding algorithm for aperture synthesis imaging, achieving high performance and addressing irregular sampling and load balancing challenges.

Contribution

The work introduces a CUDA-based, distributed GPU implementation of 3D degridding that handles irregular data sampling and load balancing for improved imaging accuracy.

Findings

Achieves up to 1.2 billion visibilities per second performance.

Effectively manages irregular and non-planar sampling issues.

Provides a scalable solution for distributed aperture synthesis imaging.

Abstract

Practical aperture synthesis imaging algorithms work by iterating between estimating the sky brightness distribution and a comparison of a prediction based on this estimate with the measured data ("visibilities"). Accuracy in the latter step is crucial but is made difficult by irregular and non-planar sampling of data by the telescope. In this work we present a GPU implementation of 3d de-gridding which accurately deals with these two difficulties and is designed for distributed operation. We address the load balancing issues caused by large variation in visibilities that need to be computed. Using CUDA and NVidia GPUs we measure performance up to 1.2 billion visibilities per second.