Porting and optimizing UniFrac for GPUs

TL;DR

This paper details the porting and optimization of the UniFrac microbiome comparison metric to GPUs, significantly reducing computation time for large datasets through code optimization and parallelization.

Contribution

The paper presents a GPU implementation of Striped UniFrac, achieving substantial speedups over CPU versions with minimal precision loss.

Findings

Reduced computation time from 13 hours to 12 minutes on GPU

Achieved sub-2-hour processing for 113k samples dataset on GPU

Provided a BSD-licensed GPU-accelerated UniFrac implementation

Abstract

UniFrac is a commonly used metric in microbiome research for comparing microbiome profiles to one another ("beta diversity"). The recently implemented Striped UniFrac added the capability to split the problem into many independent subproblems and exhibits near linear scaling. In this paper we describe steps undertaken in porting and optimizing Striped Unifrac to GPUs. We reduced the run time of computing UniFrac on the published Earth Microbiome Project dataset from 13 hours on an Intel Xeon E5-2680 v4 CPU to 12 minutes on an NVIDIA Tesla V100 GPU, and to about one hour on a laptop with NVIDIA GTX 1050 (with minor loss in precision). Computing UniFrac on a larger dataset containing 113k samples reduced the run time from over one month on the CPU to less than 2 hours on the V100 and 9 hours on an NVIDIA RTX 2080TI GPU (with minor loss in precision). This was achieved by using OpenACC for generating the GPU offload code and by improving the memory access patterns. A BSD-licensed implementation is available, which produces a C shared library linkable by any programming language.