sPEGG: high throughput eco-evolutionary simulations on commodity graphics processors

TL;DR

sPEGG is an open-source, high-throughput simulator for eco-evolutionary models that leverages commodity GPUs to achieve over 200-fold acceleration, making complex population genetics analyses more accessible.

Contribution

This work introduces sPEGG, a novel GPU-accelerated simulator that significantly speeds up eco-evolutionary simulations on standard hardware.

Findings

sPEGG achieves over 200x speedup compared to CPU-based simulations.

The simulator enables complex models to run efficiently on a single commodity GPU.

Performance is comparable to small-to-medium computer clusters.

Abstract

Integrating population genetics into community ecology theory is a major goal in ecology and evolution, but analyzing the resulting models is computationally daunting. Here we describe sPEGG ($\underline{\textrm{s}}\textrm{imulating}$ $\underline{\textrm{P}}\textrm{henotypic}$ $\underline{\textrm{E}}\textrm{volution}$ on $\underline{\textrm{G}}\textrm{eneral Purpose}$ $\underline{\textrm{G}}\textrm{raphics Processing Units}$ (GPGPUs)), an open-source, multi-species forward-time population genetics simulator. Using a single commodity GPGPU instead of a single central processor, we find sPEGG can accelerate eco-evolutionary simulations by a factor of over 200, comparable to performance on a small-to-medium sized computer cluster.