Heterogeneous CPU+iGPU processing for efficient epistasis detection

Abstract

Epistasis detection represents a fundamental problem in bio-medicine to understand the reasons for occurrence of complex phenotypic traits (diseases) across a population of individuals. Exhaustively examining all possible interactions of multiple Single-Nucleotide Polymorphisms provides the most reliable way to identify accurate solutions, but it is both computationally and memory intensive task. To tackle this challenge, this work proposes a modular and self-adaptive framework for high-performance and energy-efficient epistasis analysis on modern tightly-coupled heterogeneous platforms composed of multi-core CPUs and integrated GPUs. To fully exploit the capabilities of these systems, the proposed framework incorporates both task- and data-parallel approaches specifically tailored to enhance single and multi-objective epistasis detection on each device architecture, along with allowing efficient collaborative execution across all devices. The experimental results show the ability of the proposed framework to handle the heterogeneity of an Intel CPU+iGPU system, achieving performance and energy-efficiency gains of up to 5× and 6× in different parallel execution scenarios.