Comparative assessment of GPGPU technologies to accelerate objective functions: A case study on parsimony

Abstract Objective functions provide measurements of solution quality that represent the core calculations required to tackle NP-hard optimization problems. Since their complexity keeps growing with the introduction of more realistic data, research efforts have turned their interest into the proposal of efficient objective function implementations that take advantage of potential parallelism. This work explores GPGPU technologies to accelerate objective functions, considering as a case study the parallelization of phylogenetic parsimony calculations from DNA data. We undertake the comparative evaluation of different GPU programming models and architectures, highlighting the benefits and drawbacks of each approach through experimentation on six real-world biological datasets. Experimental results shed light on the strong relationship between the characteristics of the input data and the effective utilization of GPU resources. Furthermore, comparisons with other parallel architectures and methods point out how current and future optimization scenarios can benefit from the use of accurate, efficient GPU approaches.