Discrete event modeling and massively parallel execution of epidemic outbreak phenomena

Abstract

In complex phenomena such as epidemiological outbreaks, the intensity of inherent feedback effects and the significant role of transients in the dynamics make simulation the only effective method for proactive, reactive or post facto analysis. The spatial scale, runtime speed, and behavioral detail needed in detailed simulations of epidemic outbreaks cannot be supported by sequential or small-scale parallel execution, making it necessary to use large-scale parallel processing. Here, an optimistic parallel execution of a new discrete event formulation of a reaction–diffusion simulation model of epidemic propagation is presented to facilitate a dramatic increase in the fidelity and speed by which epidemiological simulations can be performed. Rollback support needed during optimistic parallel execution is achieved by combining reverse computation with a small amount of incremental state saving. Parallel speedup of over 5,500 and other runtime performance metrics of the system are observed with weak-scaling execution on a small (8,192-core) Blue Gene/P system, while scalability with a weak-scaling speedup of over 10,000 is demonstrated on 65,536 cores of a large Cray XT5 system. Scenarios representing large population sizes, with mobility and detailed state evolution modeled at the level of each individual, exceeding several hundreds of millions of individuals in the largest cases, are successfully exercised to verify model scalability.

Accelerating parallel agent-based epidemiological simulations

“…The use of ABMs with explicit agent mobility distinguishes the proposed research from work from those reported by Barrett et al [1], Bisset et al [2], and Perumalla et al [19]. These three investigations use a reaction-diffusion approach involving contact graphs for epidemic modeling.…”

Overcoming the Scalability Challenges of Epidemic Simulations on Blue Waters

“…The work by Permualla et al [4] has the best scaling of any individual level Epidemiology simulation known to the authors, and is based on previous EPISIMDEMICS work [1]. The system uses the same disease model and transmission function, but is based on a hierarchical social network construction similar to [12], a population of homogeneous agents, and lacks the ability to model interventions.…”

“…The fastest known simulation [4] reported a speedup of 10,000 on 64K cores (15.2% efficiency) on Cray XT5. Our approach achieves a speedup of 14,357 (22% efficiency) on the same number of cores on Cray XE6.…”

Lock-free pending event set management in time warp

“…There is a pre-defined travel time between any two locations. A reaction-diffusion model based on [20] has been employed to simulate the intralocation and inter-location spread of the disease. The reaction function determines the intra-location spread of disease.…”