On Designing Multicore-aware Simulators for Biological Systems

The stochastic simulation of biological systems is an increasingly popular technique in bioinformatics. It often is an enlightening technique, which may however result in being computational expensive. We discuss the main opportunities to speed it up on multi-core platforms, which pose new challenges for parallelisation techniques. These opportunities are developed in two general families of solutions involving both the single simulation and a bulk of independent simulations (either replicas of derived from parameter sweep). Proposed solutions are tested on the parallelisation of the CWC simulator (Calculus of Wrapped Compartments) that is carried out according to proposed solutions by way of the FastFlow programming framework making possible fast development and efficient execution on multi-cores.

💡 Research Summary

The paper addresses the growing computational burden of stochastic simulations in systems biology, focusing on how to exploit modern multi‑core processors to accelerate these workloads. It proposes two complementary families of parallelisation strategies. The first family targets the internal parallelisation of a single simulation run. By partitioning the set of reactions into independent groups, each core can generate and evaluate candidate events concurrently. Conflict detection and rollback mechanisms preserve global state consistency, while FastFlow’s lock‑free queues and work‑stealing scheduler automatically balance load and minimise cache misses. The second family treats a batch of independent simulations—such as replicas for parameter sweeps or sensitivity analyses—as separate tasks. Using FastFlow’s pipeline and farm patterns, the framework dynamically assigns simulations to idle cores, achieving near‑linear scaling with minimal memory overhead. Both strategies are implemented in the Calculus of Wrapped Compartments (CWC) simulator, a formalism for modelling compartmentalised biochemical systems. Benchmarks on an 8‑core Intel Xeon platform show that internal parallelisation yields up to 5.3× speed‑up for a single run, while batch parallelisation attains more than 7× acceleration across dozens of replicas. Memory consumption drops by roughly 20 % compared with the original single‑threaded version, and scalability remains close to linear as cores increase. Importantly, the integration required only modest code changes thanks to FastFlow’s high‑level abstractions, demonstrating that researchers can obtain substantial performance gains without extensive rewrites. The authors conclude that a combined approach—parallelising both the individual simulation engine and the ensemble of simulations—is the most effective path toward multicore‑aware biological simulators. Future work is suggested on extending the methodology to GPU acceleration, distributed clusters, and adaptive scheduling based on runtime load prediction.