Argonne’s Kevin Brown presents work at flagship simulation conference
Kevin A. Brown, an assistant computer scientist in the Mathematics and Computer Science division at the U.S. Department of Energy’s Argonne National Laboratory, co-authored three studies presented at the 39thACM SIGSIM Conference on Principles of Advanced Discrete Simulation (SIGSIM-PADS ’25). PADS is considered the premier academic conference for the simulation and modeling community.
In the session on Simulation Algorithms and Methods, Brown presented a short paper on combining parallel discrete event simulation (PDES) with surrogate models. The approach, which uses a “director” to facilitate the synchronization, incurs less than 2% overhead while enabling nearly a 2X speedup in hybrid simulations compared with PDES-only simulations. This novel framework can be leveraged to accelerate not just PDES but other high-fidelity simulation methodologies across various scientific domains.
Related Paper: Kevin A. Brown, Elkin Cruz-Camacho, Kazutomo Yoshii, Xin Wang, Zhiling Lan, Christopher D. Carothers, and Robert B. Ross, “Directing PDES and Surrogate Models in Loosely Coupled Hybrid Simulations,” doi: 10.1145/3726301.3728419
In the session on Simulation of Computer Networks, Brown was lead on a short paper describing a novel approach for developing fluid models that capture critical features of high-performance computing networks to predict the steady states of multiple conflicting flows. This work represents a first step toward developing fluid models for high-performance computing networks.
Related Paper: Kevin A. Brown, Andes Lopez, and Jason Liu, “Toward a Steady-State Fluid Model of HPC Networks,” doi: 10.1145/3726301.3728417
Also at the Simulation of Computer Networks session, Brown co-authored a study reporting on MFNetSim, a multifidelity framework that models both I/O and communication traffic on high-performance computing systems. MFNetSim’s design addresses critical trade-offs between accuracy and scalability and between stable performance and fast performance.
Related Paper: Xin Wang, Kevin A. Brown, Robert B. Ross, Christopher D. Carothers, and Zhiling Lan, “MFNetSim: A Multi-Fidelity Network Simulation Framework for Multi-Traffic Modeling of Dragonfly Systems,” doi: 10.1145/3726301.3728417
Argonne National Laboratory seeks solutions to pressing national problems in science and technology by conducting leading-edge basic and applied research in virtually every scientific discipline. Argonne is managed by UChicago Argonne, LLC for the U.S. Department of Energy’s Office of Science.
The U.S. Department of Energy’s Office of Science is the single largest supporter of basic research in the physical sciences in the United States and is working to address some of the most pressing challenges of our time. For more information, visit https://energy.gov/science.