Exploring Execution Schemes for Agent-Based Traffic Simulation on Heterogeneous Hardware

Abstract

Microscopic traffic simulation is associated with substantial runtimes, limiting the feasibility of large-scale evaluation of traffic scenarios. Even though today heterogeneous hardware comprised of CPUs, graphics processing units (GPUs) and fused CPU-GPU devices is inexpensive and widely available, common traffic simulators still rely purely on CPU-based execution, leaving substantial acceleration potentials untapped. A number of existing works have considered the execution of traffic simulations on accelerators, but have relied on simplified models of road networks and driver behaviour tailored to the given hardware platform. Thus, the existing approaches cannot directly benefit from the vast body of research on the validity of common traffic simulation models. In this paper, we explore the performance gains achievable through the use of heterogeneous hardware when relying on typical traffic simulation models used in CPUbased simulators. We propose a partial offloading approach that relies either on a dedicated GPU or a fused CPU-GPU device. Further, we present a traffic simulation running fully on a manycore GPU and discuss the challenges of this approach. Our results show that a CPU-based parallelisation closely approaches the results of partial offloading, while full offloading substantially outperforms the other approaches. We achieve a speedup of up to 28.7x over the sequential execution on a CPU.

OpenABLext: An automatic code generation framework for agent‐based simulations on CPU‐GPU‐FPGA heterogeneous platforms

“…As a proof of concept, we developed a traffic simulation based on our previous implementation. 49 Our results are shown in Figure 7D. Note that we do not include MASON or C with OpenMP as these backends are unable to support graph-based simulation spaces.…”

“…We based our implementation on the OpenABL code provided in the OpenABL repository (https://github.com/OpenABL/OpenABL). To gain insights into the performance of our co‐execution method as well as the online dispatcher, we utilize two additional simulation models: Traffic , a traffic simulation based on the implementation in Reference 49 and Crowd , which simulates the flocking behavior of people following leaders (fire wardens) during a building evacuation 50 . While the Traffic simulation is graph‐based, all others exhibit 2D/3D simulation spaces.…”

“…As described above, OpenABLext enables modelers to utilize graph‐based simulation environments. As a proof of concept, we developed a traffic simulation based on our previous implementation 49 7D.…”

Causality and Consistency of State Update Schemes in Synchronous Agent-based Simulations

“…To implement the two-state access graph, two sets of state variables are used, 1 and 2 . In contrast, the independent and ordered updates 4, agent or can be processed at the same time, e.g., making the update scheme suitable for execution on parallel computing platforms, e.g., GPUs or FPGAs [39,40].…”

Masivo: Parallel Simulation Model Based on OpenCL for Massive Public Transportation Systems’ Routes

“…For these reasons, novel works have developed new models and simulators that employ parallel architectures to get the simulation results more efficiently. Currently, we find mesoscopic [13,[56][57][58][59][60] and microscopic [61] traffic simulators based in GPUs (Graphics Processing Units), HPC (High Performance Computing) [62][63][64], and microscopic simulation based on Apache Spark [55,65]. Nevertheless, these modeling and simulation approaches do not consider each passenger as a unique entity in the simulation model.…”