A Programmable Hardware Accelerator for Simulating Dynamical Systems
- 24 June 2017
- journal article
- conference paper
- Published by Association for Computing Machinery (ACM) in ACM SIGARCH Computer Architecture News
- Vol. 45 (2), 403-415
- https://doi.org/10.1145/3140659.3080252
Abstract
The fast and energy-efficient simulation of dynamical systems defined by coupled ordinary/partial differential equations has emerged as an important problem. The accelerated simulation of coupled ODE/PDE is critical for analysis of physical systems as well as computing with dynamical systems. This paper presents a fast and programmable accelerator for simulating dynamical systems. The computing model of the proposed platform is based on multilayer cellular nonlinear network (CeNN) augmented with nonlinear function evaluation engines. The platform can be programmed to accelerate wide classes of ODEs/PDEs by modulating the connectivity within the multilayer CeNN engine. An innovative hardware architecture including data reuse, memory hierarchy, and near-memory processing is designed to accelerate the augmented multilayer CeNN. A dataflow model is presented which is supported by optimized memory hierarchy for efficient function evaluation. The proposed solver is designed and synthesized in 15nm technology for the hardware analysis. The performance is evaluated and compared to GPU nodes when solving wide classes of differential equations and the power consumption is analyzed to show orders of magnitude improvement in energy efficiency.Keywords
Funding Information
- National Science Foundation (Career Award)
- Office of Naval Research (Young Investigator Award)
This publication has 35 references indexed in Scilit:
- Information Processing Capacity of Dynamical SystemsScientific Reports, 2012
- 24-GOPS 4.5-${\rm mm}^{2}$ Digital Cellular Neural Network for Rapid Visual Attention in an Object-Recognition SoCIEEE Transactions on Neural Networks, 2010
- Modeling, stability and control of biped robots—a general frameworkAutomatica, 2004
- Spatial Pattern Formation via Reaction–Diffusion Dynamics in 32$,times,$32$,times,$4 CNN ChipIEEE Transactions on Circuits and Systems I: Regular Papers, 2004
- Simple model of spiking neuronsIEEE Transactions on Neural Networks, 2003
- Linear algebra operators for GPU implementation of numerical algorithmsACM Transactions on Graphics, 2003
- A programmable analog cellular neural network CMOS chip for high speed image processingIEEE Journal of Solid-State Circuits, 1995
- Simulating nonlinear waves and partial differential equations via CNN. I. Basic techniquesIEEE Transactions on Circuits and Systems I: Regular Papers, 1995
- Autonomous cellular neural networks: a unified paradigm for pattern formation and active wave propagationIEEE Transactions on Circuits and Systems I: Regular Papers, 1995
- On the analog computer solution of first-order partial differential equationsMathematics and Computers in Simulation, 1965