Jui-Ming Chang

Bio: Jui-Ming Chang is an academic researcher from Cadence Design Systems. The author has contributed to research in topics: Dynamic programming & Energy consumption. The author has an hindex of 12, co-authored 14 publications receiving 1011 citations. Previous affiliations of Jui-Ming Chang include University of Southern California & Hewlett-Packard.

Energy minimization using multiple supply voltages

Abstract: We present a dynamic programming technique for solving the multiple supply voltage scheduling problem in both nonpipelined and functionally pipelined data-paths. The scheduling problem refers to the assignment of a supply voltage level (selected from a fixed and known number of voltage levels) to each operation in a data flow graph so as to minimize the average energy consumption for given computation time or throughput constraints or both. The energy model is accurate and accounts for the input pattern dependencies, re-convergent fanout induced dependencies, and the energy cost of level shifters. Experimental results show that using three supply voltage levels on a number of standard benchmarks, an average energy saving of 40.19% (with a computation time constraint of 1.5 times the critical path delay) can be obtained compared to using a single supply voltage level.

Register Allocation and Binding for Low Power

Abstract: This paper describes a technique for calculating the switching activity of a set of registers shared by different data values. Based on the assumption that the joint pdf (probability density function) of the primary input random variables is known or that a suffficiently large number of input vectors has been given, the register assignment problem for minimum power consumption is formulated as a minimum cost clique covering of an appropriately defined compatibility graph (which is shown to be transitively orientable). The problem is then solved optimally (in polynomial time) using a max-cost ow algorithm. Experimental results confirm the viability and usefulness of the approach in minimizing power consumption during the register assignment phase of the behavioral synthesis process.

Energy minimization using multiple supply voltages

Abstract: We present a dynamic programming technique for solving the multiple supply voltage scheduling problem in both non-pipelined and functionally pipelined data-paths. The scheduling problem refers to the assignment of a supply voltage level to each operation in a data flow graph so as to minimize the average energy consumption for given computation time or throughput constraints or both. The energy model is accurate and accounts for the input pattern dependencies, re-convergent fanout induced dependencies, and the energy cost of level shifters. Experimental results show that using four supply voltage levels on a number of standard benchmarks, an average energy saving of 53% (with a computation time constraint of 1.5 times the critical path delay) can be obtained compared to using one fixed supply voltage level.

Module assignment for low power

Abstract: We investigate the problem of minimizing the total power consumption during the binding of operations to functional units in a scheduled data path with functional pipelining and conditional branching for data intensive applications. We first present a technique to estimate the power consumption in a functionally pipelined data path and then formulate the power optimization problem as a max cost multi commodity flow problem and solve it optimally. Our proposed method can augment most high level synthesis algorithms as a post processing step for reducing power after the optimizations for area or speed have been completed. An average power savings of 28% has been observed after we apply our method to pipelined designs that have been optimized using conventional techniques.

A Gaussian synapse circuit for analog VLSI neural networks

Abstract: Back-propagation neural networks with Gaussian function synapses have better convergence property over those with linear-multiplying synapses. In digital simulation, more computing time is spent on Gaussian function evaluation. We present a compact analog synapse cell which is not biased in the subthreshold region for fully-parallel operation. This cell can approximate a Gaussian function with accuracy around 98% in the ideal case. Device mismatch induced by fabrication process will cause some degradation to this approximation. The Gaussian synapse cell can also be used in unsupervised learning. Programmability of the proposed Gaussian synapse cell is achieved by changing the stored synapse weight W/sub ji/, the reference current and the sizes of transistors in the differential pair. >

Voltage scheduling problem for dynamically variable voltage processors

Abstract: This paper presents a model of dynamically variable voltage processors and basic theorems for power-delay optimization. A static voltage scheduling problem is also proposed and formulated as an integer linear programming (ILP) problem. In the problem, we assume that a core processor can vary its supply voltage dynamically, but can use only a single voltage level at a time. For a given application program and a dynamically variable voltage processor, a voltage scheduling which minimizes energy consumption under an execution time constraint can be found.

Energy- and performance-aware mapping for regular NoC architectures

Abstract: In this paper, we present an algorithm which automatically maps a given set of intellectual property onto a generic regular network-on-chip (NoC) architecture and constructs a deadlock-free deterministic routing function such that the total communication energy is minimized. At the same time, the performance of the resulting communication system is guaranteed to satisfy the specified design constraints through bandwidth reservation. As the main theoretical contribution, we first formulate the problem of energy- and performance-aware mapping in a topological sense, and show how the routing flexibility can be exploited to expand the solution space and improve the solution quality. An efficient branch-and-bound algorithm is then proposed to solve this problem. Experimental results show that the proposed algorithm is very fast, and significant communication energy savings can be achieved. For instance, for a complex video/audio application, 51.7% communication energy savings have been observed, on average, compared to an ad hoc implementation.

Energy-aware mapping for tile-based NoC architectures under performance constraints

Abstract: In this paper, we present an algorithm which automatically maps the IPs/cores onto a generic regular Network on Chip (NoC) architecture such that the total communication energy is minimized At the same time, the performance of the mapped system is guaranteed to satisfy the specified constraints through bandwidth reservation As the main contribution, we first formulate the problem of energy-aware mapping, in a topological sense, and then propose an efficient branch-and-bound algorithm to solve it Experimental results show that the proposed algorithm is very fast and robust, and significant energy savings can be achieved For instance, for a complex video/audio SoC design, on average, 604% energy savings have been observed compared to an ad-hoc implementation

A Survey of Neuromorphic Computing and Neural Networks in Hardware.

Abstract: Neuromorphic computing has come to refer to a variety of brain-inspired computers, devices, and models that contrast the pervasive von Neumann computer architecture This biologically inspired approach has created highly connected synthetic neurons and synapses that can be used to model neuroscience theories as well as solve challenging machine learning problems The promise of the technology is to create a brain-like ability to learn and adapt, but the technical challenges are significant, starting with an accurate neuroscience model of how the brain works, to finding materials and engineering breakthroughs to build devices to support these models, to creating a programming framework so the systems can learn, to creating applications with brain-like capabilities In this work, we provide a comprehensive survey of the research and motivations for neuromorphic computing over its history We begin with a 35-year review of the motivations and drivers of neuromorphic computing, then look at the major research areas of the field, which we define as neuro-inspired models, algorithms and learning approaches, hardware and devices, supporting systems, and finally applications We conclude with a broad discussion on the major research topics that need to be addressed in the coming years to see the promise of neuromorphic computing fulfilled The goals of this work are to provide an exhaustive review of the research conducted in neuromorphic computing since the inception of the term, and to motivate further work by illuminating gaps in the field where new research is needed