GPU-Based Parallel Computing for Fast Circuit Optimization
Chapter
Overview
Identity
Additional Document Info
Other
View All
Overview
abstract
This chapter exploits the GPU to accelerate VLSI circuit optimization. Fast circuit optimization technique is used during chip design. Although the pressure of time-to-market is almost never relieved, design complexity keeps growing along with transistor count. In addition, more and more issues need to be considered-from conventional objectives like performance and power to new concerns like process variability and transistor aging. On the other hand, the advancement of chip technology opens new avenues for boosting computing power. In this work, GPU-based parallel computing techniques is proposed for simultaneous gate sizing and threshold voltage assignment. Gate sizing is a classic approach for optimizing performance and power of combinational circuits. Different size implementations of a gate realize the same gate logic, but present a trade-offbetween the gate's input capacitance and output resistance, thus affecting the signal propagation delay on the gate's fan-in and fan-out paths, as well as the balance between timing performance and power dissipation of the circuit. It has long been a challenge to optimize a combinational circuit in a systematic, yet fast manner owing to its topological reconvergence and large size. A recent progress suggests an effective solution to the reconvergence problem. This work addresses the large problem size by exploiting GPU-based parallelism. The proposed parallel techniques are integrated with the state-of-the-art gate sizing and Vt assignment algorithm. 2011 Copyright 2011 NVIDIA Corporation and Wen-mei W. Hwu Published by Elsevier Inc. All rights reserved..
