北京大学学报(自然科学版)

基于贝叶斯推理和向量压缩技术的最大功耗分析

陈杰,李险峰,佟冬,王克义,程旭   

  1. 北京大学微处理器研发中心,北京100871;
  • 收稿日期:2008-03-05 出版日期:2009-03-20 发布日期:2009-03-20

Maximum Power Analysis Based on Bayesian Inference and Vector Compression Techniques

CHEN Jie, LI Xianfeng, TONG Dong, WANG Keyi, CHENG Xu   

  1. Microprocessor Research & Development Center, Peking University, Beijing 100871;
  • Received:2008-03-05 Online:2009-03-20 Published:2009-03-20

摘要: 针对模拟评测电路最大功耗分析速度缓慢的问题,使用贝叶斯推理功耗模型和切片分析技术进行向量压缩,优选出可能生成最大功耗的向量进行详细分析。进一步的,基于输入信号翻转密度和最大功耗生成之间的关系分析,设计自适应翻转密度与向量生成平台,结合贝叶斯向量压缩技术进行最大功耗评测。实验表明,基于切片分析的贝叶斯模型向量压缩平均加速比达1005倍,分析误差2.40%;结合自适应翻转密度计算与向量压缩的评测方法平均加速比达163倍,最大功耗分析结果相对原始序列提高1.99%。

关键词: 最大功耗评测, 贝叶斯推理, 切片分析, 翻转密度计算, 向量压缩

Abstract: To resolve oversize time consuming problem in simulation based maximum power analysis, Bayesian power model based on slice analysis is proposed. This model selects the input vector set which may generate maximum power and performs accurate power estimation for the compact sequence. The relationship between signal switch density and maximum power generation is analyzed, and then an input vector generation platform with switching density self-adaptation computing and Bayesian vector compression is proposed. The experimental results indicate that, Bayesian vector compression method results in 1005 times average estimation time speed-ups, and the average maximum-power error is 2.40%. When using vector generation method based on self-adaptation computation and Bayesian vector compression, the maximum power bottom limit can be increased with 1.99%, and average speed-ups reaches 163 times.

Key words: maximum power analyses, Bayesian inference, slice analysis, switch density computing, vector compression

中图分类号: