Wavelet analysis for microprocessor design: Experiences with wavelet-based dI/dt characterization

Russ Joseph*, Zhigang Hu, Margaret Martonosi

*Corresponding author for this work

Research output: Contribution to journalConference articlepeer-review

21 Scopus citations


As microprocessors become increasingly complex, the techniques used to analyze and predict their behavior must become increasingly rigorous. This paper applies wavelet analysis techniques to the problem of dI/dt estimation and control in modern microprocessors. While prior work has considered Bayesian phase analysis, Markov analysis, and other techniques to characterize hardware and software behavior, we know of no prior work using wavelets for characterizing computer systems. The dI/dt problem has been increasingly vexing in recent years, because of aggressive drops in supply voltage and increasingly large relative fluctuations in CPU current dissipation. Because the dI/dt problem has a natural frequency dependence (it is worst in the mid-frequency range of roughly 50-200MHz) it is natural to apply frequency-oriented techniques like wavelets to understand it. Our work proposes (i) an off-line wavelet-based estimation technique that can accurately predict a benchmark's likelihood of causing voltage emergencies, and (ii) an on-line wavelet-based control technique that uses key wavelet coefficients to predict and avert impending voltage emergencies. The off-line estimation technique works with roughly 0.94% error. The on-line control technique reduces false positives in dI/dt prediction, allowing voltage control to occur with less than 2.5% performance overhead on the SPEC benchmark suite.

Original languageEnglish (US)
Pages (from-to)36-46
Number of pages11
JournalIEEE High-Performance Computer Architecture Symposium Proceedings
StatePublished - May 24 2004
EventProceedings - 10th International Symposium on High Performance Computer Architecture - Madrid, Spain
Duration: Feb 14 2004Feb 18 2004

ASJC Scopus subject areas

  • Hardware and Architecture


Dive into the research topics of 'Wavelet analysis for microprocessor design: Experiences with wavelet-based dI/dt characterization'. Together they form a unique fingerprint.

Cite this