Efficient RTL Power Estimation for Large Designs

  • Authors:
  • Srivaths Ravi;Anand Raghunathan;Srimat Chakradhar

  • Affiliations:
  • -;-;-

  • Venue:
  • VLSID '03 Proceedings of the 16th International Conference on VLSI Design
  • Year:
  • 2003

Quantified Score

Hi-index 0.00

Visualization

Abstract

The adoption of register-transfer level (RTL) sign-off in ASIC designmethodologies, and the increasing scale of system-on-chip integration,are leading to unprecedented accuracy and efficiency demandson RT-level estimation tools. In this work, we focus onthe deployment of a simulation-based RTL power estimation toolin a commercial design flow, and describe several enhancementsthat improve its efficiency and scalability for large, industrial designs.We profile the computational effort involved in RTL powerestimation, and propose a suite of acceleration techniques, including(i) transformation of the enhanced RTL description (functionalmodel with annotations for power estimation) to be more simulator-friendly,(ii) computation vs. storage tradeoffs, and (iii) a novelvariation of statistical sampling, called partitioned sampling. Ourtechniques result in an optimized allocation of the overall computationaleffort for power estimation and minimize the computationaleffort involved in the evaluation of power models.Extensive experimental results in the context of a commercialdesign flow have yielded promising results (e.g., upto 31X reductionin power estimation time with negligible loss of accuracy) onindustrial designs of upto 1.25 million transistors. In addition toaccurate power estimation for an entire circuit, these accelerationtechniques result in superior accuracy of local power estimates forindividual components or small sub-circuits, compared to conventionalsampling or test-bench compaction techniques.