Title :
High-Level Switching Activity Prediction Through Sampled Monitored Simulation
Author :
Klein, Felipe ; Azevedo, Rodolfo ; Araujo, Guido
Author_Institution :
State Univ. of Campinas, Campinas
Abstract :
This paper presents a sample-based technique to predict the switching activity of digital circuits. It is an improvement to PowerSC, a SystemC library extension that enables a fast and easy-to-use way of gathering switching activity from SystemC descriptions. The experimental results reveal that it can dramatically reduce the monitoring time of the simulation, with a minimal loss of accuracy with respect to estimates provided by an industrial tool. Several tests realized in a case study with a real-world design obtained reductions in the monitoring time of up to 99% with average errors of no more than 0.05%.
Keywords :
circuit simulation; hardware description languages; software libraries; switching circuits; PowerSC SystemC library extension; digital circuits; high-level switching activity prediction; sampled monitored simulation; Circuit simulation; Computational modeling; Condition monitoring; Costs; Design optimization; Digital circuits; Energy consumption; Predictive models; Software libraries; Switching circuits;
Conference_Titel :
System-on-Chip, 2005. Proceedings. 2005 International Symposium on
Conference_Location :
Tampere
Print_ISBN :
0-7803-9294-9
DOI :
10.1109/ISSOC.2005.1595668