Adventures in Signal Selection

Add to your list(s) Download to your calendar using vCal

• James Davis (Imperial College)
• Monday 27 March 2017, 15:00-16:00
• Level 9 Seminar Room, EE Dept..

If you have a question about this talk, please contact George A Constantinides.

Last year, PRiME researchers introduced KAPow, a framework enabling the automatic and transparent instrumentation of arbitrary hardware systems, implemented on FPG As, for runtime power estimation of their constituent components. It was shown to be capable of achieving +/-5mW estimation accuracy and the ability to adapt to environmental and workload changes. But KAPow has a problem: often, many more signals have to be instrumented than is strictly necessary to achieve a particular estimation error. This is due to its rudimentary signal selection, based solely on predicted switching activity and devoid of information pertaining to circuit structure.

In this talk, I’ll present some of the alternative, more intelligent signal selection strategies we’ve been looking at for KAPow, including those that analyse network flow and statistical methods based on the outputs of automated simulations. We are hopeful that one or more of these will be useful in maintaining low error while simultaneously reducing the costs (in compilation time, area and/or power overheads) of instrumenting hardware systems for power estimation.

This talk is part of the CAS Talks series.

This talk is included in these lists:

• Andrea Picciau's list
• CAS Talks
• Circuits and Systems Group: Internal Seminars
• Level 9 Seminar Room, EE Dept.

Note that ex-directory lists are not shown.