*banner
 

Controlling Timing vs. Measuring Timing
Edward A. Lee

Citation
Edward A. Lee. "Controlling Timing vs. Measuring Timing". Talk or presentation, 7, June, 2015; Keynote, Workshop on Suite of Embedded Applications and Kernels (SEAK) Design Automation Conference (DAC) San Francisco, CA.

Abstract
Embedded systems differ from many general purpose and scientific computing applications in that repeatability is often more important than performance. The goal in an embedded system is not usually to get a task done as soon as possible, but rather to get a task done reliably, on time, and with minimal energy consumption. Benchmarking, however, is typically focused on performance, not repeatability. In this talk, I will argue that when the primary goal is repeatability, design decisions can be very different, and benchmarking needs to be done differently. I will describe the Berkeley PRET project, which shows that embedded processors can be designed to deliver repeatable timing, and with appropriately adjusted measures, with no loss in performance.

Electronic downloads

Citation formats  
  • HTML
    Edward A. Lee. <a
    href="http://chess.eecs.berkeley.edu/pubs/1103.html"
    ><i>Controlling Timing vs. Measuring
    Timing</i></a>, Talk or presentation,  7, June,
    2015; Keynote, Workshop on Suite of Embedded Applications
    and Kernels (SEAK) Design Automation Conference (DAC) San
    Francisco, CA.
  • Plain text
    Edward A. Lee. "Controlling Timing vs. Measuring
    Timing". Talk or presentation,  7, June, 2015; Keynote,
    Workshop on Suite of Embedded Applications and Kernels
    (SEAK) Design Automation Conference (DAC) San Francisco, CA.
  • BibTeX
    @presentation{Lee15_ControllingTimingVsMeasuringTiming,
        author = {Edward A. Lee},
        title = {Controlling Timing vs. Measuring Timing},
        day = {7},
        month = {June},
        year = {2015},
        note = {Keynote, Workshop on Suite of Embedded
                  Applications and Kernels (SEAK) Design Automation
                  Conference (DAC) San Francisco, CA.},
        abstract = {Embedded systems differ from many general purpose
                  and scientific computing applications in that
                  repeatability is often more important than
                  performance. The goal in an embedded system is not
                  usually to get a task done as soon as possible,
                  but rather to get a task done reliably, on time,
                  and with minimal energy consumption. Benchmarking,
                  however, is typically focused on performance, not
                  repeatability. In this talk, I will argue that
                  when the primary goal is repeatability, design
                  decisions can be very different, and benchmarking
                  needs to be done differently. I will describe the
                  Berkeley PRET project, which shows that embedded
                  processors can be designed to deliver repeatable
                  timing, and with appropriately adjusted measures,
                  with no loss in performance.},
        URL = {http://chess.eecs.berkeley.edu/pubs/1103.html}
    }
    

Posted by Mary Stewart on 22 Jun 2015.
For additional information, see the Publications FAQ or contact webmaster at chess eecs berkeley edu.

Notice: This material is presented to ensure timely dissemination of scholarly and technical work. Copyright and all rights therein are retained by authors or by other copyright holders. All persons copying this information are expected to adhere to the terms and constraints invoked by each author's copyright.

©2002-2018 Chess