WattWatcher: Fine-Grained Power Estimation on Live Multicore Systems Using Configurable Models
Power is a first-order design constraint in any modern computing platform. Extensive research has focused on estimating power to guide advances in power management schemes, thermal hot spots, and voltage noise. However, most power estimation tools possess drawbacks related to the level of detail and amount of training needed for reliable results. This paper introduces WattWatcher, a power measurement framework that provides fine-grained power traces by passing predefined performance counters and a hardware descriptor file into configurable back-end power models. We show that WattWatcher has a mean absolute percentage error of 2.67% aggregated over all benchmarks when compared to measured power consumption on SPEC CPU 2006 and PARSEC benchmarks across three different machines.
Sunday, Sept. 20, 2015, 8 a.m. — Tuesday, Sept. 22, 2015, 10 p.m. CT
Austin, TX, United States
Technical conference and networking event for SRC members and students.