Run-time platform variability presents a number of challenges to the system software in order that a run-time environment is presented to applications that sufficiently masks dynamic platform variability (including fabrication variability), whilst allowing applications to tune overall system performance to exploit key aspects of dynamic energy usage and platform variability. The approach taken within the Touchmore project is to model key aspects of the platform in order that performance and variability can be understood and exploited by the system software. In turn, the system software (comprising OS and run-time) utilises the model so that aspects of variability and energy usage are abstracted from the platform, then monitored and controlled in order to meet policy goals, eg. energy minimisation. This paper documents aspects of the modeling and system software structure to show how the Touchmore project is managing energy and platform variability using customisation of the application, system software and toolchain.
Neil C. Audsley, Ian Gray, ACQUAVIVA, A., Ralph Haines (2012). ToucHMore toolchain and system software for energy and variability customisation. IEEE - INST ELECTRICAL ELECTRONICS ENGINEERS INC [10.1109/HLDVT.2012.6418257].
ToucHMore toolchain and system software for energy and variability customisation
ACQUAVIVA, ANDREA;
2012
Abstract
Run-time platform variability presents a number of challenges to the system software in order that a run-time environment is presented to applications that sufficiently masks dynamic platform variability (including fabrication variability), whilst allowing applications to tune overall system performance to exploit key aspects of dynamic energy usage and platform variability. The approach taken within the Touchmore project is to model key aspects of the platform in order that performance and variability can be understood and exploited by the system software. In turn, the system software (comprising OS and run-time) utilises the model so that aspects of variability and energy usage are abstracted from the platform, then monitored and controlled in order to meet policy goals, eg. energy minimisation. This paper documents aspects of the modeling and system software structure to show how the Touchmore project is managing energy and platform variability using customisation of the application, system software and toolchain.I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.