BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//Talks.cam//talks.cam.ac.uk//
X-WR-CALNAME:Talks.cam
BEGIN:VEVENT
SUMMARY:Predictive Models for Microarchitectural Adaptation - Dr Timothy J
 ones\, University of Cambridge Computer Lab.
DTSTART:20110204T160000Z
DTEND:20110204T170000Z
UID:TALK29630@talks.cam.ac.uk
CONTACT:Prof Simon Moore
DESCRIPTION:Adaptive microarchitectures are a promising solution for desig
 ning high-performance\, power-efficient microprocessors.  They offer the a
 bility to tailor computational resources to the specific requirements of d
 ifferent programs or program phases.  They have the potential to adapt the
  hardware cost-effectively at runtime to any application's needs.  However
 \, one of the key challenges is how to dynamically determine the best arch
 itecture configuration at any given time\, for any \nnew workload.\n\nThis
  talk will describe a novel control mechanism based on a predictive model 
 for microarchitectural adaptivity control.  This model is able to efficien
 tly control adaptivity by monitoring the behaviour of an application's dif
 ferent phases at runtime.  It will show that using this  model on SPEC 200
 0 can double the energy/performance efficiency of the processor when compa
 red to the best static configuration tuned for the \nwhole benchmark suite
 .  This represents 74% of the improvement available if we knew the best mi
 croarchitecture for each program phase ahead of \ntime.  In addition\, it 
 will show that the overheads associated with the  implementation of this s
 cheme have a negligible impact on performance and power.\n
LOCATION:SC04\, Computer Laboratory\, William Gates Building
END:VEVENT
END:VCALENDAR