21-02-2012, 04:09 PM
Power Optimization and Management in Embedded Systems1
[attachment=17639]
Introduction
A system consists of a set of components that provide a useful
behavior or service. System design is the process of
implementing the desired behavior or service while optimizing
some objectives and satisfying some constraints. For example,
a high-end desktop computer system is optimized for maximum
clock rate while meeting a price target. On the other hand, a
battery-operated portable electronic system for information
processing and wireless communication may be optimized for
minimum power dissipation subject to a required performance.
Power-aware software compilation
Power analysis techniques have been proposed for embedded
software based on instruction-level characterization [1] and
simulation of the underlying hardware [2]. Techniques to
improve the efficiency of software power analysis through
statistical profiling have been proposed in [3]. The SOC design
paradigm has fueled research in co-estimation of hardware and
software power consumption [4]-[7]. Optimizations of
embedded software power consumption through compiler
optimizations [8], source-level transformations [9], and
memory management schemes [10] [11] have been reported.
Dynamic power management
Dynamic power management – which refers to selective shutoff
or slow-down of system components that are idle or
underutilized – has proven to be a particularly effective
technique for reducing power dissipation in such systems.
Incorporating a dynamic power management scheme in the
design of an already-complex system is a difficult process that
may require many design iterations and careful debugging and
validation. The goal of a dynamic power management policy is
to reduce the power consumption of an electronic system by
putting system components into different states, each
representing a certain performance and power consumption
level. The policy determines the type and timing of these
transitions based on the system history, workload, and
performance constraints.
Low power bus encoding
The major building blocks of a computer system include the
processor, the memory chips, and the communication channels
dedicated to providing the means for data transfer between the
processor and the memory. These channels tend to support
heavy traffic and often constitute the performance bottleneck in
many systems. At the same time, the energy dissipation per
memory bus access is quite high, which in turn limits the power
efficiency of the overall system.
