EECE **** Embedded System Design. Embedded Systems. Power/Energy Aware Embedded Systems Dynamic Voltage Scheduling Dynamic Power Management. Power Consumption of a Gate. Recap from chapter 3: Fundamentals of dynamic voltage scaling (DVS).
Download Policy: Content on the Website is provided to you AS IS for your information and personal use and may not be sold / licensed / shared on other websites without getting consent from its author.While downloading, if for some reason you are not able to download a presentation, the publisher may have deleted the file from their server.
Power consumption of CMOScircuits (ignoring leakage):
Delay for CMOS circuits:
Decreasing Vdd reduces P quadratically,while the run-time of algorithms is only linearly increased(ignoring the effects of the memory system).
Saving Energy under given Time Constraints:
Energy / Cycle [nJ]
Normal Mode scaling (DVS)
Spatial vs. Dynamic
Supply Voltage Management
Analogy of biological blood systems:
Not all components require
Required performance may change over time
• Static Voltage Scheduling
• Dynamic Voltage Scheduling
Voltage Control and Task Scheduling by Operating System to minimize energy consumption
Okuma, Ishihara, and Yasuura: “Real-Time Task Scheduling for a Variable Voltage Processor”, Proc. of the 1999 International Symposium on System Synthesis (ISSS'99)
Problem for Operating Systems scaling (DVS)
What is the optimum supply voltage assignment for each task in order to obtain minimum energy consumption?
device states scaling (DVS)
Tsd: shutdown delay Twu: wakeup delay
Tbs: time before shutdown Tbw: time before wakeupThe opportunity:Reduce power according to workload
Desired: Shutdown only during long idle times
Tradeoff between savings and overhead
B scaling (DVS)
timeAdaptive Stochastic Models
Sliding Window (SW): [Chung DATE 99]
P : average power
Nsd: number of shutdowns
Nwd : wrong shutdowns (actually waste energy)
Coordinate multiple workload sources
Power manager is notified when processes change state
We use processes to represent requesters
requester = process
t scaling (DVS)1
Rearrange task execution to cluster
similar utilization and idle periods
T: time quantum
Device-centric, shutdown based
Good research platform (several partial implementations, es. U. Delft, Compaq, etc.)
Quite high-overhead for low-end embedded systems
Good research platform (HP-Unibo implementation)
Lower overhead than Linux, modular
2.5 hours scaling (DVS)Application Aware DPM –Example: Communication Power
NICs powered by portables reduce battery life
Higher bit rates lead to higher power consumption
Proper use of PHY layer services by MAC is critical!
Off mode power savings scaling (DVS)
Low water mark
LWM / Buffer characteristics scaling (DVS)
Where to put the LWM?
How long should the buffer be?
Buffering Strategies should be Power Aware!
C scaling (DVS)omparison
Approximate processing [Chandrakasan98-01]
Tradeoff quality for energy (es. lossy compression)
Design algorithms for graceful degradation
Enforce power-efficiency in programming
Avoid repetitive polling [Intel98]
Use event-based activation (interrupts)
Localize computation whenever possible
Helps shutdown of peripherals
Helps shutdown of memories