Yi Xiang Committee members: Dr. Sudeep Pasricha (Academic Advisor) Dr. Anura Jayasumana Dr. H. J. Siegel Dr. Michelle Strout
OUTLINE Introduction and Preview of Contributions Contribution I: Semi-Dynamic Scheduling for Independent Tasks: Hybrid Energy Storage, Process Variation, and Thermal Management Contribution II: Template-Based Scheduling for Task Graphs: Slack Reclamation, Soft Errors, and Hard Failures Contribution III: Mixed-Criticality Scheduling on Heterogeneous Cores: Soft Deadline, Near-Threshold/Super-Threshold Computing Outline Conclusion 2
What is Energy Harvesting? • We have been doing this for a very long time: wind + mill windmill Introduction and Preview of Contributions water + wheel waterwheel solar + core solarcore? 3 Chao Li et al., “ SolarCore : Solar Energy Driven Multi-Core Architecture Power Management”, IEEE International Symposium on High Performance Computer Architecture (HPCA), pp. 205-216, 2011.
Energy Harvesting for Electronic Systems • Collect energy from ambient sources • Including solar, radio frequency, magnetic, vibration, thermoelectric, etc. Introduction and Preview of Contributions • To support energy autonomy for electronic devices • Wearable electronics, wireless sensor networks, etc. 4
Application of Energy Harvesting: Project Loon • Limited energy availability for electronic devices: • Electricity grid is not available everywhere Introduction and Preview of Contributions Project Loon o Global network-on-balloon o Provide internet access for rural and remote areas o Float in atmosphere o Powered by solar energy harvesting For many applications energy is not readily available, motivating the use of energy harvesting 5
Energy Harvesting and Batteries User Demands: • Energy constraints for embedded devices: High performance • Limited energy capacity of batteries Large screen size High resolution GPS Camera Introduction and Preview of Contributions Biometric sensors External 24X7 battery life? Battery Pack • Replacing battery can be inconvenient, costly, or even impractical Pervasive computing: o Billions of sensors o Scattered everywhere o Batteries: o costly or impossible maintenance o toxic to environment 6 o Need energy harvesting to achieve energy autonomy
Solar Energy Harvesting • Solar energy harvesting as power supply • Photovoltaic (PV) panel to scavenge energy from solar radiation Introduction and Preview of Contributions • Why choose solar energy harvesting? 7
Why Solar Energy Harvesting? • Advantages of solar energy harvesting as power supply • High power density Energy Source Typical Power Density Introduction and Preview of Contributions 60 μ W/cm 2 thermal gradient 4 μ W/cm 3 vibration 1 μ W/cm 2 radio frequency 100 mW/cm 2 solar radiation • Varied scales of PV panels and systems 8
Challenges with Solar Energy Harvesting for Embedded Systems Harvesting power trace of an PV array in one day Provided by N ational R enewable E nergy L aboratory ( NREL ), Golden, Colorado Introduction and Preview of Contributions • Solar radiation can vary dramatically with environment change • Energy shortage at times • Hard to predict 9 As a result, hard to find a performance- and energy-optimal schedule for workload running on energy harvesting-aware embedded systems
Focus of this Dissertation Design a holistic framework for performance- and energy-optimal scheduling and allocation of workload (task, communication) on energy harvesting-aware multicore embedded system platforms • Solar energy harvesting as the only power source Introduction and Preview of Contributions • Batteries/supercapacitors used for temporal energy storage • Multi-core processors with frequency scaling capability • Real-time periodic workloads with deadline constraints 10
Real-Time Workload with Different Timing Constraints • Hard deadline constraint • Any task miss → total system failure • Firm deadline constraint: Introduction and Preview of Contributions • Every task miss → inevitable performance penalty • Soft deadline constraint: • Each task miss → possible performance penalty Main objective: Minimizing miss rate/penalty for task set with firm/soft deadlines 11
Overview of Proposed Framework Real-Time Workloads independent tasks Objective task graphs minimize miss rate/penalty multithreaded tasks Constraints timing energy Multicore Platforms Semi-Dynamic core variation Workload and Platform temperature homogeneous Management Framework soft error hard error heterogeneous firm deadline soft deadline Energy Harvesting Systems • task-to-core mapping photovoltaic panels • intra-core scheduling t A t D batteries hybrid supercapacitors off • communication mapping • voltage-frequency selection t C t B t F • dynamic power management t E off off 12
OUTLINE Introduction and Preview of Contributions Contribution I: Semi-Dynamic Scheduling for Independent Tasks: Hybrid Energy Storage, Process Variation, and Thermal Management Contribution II: Template-Based Scheduling for Task Graphs: Slack Reclamation, Soft Errors, and Hard Failures Contribution III: Mixed-Criticality Scheduling on Heterogeneous Cores: Soft Deadline, Near-Threshold/Super-Threshold Computing Outline Conclusion 13
Contribution I: Contribution I: Semi-Dynamic Scheduling for Independent Tasks Semi-Dynamic Scheduling for Independent Tasks • Main Objective • To reduce miss rate of independent tasks under varying and stringent energy harvesting conditions • Contributions • A semi-dynamic algorithm ( SDA ) that results in lower task miss rates compared to best known prior work • Efficient utilization of multicore systems • Management of battery/supercapacitor hybrid storage system • Awareness of discrete frequency levels, process variations, thermal issues 14
Workload: Periodic Independent Task Set Contribution I: Semi-Dynamic Scheduling for Independent Tasks • Multiple independent periodic tasks with firm deadlines An example of periodic task set • A task miss: missing deadline of a task instance • Minimize miss rate → utilize energy as efficient as possible 15
Contribution I: Semi-Dynamic Scheduling for Independent Tasks Related Work: Utilization-Based Algorithm (UTB) • Task utilization = execution time at f max task period • f opt = f max × U • The lowest frequency that is sufficient to meet all task deadlines, with E arliest D eadline F irst ( EDF ) scheduling • For energy efficiency: minimize frequency fluctuations • Main drawback: dynamic task dropping and slowing down on energy shortage 16 J. Lu et al., “Scheduling and Mapping of Periodic Tasks on Multi-Core Embedded Systems with Energy Harvesting", IEEE International Green Computing Conference (IGCC), pp. 1-6, 2011.
Contribution I: Semi-Dynamic Scheduling for Independent Tasks Motivation: Address Limitation of UTB 6 tasks finished • Proposed solution: a semi-dynamic window-based scheduling 9 tasks finished • Estimate/predict energy budget • Preset execution strategy with uniform speed HOW? A spike/dip in harvesting power can make the prediction inaccurate. • Divide execution process into schedule windows Any mispredictions can only affect prediction accuracy of one schedule window 17
Proposed Semi-Dynamic Framework Contribution I: Semi-Dynamic Scheduling for Independent Tasks • During system execution for a long duration of time • Slice execution time into time windows of k minutes • At reschedule point, predict/obtain energy budget for next time window • Reject tasks based on energy budget. Then allocate the rest • Execute accepted tasks with uniform optimal frequency Semi-Dynamic : reschedule → execute → reschedule → execute … • k minutes 18
Contribution I: Semi-Dynamic Scheduling for Independent Tasks Experiment Setup • Design a simulation environment to capture workload executing on multicore embedded system platform • Historical weather data (solar radiation intensity, temperature) provided by National Renewable Energy Laboratory (NREL) • System only operates from 6:00 AM to 6:30 PM • 50 periodic task sets are randomly generated for each comparison • Implementation of our proposed S emi- D ynamic A lgorithm ( SDA ), together with Ut ilization- B ased Algorithm (UTB) UTB: Ut ilization- B ased Algorithm J. Lu et al., “Scheduling and mapping of periodic tasks on multi- core embedded systems with energy harvesting”, IGCC 2011. 19
Contribution I: Semi-Dynamic Scheduling for Independent Tasks Simulation Results with Heavy Workload 100% utilization: very intensive workload • SDA outperforms UTB • Advantage expands with increasing number of cores 20 • Up to 70% miss rate reduction compared to UTB
Contribution I: Semi-Dynamic Scheduling for Independent Tasks Advantage of SDA: miss UTB - miss SD A • More miss rate reduction when power budget is low or fluctuating 21
Contribution I: Semi-Dynamic Scheduling for Independent Tasks Related Topic I: Hybrid Energy Storage • Pros and cons of different storage medium types battery-only Supercapacitor-only high energy density low energy density low power density high power density less recharge cycles more recharge cycles Solution: Battery-supercapacitor hybrid storage system 22
Recommend
More recommend
