Low Power Design Thomas Ebi and Prof. Dr. J. Henkel Thomas Ebi and - PowerPoint PPT Presentation

1 introduction Low Power Design Thomas Ebi and Prof. Dr. J. Henkel Thomas Ebi and Prof. Dr. J. Henkel CES CES - Chair for Embedded Systems Chair for Embedded Systems KIT, Germany KIT, Germany I. Introduction and Energy/Power Sources I. Introduction and Energy/Power Sources http://ces.itec.kit.edu T. Ebi and J. Henkel, KIT, SS13

2 introduction Overview: today  Reason for Low Power Design: motivation Reason for Low Power Design: motivation  Specific need for low power in embedded systems: Specific need for low power in embedded systems: examples examples  Battery issues (re Battery issues (re-chargeable batteries) chargeable batteries)  Power/energy sources Power/energy sources http://ces.itec.kit.edu T. Ebi and J. Henkel, KIT, SS13

3 introduction Why design for low power/energy? Portable Systems  Thermal Considerations  Notebooks, smartphones,  10 o C increase in operating  tablets, cameras, etc. temperature => component 32% of PC market, and growing  failure rate doubles Battery-driven - long battery life  Packaging: ceramic vs . plastic  crucial Cooling requirements  System cost, weight limited by  batteries Increasing levels of  40W, 10 hrs @ 20-35 W- integration / clock  hr/pound = 7-20 pounds (Src: A. Raghunathan, NEC) frequencies make the Slow growth in battery  technology problem worse Must reduce energy drain LOW  10cm 2 , 500 MHz => 315Watts  from batteries POWER Reliability Issues  Environmental Concerns Electro-migration   EPA estimate: 80% of office IR drops on supply lines   equipment electricity is used Inductive effects  in computers Tied to peak/average  “ Energy Star ” program to  power consumption recognize power efficient PCs Power management standard  for desktops and laptops Drive towards “ Green PC ”  http://ces.itec.kit.edu T. Ebi and J. Henkel, KIT, SS13

4 introduction (Src: F. Pollack, Intel http://ces.itec.kit.edu T. Ebi and J. Henkel, KIT, SS13

5 introduction Power consumption: motivation Pentium Crusoe Pentium 4 Crusoe Processor (source: www.transmeta.com) http://ces.itec.kit.edu T. Ebi and J. Henkel, KIT, SS13

6 introduction Energy vs . Programmability  Large (100X Large (100X – – 1000X) gap in energy efficiency between 1000X) gap in energy efficiency between fully programmable and fully custom implementations fully programmable and fully custom implementations  Ample scope for tradeoffs Ample scope for tradeoffs Source: Rabaey et. al ., IEEE Computer, July 2000 http://ces.itec.kit.edu T. Ebi and J. Henkel, KIT, SS13

7 introduction Power consumption by processing type Operations/Watt [MOPS/mW] Ambient Intelligence 10 DSP-ASIPs 1 µPs 0.1 poor design generation 0.01 techniques Technology 1.0µ 0.5µ 0.25µ 0.13µ 0.07µ (Src:[Marw03]) http://ces.itec.kit.edu T. Ebi and J. Henkel, KIT, SS13

8 introduction Relationship between Power and Energy P E t   E P dt Energy: 1 Ws = 1 VAs = 1 Joule = 1 Nm http://ces.itec.kit.edu T. Ebi and J. Henkel, KIT, SS13

9 introduction Power vs. Energy  Minimizing the power consumption is important for  the design of the power supply  the design of voltage regulators  the dimensioning of interconnect  short term cooling  Minimizing the energy consumption:  Limited availability of energy (mobile systems, try to maximize the amount of computation that can be accomplished with a given amount of energy) through:  limited battery capacities (only slowly improving)  very high costs of energy (solar panels, in space)  cooling  high costs  limited space  dependability  long lifetimes, low temperatures (Src:[Marw03]) http://ces.itec.kit.edu T. Ebi and J. Henkel, KIT, SS13

10 introduction HW Power Consumption 1 2 ) ( Power Cap Switching = . _ _Power 2 C V . . . A f L dd + Leakage/Static Power + … High-level synthesis, Power analysis RTL optimizations iteration times Architecture-level Power models Decreasing design iteration times power analysis for macroblocks, seconds - minutes control logic Behavior level Logic synthesis Register-transfer level Power models Logic-level minutes - hours for gates, cells, power analysis nets Transistor-level/ Logic level Layout synthesis hours - days Transistor-level Transistor level power analysis (src: A. Raghunathan, NEC) http://ces.itec.kit.edu T. Ebi and J. Henkel, KIT, SS13

11 introduction Power/Energy-Conscious Applications -Some examples Some examples- http://ces.itec.kit.edu T. Ebi and J. Henkel, KIT, SS13

12 introduction Example 1: E-Textiles - Smart Shirt - Source: [Marc03] http://ces.itec.kit.edu T. Ebi and J. Henkel, KIT, SS13

13 introduction Example 2: Medical Diagnostics (source: Jan Madsen DTU) http://ces.itec.kit.edu T. Ebi and J. Henkel, KIT, SS13

14 introduction Example 3: Sensor Networks Disaster Prevention & Energy-efficient Manufacturing plants & Power distribution Emergency buildings • Improve reliability, operating efficiency Response • $55 B / year opportunity in the US Health care • Unwired operating “ Smart ” environments Traffic control rooms • Homes, Offices, Schools, … • Reduce commute time • Early detection of • Convenience, Productivity, Security by 15 min => $15B/year cardiac attacks in California alone (source: A. Raghunathan, NEC) http://ces.itec.kit.edu T. Ebi and J. Henkel, KIT, SS13

15 introduction More examples  Banking & Money transfer Banking & Money transfer smart cards, … smart cards, …  Consumer Consumer cell phone, MP3 player, PDA, … cell phone, MP3 player, PDA, …  Clothing Clothing electronic textiles electronic textiles  Environment Environment sensor networks sensor networks  Healthcare Healthcare hearings aids, pace maker, … hearings aids, pace maker, …  Telecom Systems Telecom Systems satellite, … satellite, …  … http://ces.itec.kit.edu T. Ebi and J. Henkel, KIT, SS13

16 introduction Power/Energy Sources http://ces.itec.kit.edu T. Ebi and J. Henkel, KIT, SS13

17 introduction Problem of battery capacity in comparison Algorithmic Complexity (src: A. Cuomo, ST Micro, Stockholm, Sept.8, 2004) (Shannon ’ s Law) 10000000 1000000 3G 100000 Processor Performance (Moore ’ s Law) 10000 2G 1000 100 Battery Capacity 10 1G 1 http://ces.itec.kit.edu T. Ebi and J. Henkel, KIT, SS13

18 introduction Primary/Secondary Batteries  Primary batteries Primary batteries  + availability + availability  + no re + no re-charging required charging required  + often higher density compared to secondary batteries (later) + often higher density compared to secondary batteries (later)  - cannot be re cannot be re-charged (replacement of cartridge etc. instead) charged (replacement of cartridge etc. instead)  - - user always needs to carry replacement batteries user always needs to carry replacement batteries  - - form form-factor often unfavorable (not flat as desired) factor often unfavorable (not flat as desired)  Secondary batteries Secondary batteries  Ni Ni-Cd (nickel Cd (nickel-cadmium), NiMH (nickel cadmium), NiMH (nickel-metal metal-hydride, Lithium hydride, Lithium-Ion, Ion, Lithium Lithium-polymer polymer  + can be re + can be re-charged charged  - - lesser energy density compared to primary (it is constantly lesser energy density compared to primary (it is constantly increasing but increasing but “ plateauing plateauing ” i.e. cannot be significantly improved i.e. cannot be significantly improved any more. Lithium-Ion: has increased around 8 any more. Lithium Ion: has increased around 8-10% in the last 10 10% in the last 10 years (every year) years (every year) http://ces.itec.kit.edu T. Ebi and J. Henkel, KIT, SS13

19 introduction Metrics: Energy density: - gravimetric, volumetric -  Gravimetric: Wh/kg Gravimetric: Wh/kg -> Watt * hours / kg > Watt * hours / kg  Volumetric: Wh/l Volumetric: Wh/l -> Watt * hours / liter > Watt * hours / liter 200 500 400 150 300 wh/kg wh/l 100 200 Sanyo Sanyo 50 100 Toshiba Toshiba 0 0 1994 1996 1998 2000 2002 2004 2006 1994 1996 1998 2000 2002 2004 2006 (src: [Blo04]) shown Lithium-Ion technology http://ces.itec.kit.edu T. Ebi and J. Henkel, KIT, SS13

20 introduction Metrics: cost - secondary batteries - Li-ion prism. average price $1,27  Average cost of Average cost of Lithium Lithium-Ion Ion Li-ion cyl. average price $0,45 technology (currently technology (currently 2005) is ~0.5 2005) is ~0.5 NiMH average price $0,55 USD/Wh USD/Wh NiCd average price $0,55 0,5 1 1,5 0 (src: [Blo04]) $ per Wh 10 9 Li-ion (average)  Will decrease further Will decrease further 8 Li-ion Cylindrical Li-ion Prismatic but curve is predicted but curve is predicted 7 US $/cell Li-ion Polym r e 6 to flatten in the near to flatten in the near 5 future future 4 3 2 1 0 1999 2000 2001 2002 2003 2004 2005 2006 http://ces.itec.kit.edu T. Ebi and J. Henkel, KIT, SS13