SlideShare une entreprise Scribd logo
1  sur  29
Presented By Adam Corbin and Michael Hughes
Outline
 Introduction
 Energy-Aware Processors
 OS Energy Management
 Power Reduction through Architectural
  Design
 Reducing Power through Circuitry
 Smart Batteries
 Tools and Packages for Low-Power
  Design
Introduction
   Mobile devices have limited battery
    power
     Remote areas to charge batteries
 Battery life is only expected to grow
  10-20% over the next 10 years
 Software solutions
 Hardware solutions
     Batteries
     Circuitry
Energy-Aware Processors
 Dynamic Voltage/Frequency Scaling
  (DVFS)
 Scheduling Algorithms
 Computational Offloading
     Mobile Wireless Ad-Hoc Networks (MANETs)
Dynamic Voltage/Frequency
Scaling
   One of the more popular power
    management schemes
     Easy implementation
     Performance suffers
     Most effective when hardware is more than
      adequate for the tasks being run
     In resource-constrained mobile computing
      systems with slower processors there may not
      be much room DVFS techniques to save
      energy
Scheduling Algorithms
   Classic computer science problem
     Mobile devices add to the complexity when
     battery energy utilization is considered
   Algorithms are tailored to the specific
    needs of the device
     Energy saving vs. processing performance or a
     combination of the two
Computational Offloading
 Resource-limited computing device defers
  computation to a more capable device
 A solution for Mobile Wireless Ad-Hoc
  Networks (MANETs)
     Distribute computational tasks amongst
      devices
     Improvement in both energy saving and
      processing performance
OS Energy Management
 Dynamic Software Management (DSM)
  Frameworks
 Constrained Power Management (CPM)
Dynamic Software
Management Frameworks
   Goals
     Improve battery utilization
     Avoid competition for limited energy
     resources from multiple applications
   Energy demand prediction
     Critical component of framework
     Uses energy models for prediction of
     application energy demand
Constrained Power
Management (CPM)
   Goals
     Manage both system resources and power of
      an entire platform
     Can optimize energy or performance
   Quality of Service (QoS)
     CPM collects QoS requirements from
     applications and coordinates device drivers to
     support requested QoS
Power Reduction through
Architectural Design
   Caches
     Typically 30-60% of total processor area
     20-50% of processor power consumption
 Filter Caches & On-demand Selective
  Cache
 Speculative Activation
 On-chip Memory
Filter Caches & On-demand
Selective Cache
   Filter Caches
     Reduce dynamic and static power
      consumption
     Shrink the size of L1 and L2 cache
     Significant decrease in performance
   On-demand Selective Cache
     Shuts down parts of the cache according to
      demand
     Suffers from same decrease in peformance
Speculative Activation
   Attempt to make a prediction of where
    the required data may be located
     If prediction is correct, latency and dynamic
      power consumption become similar to direct-
      mapped cache
     If prediction is incorrect, cache is accessed
      twice and results in additional latency
On-chip Memory
   Minimize external data communication
     Eliminates external frame buffers
   No off-chip memory support allows for
    high-quality processing at a minimal
    power level
Reducing Power though Circuitry
 Near-Threshold Computing
 Razor-Thin Margins
 Smaller Transistors
 All-Digital Phase Locking
 Smart Converters
 Next-Generation Dynamic Ram
 SmartReflix power and performance
Near-Threshold Computing
 Lower voltage
 Pros: Energy consumption
 Cons: performance loss, memory and
  logic failures
 Moore’s law
     Energy efficiency problem
   32nm chip example
     Solution to improve performance
Razor-Thin Margins
 Run processors at the threshold voltage
 Reasons why chips run at higher voltages
     Incorporate error detection
   Research from U. Of Michigan
     Arm Cortex –M3
      ○ 60% more energy efficient
Smaller Transistors
   2-D vs. 3-D Transistor
Smaller Transistors Cont.
 37% increase in
  performance
 Moore’s law
 Lower voltage
     50% less power
   Less leakage
All-Digital Phase Locking
   Phase-locked loops
     Lock in and track input signal
     Pace car analogy
 Used for synchronize the processor and
  the clock
 Converting from analog to digital saves
  power
Smart Converters
 The analog to digital converts use a
  switched capacitor
 Oregon State University
     Loop of inverters
   Chiao Tung University in Taiwan
     Separate signal in 2 stages
      ○ 1 simple processing
      ○ 2 fine tuning
Next Generation of Dynamic Ram
   Samsung and Hynix on DDR4 ram
     Drop voltage from 1.5V to 1.2V
 Includes better clocking
 Faster algorithms for encoding data
 End result
     Consume less energy
     Higher bandwidth
Smart Batteries
   Background Info
     Memory effect
      ○ Solution: lithium ion
   Advanced Power Management(APM)
     Mobile device uses information
   Polling for suspicious activity
     Avoid stalling of devices and wasted energy
SmartReflix power and performance
   Static Techniques
     Power-gating


   Dynamic Techniques
     Dynamic power switching
     Adaptive voltage scaling
     Dynamic voltage/frequency scaling
SimpleScalar (Tools)
 Tool set
 Simulator to compare two different
  machines
 Extended to the Power Analyzer
     For the ARM platform
     Exam the performance/power trade-offs
   Extended to the Wattch Project
     Web app
PowerScope (Tools)
   Step 1. Gathers information based on
    processes
PowerScope Cont
   Step 2. Of- Line Analysis
PowerScope Cont.
   Output
Conclusions

Contenu connexe

Tendances

Intelligent Power Allocation
Intelligent Power AllocationIntelligent Power Allocation
Intelligent Power AllocationChiou-Nan Chen
 
What is a data center
What is a data centerWhat is a data center
What is a data centerLivin Jose
 
Genesys System - 8dec2010
Genesys System - 8dec2010Genesys System - 8dec2010
Genesys System - 8dec2010Agora Group
 
A survey on dynamic energy management at virtualization level in cloud data c...
A survey on dynamic energy management at virtualization level in cloud data c...A survey on dynamic energy management at virtualization level in cloud data c...
A survey on dynamic energy management at virtualization level in cloud data c...csandit
 
Making the grid more efficient, flexible and secure
Making the grid more efficient, flexible and secureMaking the grid more efficient, flexible and secure
Making the grid more efficient, flexible and secureSchneider Electric
 
What is data center availability modes slide
What is data center availability modes slideWhat is data center availability modes slide
What is data center availability modes slideLivin Jose
 
Implementing energy efficient data centers
Implementing energy efficient  data centersImplementing energy efficient  data centers
Implementing energy efficient data centersSchneider Electric India
 
Advanced Weather Systems for Mining Operations
Advanced Weather Systems for Mining OperationsAdvanced Weather Systems for Mining Operations
Advanced Weather Systems for Mining OperationsSchneider Electric
 
How Data Center Infrastructure Management Software Improves Planning and Cuts...
How Data Center Infrastructure Management Software Improves Planning and Cuts...How Data Center Infrastructure Management Software Improves Planning and Cuts...
How Data Center Infrastructure Management Software Improves Planning and Cuts...Schneider Electric
 
NetSure ITM 48V DC UPS
NetSure ITM 48V DC UPSNetSure ITM 48V DC UPS
NetSure ITM 48V DC UPSmmurrill
 
Power Protection for Digital Medical Imaging and Diagnostic Equipment
Power Protection for Digital Medical Imaging and Diagnostic EquipmentPower Protection for Digital Medical Imaging and Diagnostic Equipment
Power Protection for Digital Medical Imaging and Diagnostic EquipmentSchneider Electric
 
Underground Self-Healing Solution
Underground Self-Healing Solution Underground Self-Healing Solution
Underground Self-Healing Solution Schneider Electric
 
A Survey on Low Power VLSI Designs
A Survey on Low Power VLSI Designs A Survey on Low Power VLSI Designs
A Survey on Low Power VLSI Designs IJEEE
 
Smart grid technologies across the globe
Smart grid technologies across the globeSmart grid technologies across the globe
Smart grid technologies across the globeSchneider Electric
 

Tendances (19)

Intelligent Power Allocation
Intelligent Power AllocationIntelligent Power Allocation
Intelligent Power Allocation
 
What is a data center
What is a data centerWhat is a data center
What is a data center
 
Types of Electrical Meters in Data Centers
Types of Electrical Meters in Data CentersTypes of Electrical Meters in Data Centers
Types of Electrical Meters in Data Centers
 
Genesys System - 8dec2010
Genesys System - 8dec2010Genesys System - 8dec2010
Genesys System - 8dec2010
 
A survey on dynamic energy management at virtualization level in cloud data c...
A survey on dynamic energy management at virtualization level in cloud data c...A survey on dynamic energy management at virtualization level in cloud data c...
A survey on dynamic energy management at virtualization level in cloud data c...
 
Aqeel
AqeelAqeel
Aqeel
 
Making the grid more efficient, flexible and secure
Making the grid more efficient, flexible and secureMaking the grid more efficient, flexible and secure
Making the grid more efficient, flexible and secure
 
The Dreamwatts Smart Meter
The Dreamwatts Smart MeterThe Dreamwatts Smart Meter
The Dreamwatts Smart Meter
 
What is data center availability modes slide
What is data center availability modes slideWhat is data center availability modes slide
What is data center availability modes slide
 
Implementing energy efficient data centers
Implementing energy efficient  data centersImplementing energy efficient  data centers
Implementing energy efficient data centers
 
Advanced Weather Systems for Mining Operations
Advanced Weather Systems for Mining OperationsAdvanced Weather Systems for Mining Operations
Advanced Weather Systems for Mining Operations
 
How Data Center Infrastructure Management Software Improves Planning and Cuts...
How Data Center Infrastructure Management Software Improves Planning and Cuts...How Data Center Infrastructure Management Software Improves Planning and Cuts...
How Data Center Infrastructure Management Software Improves Planning and Cuts...
 
Indonesia JCM
Indonesia JCM Indonesia JCM
Indonesia JCM
 
NetSure ITM 48V DC UPS
NetSure ITM 48V DC UPSNetSure ITM 48V DC UPS
NetSure ITM 48V DC UPS
 
Power Protection for Digital Medical Imaging and Diagnostic Equipment
Power Protection for Digital Medical Imaging and Diagnostic EquipmentPower Protection for Digital Medical Imaging and Diagnostic Equipment
Power Protection for Digital Medical Imaging and Diagnostic Equipment
 
Underground Self-Healing Solution
Underground Self-Healing Solution Underground Self-Healing Solution
Underground Self-Healing Solution
 
DCD_FOCUS on POWER Article
DCD_FOCUS on POWER ArticleDCD_FOCUS on POWER Article
DCD_FOCUS on POWER Article
 
A Survey on Low Power VLSI Designs
A Survey on Low Power VLSI Designs A Survey on Low Power VLSI Designs
A Survey on Low Power VLSI Designs
 
Smart grid technologies across the globe
Smart grid technologies across the globeSmart grid technologies across the globe
Smart grid technologies across the globe
 

Similaire à Mobile computing edited

3-Anandi.ppt
3-Anandi.ppt3-Anandi.ppt
3-Anandi.pptECEHoD16
 
LPflow_updated.ppt
LPflow_updated.pptLPflow_updated.ppt
LPflow_updated.pptssuser36861c
 
MRI Energy-Efficient Cloud Computing
MRI Energy-Efficient Cloud ComputingMRI Energy-Efficient Cloud Computing
MRI Energy-Efficient Cloud ComputingRoger Rafanell Mas
 
A verilog based simulation methodology for estimating statistical test for th...
A verilog based simulation methodology for estimating statistical test for th...A verilog based simulation methodology for estimating statistical test for th...
A verilog based simulation methodology for estimating statistical test for th...ijsrd.com
 
A Brief Survey of Current Power Limiting Strategies
A Brief Survey of Current Power Limiting StrategiesA Brief Survey of Current Power Limiting Strategies
A Brief Survey of Current Power Limiting StrategiesIRJET Journal
 
Computer Architecture and Organization
Computer Architecture and OrganizationComputer Architecture and Organization
Computer Architecture and Organizationssuserdfc773
 
FPGA Based Data Processing for Real-time WSN Applications:
FPGA Based Data Processing for Real-time WSN Applications: FPGA Based Data Processing for Real-time WSN Applications:
FPGA Based Data Processing for Real-time WSN Applications: Ilham Amezzane
 
How to achieve 95%+ Accurate power measurement during architecture exploration?
How to achieve 95%+ Accurate power measurement during architecture exploration? How to achieve 95%+ Accurate power measurement during architecture exploration?
How to achieve 95%+ Accurate power measurement during architecture exploration? Deepak Shankar
 
Adiabatic technique based low power synchronous counter design
Adiabatic technique based low power synchronous counter  designAdiabatic technique based low power synchronous counter  design
Adiabatic technique based low power synchronous counter designIJECEIAES
 
Automatic power factor correction
Automatic power factor correction Automatic power factor correction
Automatic power factor correction VIKAS KUMAR MANJHI
 

Similaire à Mobile computing edited (20)

Ip so c-30sept2010
Ip so c-30sept2010Ip so c-30sept2010
Ip so c-30sept2010
 
3-Anandi.ppt
3-Anandi.ppt3-Anandi.ppt
3-Anandi.ppt
 
LPVLSI.ppt
LPVLSI.pptLPVLSI.ppt
LPVLSI.ppt
 
Low power methods.ppt
Low power methods.pptLow power methods.ppt
Low power methods.ppt
 
Anandi.ppt
Anandi.pptAnandi.ppt
Anandi.ppt
 
LPflow_updated.ppt
LPflow_updated.pptLPflow_updated.ppt
LPflow_updated.ppt
 
MRI Energy-Efficient Cloud Computing
MRI Energy-Efficient Cloud ComputingMRI Energy-Efficient Cloud Computing
MRI Energy-Efficient Cloud Computing
 
Low power vlsi design
Low power vlsi designLow power vlsi design
Low power vlsi design
 
Wireless Sensor Network
Wireless Sensor NetworkWireless Sensor Network
Wireless Sensor Network
 
Reduce system energy by smart CPU management
Reduce system energy by smart CPU managementReduce system energy by smart CPU management
Reduce system energy by smart CPU management
 
A verilog based simulation methodology for estimating statistical test for th...
A verilog based simulation methodology for estimating statistical test for th...A verilog based simulation methodology for estimating statistical test for th...
A verilog based simulation methodology for estimating statistical test for th...
 
A Brief Survey of Current Power Limiting Strategies
A Brief Survey of Current Power Limiting StrategiesA Brief Survey of Current Power Limiting Strategies
A Brief Survey of Current Power Limiting Strategies
 
Computer Architecture and Organization
Computer Architecture and OrganizationComputer Architecture and Organization
Computer Architecture and Organization
 
FPGA Based Data Processing for Real-time WSN Applications:
FPGA Based Data Processing for Real-time WSN Applications: FPGA Based Data Processing for Real-time WSN Applications:
FPGA Based Data Processing for Real-time WSN Applications:
 
How to achieve 95%+ Accurate power measurement during architecture exploration?
How to achieve 95%+ Accurate power measurement during architecture exploration? How to achieve 95%+ Accurate power measurement during architecture exploration?
How to achieve 95%+ Accurate power measurement during architecture exploration?
 
Low power vlsi design
Low power vlsi designLow power vlsi design
Low power vlsi design
 
Adiabatic technique based low power synchronous counter design
Adiabatic technique based low power synchronous counter  designAdiabatic technique based low power synchronous counter  design
Adiabatic technique based low power synchronous counter design
 
Embedded system
Embedded systemEmbedded system
Embedded system
 
Embeddedsystem
EmbeddedsystemEmbeddedsystem
Embeddedsystem
 
Automatic power factor correction
Automatic power factor correction Automatic power factor correction
Automatic power factor correction
 

Dernier

Enterprise Knowledge Graphs - Data Summit 2024
Enterprise Knowledge Graphs - Data Summit 2024Enterprise Knowledge Graphs - Data Summit 2024
Enterprise Knowledge Graphs - Data Summit 2024Enterprise Knowledge
 
Salesforce Adoption – Metrics, Methods, and Motivation, Antone Kom
Salesforce Adoption – Metrics, Methods, and Motivation, Antone KomSalesforce Adoption – Metrics, Methods, and Motivation, Antone Kom
Salesforce Adoption – Metrics, Methods, and Motivation, Antone KomCzechDreamin
 
AI presentation and introduction - Retrieval Augmented Generation RAG 101
AI presentation and introduction - Retrieval Augmented Generation RAG 101AI presentation and introduction - Retrieval Augmented Generation RAG 101
AI presentation and introduction - Retrieval Augmented Generation RAG 101vincent683379
 
Secure Zero Touch enabled Edge compute with Dell NativeEdge via FDO _ Brad at...
Secure Zero Touch enabled Edge compute with Dell NativeEdge via FDO _ Brad at...Secure Zero Touch enabled Edge compute with Dell NativeEdge via FDO _ Brad at...
Secure Zero Touch enabled Edge compute with Dell NativeEdge via FDO _ Brad at...FIDO Alliance
 
IESVE for Early Stage Design and Planning
IESVE for Early Stage Design and PlanningIESVE for Early Stage Design and Planning
IESVE for Early Stage Design and PlanningIES VE
 
Powerful Start- the Key to Project Success, Barbara Laskowska
Powerful Start- the Key to Project Success, Barbara LaskowskaPowerful Start- the Key to Project Success, Barbara Laskowska
Powerful Start- the Key to Project Success, Barbara LaskowskaCzechDreamin
 
Designing for Hardware Accessibility at Comcast
Designing for Hardware Accessibility at ComcastDesigning for Hardware Accessibility at Comcast
Designing for Hardware Accessibility at ComcastUXDXConf
 
Speed Wins: From Kafka to APIs in Minutes
Speed Wins: From Kafka to APIs in MinutesSpeed Wins: From Kafka to APIs in Minutes
Speed Wins: From Kafka to APIs in Minutesconfluent
 
Oauth 2.0 Introduction and Flows with MuleSoft
Oauth 2.0 Introduction and Flows with MuleSoftOauth 2.0 Introduction and Flows with MuleSoft
Oauth 2.0 Introduction and Flows with MuleSoftshyamraj55
 
ASRock Industrial FDO Solutions in Action for Industrial Edge AI _ Kenny at A...
ASRock Industrial FDO Solutions in Action for Industrial Edge AI _ Kenny at A...ASRock Industrial FDO Solutions in Action for Industrial Edge AI _ Kenny at A...
ASRock Industrial FDO Solutions in Action for Industrial Edge AI _ Kenny at A...FIDO Alliance
 
Intro in Product Management - Коротко про професію продакт менеджера
Intro in Product Management - Коротко про професію продакт менеджераIntro in Product Management - Коротко про професію продакт менеджера
Intro in Product Management - Коротко про професію продакт менеджераMark Opanasiuk
 
Linux Foundation Edge _ Overview of FDO Software Components _ Randy at Intel.pdf
Linux Foundation Edge _ Overview of FDO Software Components _ Randy at Intel.pdfLinux Foundation Edge _ Overview of FDO Software Components _ Randy at Intel.pdf
Linux Foundation Edge _ Overview of FDO Software Components _ Randy at Intel.pdfFIDO Alliance
 
ECS 2024 Teams Premium - Pretty Secure
ECS 2024   Teams Premium - Pretty SecureECS 2024   Teams Premium - Pretty Secure
ECS 2024 Teams Premium - Pretty SecureFemke de Vroome
 
AI revolution and Salesforce, Jiří Karpíšek
AI revolution and Salesforce, Jiří KarpíšekAI revolution and Salesforce, Jiří Karpíšek
AI revolution and Salesforce, Jiří KarpíšekCzechDreamin
 
Measures in SQL (a talk at SF Distributed Systems meetup, 2024-05-22)
Measures in SQL (a talk at SF Distributed Systems meetup, 2024-05-22)Measures in SQL (a talk at SF Distributed Systems meetup, 2024-05-22)
Measures in SQL (a talk at SF Distributed Systems meetup, 2024-05-22)Julian Hyde
 
The Value of Certifying Products for FDO _ Paul at FIDO Alliance.pdf
The Value of Certifying Products for FDO _ Paul at FIDO Alliance.pdfThe Value of Certifying Products for FDO _ Paul at FIDO Alliance.pdf
The Value of Certifying Products for FDO _ Paul at FIDO Alliance.pdfFIDO Alliance
 
How we scaled to 80K users by doing nothing!.pdf
How we scaled to 80K users by doing nothing!.pdfHow we scaled to 80K users by doing nothing!.pdf
How we scaled to 80K users by doing nothing!.pdfSrushith Repakula
 
Strategic AI Integration in Engineering Teams
Strategic AI Integration in Engineering TeamsStrategic AI Integration in Engineering Teams
Strategic AI Integration in Engineering TeamsUXDXConf
 
Connecting the Dots in Product Design at KAYAK
Connecting the Dots in Product Design at KAYAKConnecting the Dots in Product Design at KAYAK
Connecting the Dots in Product Design at KAYAKUXDXConf
 
Free and Effective: Making Flows Publicly Accessible, Yumi Ibrahimzade
Free and Effective: Making Flows Publicly Accessible, Yumi IbrahimzadeFree and Effective: Making Flows Publicly Accessible, Yumi Ibrahimzade
Free and Effective: Making Flows Publicly Accessible, Yumi IbrahimzadeCzechDreamin
 

Dernier (20)

Enterprise Knowledge Graphs - Data Summit 2024
Enterprise Knowledge Graphs - Data Summit 2024Enterprise Knowledge Graphs - Data Summit 2024
Enterprise Knowledge Graphs - Data Summit 2024
 
Salesforce Adoption – Metrics, Methods, and Motivation, Antone Kom
Salesforce Adoption – Metrics, Methods, and Motivation, Antone KomSalesforce Adoption – Metrics, Methods, and Motivation, Antone Kom
Salesforce Adoption – Metrics, Methods, and Motivation, Antone Kom
 
AI presentation and introduction - Retrieval Augmented Generation RAG 101
AI presentation and introduction - Retrieval Augmented Generation RAG 101AI presentation and introduction - Retrieval Augmented Generation RAG 101
AI presentation and introduction - Retrieval Augmented Generation RAG 101
 
Secure Zero Touch enabled Edge compute with Dell NativeEdge via FDO _ Brad at...
Secure Zero Touch enabled Edge compute with Dell NativeEdge via FDO _ Brad at...Secure Zero Touch enabled Edge compute with Dell NativeEdge via FDO _ Brad at...
Secure Zero Touch enabled Edge compute with Dell NativeEdge via FDO _ Brad at...
 
IESVE for Early Stage Design and Planning
IESVE for Early Stage Design and PlanningIESVE for Early Stage Design and Planning
IESVE for Early Stage Design and Planning
 
Powerful Start- the Key to Project Success, Barbara Laskowska
Powerful Start- the Key to Project Success, Barbara LaskowskaPowerful Start- the Key to Project Success, Barbara Laskowska
Powerful Start- the Key to Project Success, Barbara Laskowska
 
Designing for Hardware Accessibility at Comcast
Designing for Hardware Accessibility at ComcastDesigning for Hardware Accessibility at Comcast
Designing for Hardware Accessibility at Comcast
 
Speed Wins: From Kafka to APIs in Minutes
Speed Wins: From Kafka to APIs in MinutesSpeed Wins: From Kafka to APIs in Minutes
Speed Wins: From Kafka to APIs in Minutes
 
Oauth 2.0 Introduction and Flows with MuleSoft
Oauth 2.0 Introduction and Flows with MuleSoftOauth 2.0 Introduction and Flows with MuleSoft
Oauth 2.0 Introduction and Flows with MuleSoft
 
ASRock Industrial FDO Solutions in Action for Industrial Edge AI _ Kenny at A...
ASRock Industrial FDO Solutions in Action for Industrial Edge AI _ Kenny at A...ASRock Industrial FDO Solutions in Action for Industrial Edge AI _ Kenny at A...
ASRock Industrial FDO Solutions in Action for Industrial Edge AI _ Kenny at A...
 
Intro in Product Management - Коротко про професію продакт менеджера
Intro in Product Management - Коротко про професію продакт менеджераIntro in Product Management - Коротко про професію продакт менеджера
Intro in Product Management - Коротко про професію продакт менеджера
 
Linux Foundation Edge _ Overview of FDO Software Components _ Randy at Intel.pdf
Linux Foundation Edge _ Overview of FDO Software Components _ Randy at Intel.pdfLinux Foundation Edge _ Overview of FDO Software Components _ Randy at Intel.pdf
Linux Foundation Edge _ Overview of FDO Software Components _ Randy at Intel.pdf
 
ECS 2024 Teams Premium - Pretty Secure
ECS 2024   Teams Premium - Pretty SecureECS 2024   Teams Premium - Pretty Secure
ECS 2024 Teams Premium - Pretty Secure
 
AI revolution and Salesforce, Jiří Karpíšek
AI revolution and Salesforce, Jiří KarpíšekAI revolution and Salesforce, Jiří Karpíšek
AI revolution and Salesforce, Jiří Karpíšek
 
Measures in SQL (a talk at SF Distributed Systems meetup, 2024-05-22)
Measures in SQL (a talk at SF Distributed Systems meetup, 2024-05-22)Measures in SQL (a talk at SF Distributed Systems meetup, 2024-05-22)
Measures in SQL (a talk at SF Distributed Systems meetup, 2024-05-22)
 
The Value of Certifying Products for FDO _ Paul at FIDO Alliance.pdf
The Value of Certifying Products for FDO _ Paul at FIDO Alliance.pdfThe Value of Certifying Products for FDO _ Paul at FIDO Alliance.pdf
The Value of Certifying Products for FDO _ Paul at FIDO Alliance.pdf
 
How we scaled to 80K users by doing nothing!.pdf
How we scaled to 80K users by doing nothing!.pdfHow we scaled to 80K users by doing nothing!.pdf
How we scaled to 80K users by doing nothing!.pdf
 
Strategic AI Integration in Engineering Teams
Strategic AI Integration in Engineering TeamsStrategic AI Integration in Engineering Teams
Strategic AI Integration in Engineering Teams
 
Connecting the Dots in Product Design at KAYAK
Connecting the Dots in Product Design at KAYAKConnecting the Dots in Product Design at KAYAK
Connecting the Dots in Product Design at KAYAK
 
Free and Effective: Making Flows Publicly Accessible, Yumi Ibrahimzade
Free and Effective: Making Flows Publicly Accessible, Yumi IbrahimzadeFree and Effective: Making Flows Publicly Accessible, Yumi Ibrahimzade
Free and Effective: Making Flows Publicly Accessible, Yumi Ibrahimzade
 

Mobile computing edited

  • 1. Presented By Adam Corbin and Michael Hughes
  • 2. Outline  Introduction  Energy-Aware Processors  OS Energy Management  Power Reduction through Architectural Design  Reducing Power through Circuitry  Smart Batteries  Tools and Packages for Low-Power Design
  • 3. Introduction  Mobile devices have limited battery power  Remote areas to charge batteries  Battery life is only expected to grow 10-20% over the next 10 years  Software solutions  Hardware solutions  Batteries  Circuitry
  • 4. Energy-Aware Processors  Dynamic Voltage/Frequency Scaling (DVFS)  Scheduling Algorithms  Computational Offloading  Mobile Wireless Ad-Hoc Networks (MANETs)
  • 5. Dynamic Voltage/Frequency Scaling  One of the more popular power management schemes  Easy implementation  Performance suffers  Most effective when hardware is more than adequate for the tasks being run  In resource-constrained mobile computing systems with slower processors there may not be much room DVFS techniques to save energy
  • 6. Scheduling Algorithms  Classic computer science problem  Mobile devices add to the complexity when battery energy utilization is considered  Algorithms are tailored to the specific needs of the device  Energy saving vs. processing performance or a combination of the two
  • 7. Computational Offloading  Resource-limited computing device defers computation to a more capable device  A solution for Mobile Wireless Ad-Hoc Networks (MANETs)  Distribute computational tasks amongst devices  Improvement in both energy saving and processing performance
  • 8. OS Energy Management  Dynamic Software Management (DSM) Frameworks  Constrained Power Management (CPM)
  • 9. Dynamic Software Management Frameworks  Goals  Improve battery utilization  Avoid competition for limited energy resources from multiple applications  Energy demand prediction  Critical component of framework  Uses energy models for prediction of application energy demand
  • 10. Constrained Power Management (CPM)  Goals  Manage both system resources and power of an entire platform  Can optimize energy or performance  Quality of Service (QoS)  CPM collects QoS requirements from applications and coordinates device drivers to support requested QoS
  • 11. Power Reduction through Architectural Design  Caches  Typically 30-60% of total processor area  20-50% of processor power consumption  Filter Caches & On-demand Selective Cache  Speculative Activation  On-chip Memory
  • 12. Filter Caches & On-demand Selective Cache  Filter Caches  Reduce dynamic and static power consumption  Shrink the size of L1 and L2 cache  Significant decrease in performance  On-demand Selective Cache  Shuts down parts of the cache according to demand  Suffers from same decrease in peformance
  • 13. Speculative Activation  Attempt to make a prediction of where the required data may be located  If prediction is correct, latency and dynamic power consumption become similar to direct- mapped cache  If prediction is incorrect, cache is accessed twice and results in additional latency
  • 14. On-chip Memory  Minimize external data communication  Eliminates external frame buffers  No off-chip memory support allows for high-quality processing at a minimal power level
  • 15. Reducing Power though Circuitry  Near-Threshold Computing  Razor-Thin Margins  Smaller Transistors  All-Digital Phase Locking  Smart Converters  Next-Generation Dynamic Ram  SmartReflix power and performance
  • 16. Near-Threshold Computing  Lower voltage  Pros: Energy consumption  Cons: performance loss, memory and logic failures  Moore’s law  Energy efficiency problem  32nm chip example  Solution to improve performance
  • 17. Razor-Thin Margins  Run processors at the threshold voltage  Reasons why chips run at higher voltages  Incorporate error detection  Research from U. Of Michigan  Arm Cortex –M3 ○ 60% more energy efficient
  • 18. Smaller Transistors  2-D vs. 3-D Transistor
  • 19. Smaller Transistors Cont.  37% increase in performance  Moore’s law  Lower voltage  50% less power  Less leakage
  • 20. All-Digital Phase Locking  Phase-locked loops  Lock in and track input signal  Pace car analogy  Used for synchronize the processor and the clock  Converting from analog to digital saves power
  • 21. Smart Converters  The analog to digital converts use a switched capacitor  Oregon State University  Loop of inverters  Chiao Tung University in Taiwan  Separate signal in 2 stages ○ 1 simple processing ○ 2 fine tuning
  • 22. Next Generation of Dynamic Ram  Samsung and Hynix on DDR4 ram  Drop voltage from 1.5V to 1.2V  Includes better clocking  Faster algorithms for encoding data  End result  Consume less energy  Higher bandwidth
  • 23. Smart Batteries  Background Info  Memory effect ○ Solution: lithium ion  Advanced Power Management(APM)  Mobile device uses information  Polling for suspicious activity  Avoid stalling of devices and wasted energy
  • 24. SmartReflix power and performance  Static Techniques  Power-gating  Dynamic Techniques  Dynamic power switching  Adaptive voltage scaling  Dynamic voltage/frequency scaling
  • 25. SimpleScalar (Tools)  Tool set  Simulator to compare two different machines  Extended to the Power Analyzer  For the ARM platform  Exam the performance/power trade-offs  Extended to the Wattch Project  Web app
  • 26. PowerScope (Tools)  Step 1. Gathers information based on processes
  • 27. PowerScope Cont  Step 2. Of- Line Analysis

Notes de l'éditeur

  1. By adding parallel processing this could help in preformance
  2. The one side vs 3 side Tri gate transistor
  3. http://www.treehugger.com/gadgets/intel-announces-revolutionary-3d-transistors-50-more-energy-efficient-than-previous-generation.html - ideal for use in small handheld devices, which operate using less energy to "switch" back and forth between states. -200 mV lower voltage compared to the -the lower the voltage the higher the leakage but in this design there is less leakage at lower voltages compared to the 2d http://news.cnet.com/8301-13924_3-20059431-64.html http://www.anandtech.com/show/4333/intels-silvermont-a-new-atom-architecture
  4. Pace car analagy- http://en.wikipedia.org/wiki/Phase-locked_loop The All-digital phase locking are easier to make Paper- http://web.mit.edu/~bdaya/www/All%20Digital%20Phase%20Locked%20Loop%20Design%20and%20Implementation.pdf
  5. Power gating: “The basic strategy of power gating is to provide two power modes: a low power mode and an active mode. The goal is to switch between these modes at the appropriate time and in the appropriate manner to maximize power savings while minimizing the impact to performance.” http://nanocad.ee.ucla.edu/pub/Main/SnippetTutorial/PG.pdf Dynamic power switching- basically the same as power gating as it determines when a device has completed its current computational tasks and, if it's not needed at the moment, then puts the device into a low power state  Adaptive voltage scaling-by reducing the voltage you can gain up to 60% less energy consumption but you need to understand the performance at each voltage level. Once this information is gained then the voltage and be scaled based on what the device is doing http://www.ti.com/lit/ml/slyb186/slyb186.pdf Dynamic Vol/Freq Scaling- Similar to the past examples they will be scaled based on the activity of the device
  6. http://www.ecs.umass.edu/ece/koren/architecture/Simplescalar/SimpleScalar_introduction.htm Has high-level function and low –level detailed simulator   users can build modeling applications that simulate real programs running on a range of modern processors and systems. Power Analyzer- University of Michigan and University of Colorado Wattch project- Princeton University- would connect to the system and estimate power that the system had consumed. Also has multi meter results from the CPU
  7. http://notrump.eecs.umich.edu/papers/pscope99.pdf http://web.cs.wpi.edu/~emmanuel/courses/cs525m/S06/slides/powerscope_wk4.pdf The data collection phase uses a multi-meter [34] connected by a Hewlett-Packard interface bus (HPIB) connection to a data collection computer running an energy monitor component. A profiling computer running an application and a system monitor component are getting power by being connected to the multi-meter for measurement. The profiling computer also provides the trigger for the multi-meter to start measuring.
  8. The offline analysis phase uses a third component called the energy analyzer, which will take the results produced in the data collection phase by the system monitor and the energy monitor and correlate the two to produce the energy profile of the application.
  9. With this information the user can focus on the process that are consuming the most amount of power. In some cases there were 50% power reduction with minimal loss of performance