The University of Adelaide, School of Computer Science 9 April 2019 1.5 Power and Energy Introduction Problem: Get power in, get power out Power and Energy: A systems Perspective How we should think about Performance, Power, and Energy: From the viewpoint of a system designer, there are three primary concerns: What is the maximum power. What is the sustained power consumption, widely called the thermal design power (TDP)  Determines the cooling requirement. The third factor for both (Designers and Users) is Energy and Energy efficiency. Power: is simply Energy per Unit time. 1 watt = 1 Joule/second Energy is the capacity to do work. Energy is power integrated over time. Power is the rate at which work is done, or energy is transmitted. Chapter 2 — Instructions: Language of the Computer

1.5 Power and Energy (continue) Example: I left a 60W light bulb on for 30 days, which raised my electric bill by 43.2 kWh (kilowatt-hours). Example: My car's battery can provide 500 amps at 12 volts, which equals 6kW of power. We use Energy to compare the efficiency of two Processors (not Power). Example: Processor A may have a 20% higher average power consumption than Processor B, but A executes the task in only 70% of the time needed by B, Thus, its Energy consumption =1.2 × 0.7 = 0.84 which is clearly better.

1.5 Power and Energy (continue)  

1.5 Power and Energy (continue) Example: Some microprocessors today are designed to have adjustable voltage, so a 15% reduction in voltage may result in a 15% reduction in frequency. What would be the impact on dynamic energy and on dynamic power?

1.5 Power and Energy (continue) The University of Adelaide, School of Computer Science 9 April 2019 1.5 Power and Energy (continue) Intel 80386 consumed ~ 2 W 3.3 GHz Intel Core i7 consumes 130 W Heat must be dissipated from 1.5 x 1.5 cm chip This is the limit of what can be cooled by air Trends in Power and Energy Chapter 2 — Instructions: Language of the Computer

1.5 Power and Energy (continue) The University of Adelaide, School of Computer Science 9 April 2019 1.5 Power and Energy (continue) Techniques for reducing power: Do nothing well Dynamic Voltage-Frequency Scaling Low power state for DRAM, disks Overclocking, turning off cores Trends in Power and Energy Chapter 2 — Instructions: Language of the Computer

1.5 Power and Energy (continue) The University of Adelaide, School of Computer Science 9 April 2019 1.5 Power and Energy (continue) Static power consumption Currentstatic x Voltage Scales with number of transistors To reduce: power gating Trends in Power and Energy Chapter 2 — Instructions: Language of the Computer

The University of Adelaide, School of Computer Science 9 April 2019 1.6 Trends in Cost Trends in Cost The Impact of Time, Volume, and Commoditization Cost decreases over time Cost driven down by learning curve Learning curve improves Yield Designs that have twice the yield will have half the cost. DRAM: price closely tracks cost Microprocessors: Prices also drop over time, but, because they are less standardized than DRAMs, the relationship between price and cost is more complex. Volume As a rule of thumb, some designers have estimated that cost decreases about 10% for each doubling of volume. Commoditization Many vendors ship virtually identical products  the market is highly competitive. Learning Curve Chapter 2 — Instructions: Language of the Computer

Integrated Circuit Cost The University of Adelaide, School of Computer Science 9 April 2019 Integrated Circuit Cost Trends in Cost   Chapter 2 — Instructions: Language of the Computer

Examples:

The University of Adelaide, School of Computer Science 9 April 2019 1.7 Dependability Dependability   Chapter 2 — Instructions: Language of the Computer

1.8 Measuring Performance The University of Adelaide, School of Computer Science 9 April 2019 1.8 Measuring Performance Typical performance metrics: Response time Throughput Speedup of X relative to Y Execution timeY / Execution timeX Execution time Wall clock time: includes all system overheads CPU time: only computation time Benchmarks Kernels (e.g. matrix multiply) Toy programs (e.g. sorting) Synthetic benchmarks (e.g. Dhrystone) Benchmark suites (e.g. SPEC06fp, TPC-C) Measuring Performance Chapter 2 — Instructions: Language of the Computer

1.9 Principles of Computer Design The University of Adelaide, School of Computer Science 9 April 2019 1.9 Principles of Computer Design Principles Take Advantage of Parallelism e.g. multiple processors, disks, memory banks, pipelining, multiple functional units Principle of Locality Reuse of data and instructions Focus on the Common Case Amdahl’s Law Chapter 2 — Instructions: Language of the Computer

Principles of Computer Design The University of Adelaide, School of Computer Science 9 April 2019 Principles of Computer Design Principles The Processor Performance Equation Chapter 2 — Instructions: Language of the Computer

Principles of Computer Design The University of Adelaide, School of Computer Science 9 April 2019 Principles of Computer Design Principles Different instruction types having different CPIs Chapter 2 — Instructions: Language of the Computer