CS267/E233 Applications of Parallel Computers Lecture 1: Introduction

  • Published on
    14-Jan-2016

  • View
    51

  • Download
    7

Embed Size (px)

DESCRIPTION

CS267/E233 Applications of Parallel Computers Lecture 1: Introduction. James Demmel demmel@cs.berkeley.edu www.cs.berkeley.edu/~demmel/cs267_Spr05. Outline. Introduction Large important problems require powerful computers Why powerful computers must be parallel processors - PowerPoint PPT Presentation

Transcript

  • CS267/E233Applications of Parallel Computers

    Lecture 1: Introduction

    James Demmeldemmel@cs.berkeley.eduwww.cs.berkeley.edu/~demmel/cs267_Spr05

    CS267 Lecture 1: Intro

  • OutlineIntroductionLarge important problems require powerful computers Why powerful computers must be parallel processors Why writing (fast) parallel programs is hardPrinciples of parallel computing performanceStructure of the course

    CS267 Lecture 1: Intro

  • Why we need powerful computers

    CS267 Lecture 1: Intro

  • Units of Measure in HPCHigh Performance Computing (HPC) units are:Flops: floating point operationsFlops/s: floating point operations per secondBytes: size of data (a double precision floating point number is 8)Typical sizes are millions, billions, trillionsMegaMflop/s = 106 flop/secMbyte = 220 = 1048576 ~ 106 bytesGigaGflop/s = 109 flop/secGbyte = 230 ~ 109 bytesTeraTflop/s = 1012 flop/secTbyte = 240 ~ 1012 bytes PetaPflop/s = 1015 flop/secPbyte = 250 ~ 1015 bytesExaEflop/s = 1018 flop/secEbyte = 260 ~ 1018 bytesZettaZflop/s = 1021 flop/secZbyte = 270 ~ 1021 bytesYottaYflop/s = 1024 flop/secYbyte = 280 ~ 1024 bytes

    CS267 Lecture 1: Intro

  • Simulation: The Third Pillar of Science Traditional scientific and engineering paradigm:Do theory or paper design.Perform experiments or build system.Limitations:Too difficult -- build large wind tunnels.Too expensive -- build a throw-away passenger jet.Too slow -- wait for climate or galactic evolution.Too dangerous -- weapons, drug design, climate experimentation.Computational science paradigm:Use high performance computer systems to simulate the phenomenonBase on known physical laws and efficient numerical methods.

    CS267 Lecture 1: Intro

  • Some Particularly Challenging ComputationsScienceGlobal climate modelingBiology: genomics; protein folding; drug designAstrophysical modelingComputational ChemistryComputational Material Sciences and NanosciencesEngineeringSemiconductor designEarthquake and structural modelingComputation fluid dynamics (airplane design)Combustion (engine design)Crash simulationBusinessFinancial and economic modelingTransaction processing, web services and search enginesDefenseNuclear weapons -- test by simulationsCryptography

    CS267 Lecture 1: Intro

  • Economic Impact of HPCAirlines:System-wide logistics optimization systems on parallel systems.Savings: approx. $100 million per airline per year.Automotive design:Major automotive companies use large systems (500+ CPUs) for:CAD-CAM, crash testing, structural integrity and aerodynamics.One company has 500+ CPU parallel system.Savings: approx. $1 billion per company per year.Semiconductor industry:Semiconductor firms use large systems (500+ CPUs) fordevice electronics simulation and logic validation Savings: approx. $1 billion per company per year.Securities industry:Savings: approx. $15 billion per year for U.S. home mortgages.

    CS267 Lecture 1: Intro

  • $5B World Market in Technical ComputingSource: IDC 2004, from NRC Future of Supercomputer Report

    CS267 Lecture 1: Intro

  • Global Climate Modeling ProblemProblem is to compute:f(latitude, longitude, elevation, time) temperature, pressure, humidity, wind velocity Approach:Discretize the domain, e.g., a measurement point every 10 kmDevise an algorithm to predict weather at time t+dt given tUses:Predict major events, e.g., El NinoUse in setting air emissions standards

    Source: http://www.epm.ornl.gov/chammp/chammp.html

    CS267 Lecture 1: Intro

  • Global Climate Modeling ComputationOne piece is modeling the fluid flow in the atmosphereSolve Navier-Stokes equationsRoughly 100 Flops per grid point with 1 minute timestepComputational requirements:To match real-time, need 5 x 1011 flops in 60 seconds = 8 Gflop/sWeather prediction (7 days in 24 hours) 56 Gflop/sClimate prediction (50 years in 30 days) 4.8 Tflop/sTo use in policy negotiations (50 years in 12 hours) 288 Tflop/sTo double the grid resolution, computation is 8x to 16x State of the art models require integration of atmosphere, ocean, sea-ice, land models, plus possibly carbon cycle, geochemistry and moreCurrent models are coarser than this

    CS267 Lecture 1: Intro

  • High Resolution Climate Modeling on NERSC-3 P. Duffy, et al., LLNL

    CS267 Lecture 1: Intro

  • A 1000 Year Climate SimulationWarren Washington and Jerry Meehl, National Center for Atmospheric Research; Bert Semtner, Naval Postgraduate School; John Weatherly, U.S. Army Cold Regions Research and Engineering Lab Laboratory et al. http://www.nersc.gov/news/science/bigsplash2002.pdfDemonstration of the Community Climate Model (CCSM2)A 1000-year simulation shows long-term, stable representation of the earths climate. 760,000 processor hours usedTemperature change shown

    CS267 Lecture 1: Intro

  • Climate Modeling on the Earth Simulator SystemDevelopment of ES started in 1997 in order to make a comprehensive understanding of global environmental changes such as global warming.26.58Tflops was obtained by a global atmospheric circulation code.35.86Tflops (87.5% of the peak performance) is achieved in the Linpack benchmark.Its construction was completed at the end of February, 2002 and the practical operation started from March 1, 2002

    CS267 Lecture 1: Intro

  • Astrophysics: Binary Black Hole DynamicsMassive supernova cores collapse to black holes. At black hole center spacetime breaks down. Critical test of theories of gravity General Relativity to Quantum Gravity. Indirect observation most galaxies have a black hole at their center.Gravity waves show black hole directly including detailed parameters.Binary black holes most powerful sources of gravity waves. Simulation extraordinarily complex evolution disrupts the spacetime !

    CS267 Lecture 1: Intro

  • CS267 Lecture 1: Intro

  • Heart SimulationProblem is to compute blood flow in the heartApproach:Modeled as an elastic structure in an incompressible fluid.The immersed boundary method due to Peskin and McQueen.20 years of development in modelMany applications other than the heart: blood clotting, inner ear, paper making, embryo growth, and othersUse a regularly spaced mesh (set of points) for evaluating the fluidUsesCurrent model can be used to design artificial heart valvesCan help in understand effects of disease (leaky valves)Related projects look at the behavior of the heart during a heart attackUltimately: real-time clinical work

    CS267 Lecture 1: Intro

  • Heart Simulation CalculationThe involves solving Navier-Stokes equations64^3 was possible on Cray YMP, but 128^3 required for accurate model (would have taken 3 years).Done on a Cray C90 -- 100x faster and 100x more memoryUntil recently, limited to vector machinesNeeds more features:Electrical model of the heart, and details of muscles, E.g., Chris JohnsonAndrew McCullochLungs, circulatory systems

    CS267 Lecture 1: Intro

  • Heart SimulationAnimation of lower portion of the heartSource: www.psc.org

    CS267 Lecture 1: Intro

  • Parallel Computing in Data AnalysisFinding information amidst large quantities of dataGeneral themes of sifting through large, unstructured data sets:Has there been an outbreak of some medical condition in a community?Which doctors are most likely involved in fraudulent charging to medicare?When should white socks go on sale?What advertisements should be sent to you?Data collected and stored at enormous speeds (Gbyte/hour)remote sensor on a satellitetelescope scanning the skiesmicroarrays generating gene expression datascientific simulations generating terabytes of dataNSA analysis of telecommunications

    CS267 Lecture 1: Intro

  • Why powerful computers are parallel

    CS267 Lecture 1: Intro

  • Tunnel Vision by ExpertsI think there is a world market for maybe five computers.Thomas Watson, chairman of IBM, 1943.There is no reason for any individual to have a computer in their homeKen Olson, president and founder of Digital Equipment Corporation, 1977.640K [of memory] ought to be enough for anybody.Bill Gates, chairman of Microsoft,1981.Slide source: Warfield et al.

    CS267 Lecture 1: Intro

  • Technology Trends: Microprocessor Capacity2X transistors/Chip Every 1.5 yearsCalled Moores Law

    Moores LawMicroprocessors have become smaller, denser, and more powerful.Gordon Moore (co-founder of Intel) predicted in 1965 that the transistor density of semiconductor chips would double roughly every 18 months. Slide source: Jack Dongarra

    CS267 Lecture 1: Intro

  • Impact of Device ShrinkageWhat happens when the feature size (transistor size) shrinks by a factor of x ?Clock rate goes up by x because wires are shorteractually less than x, because of power consumptionTransistors per unit area goes up by x2Die size also tends to increasetypically another factor of ~xRaw computing power of the chip goes up by ~ x4 !of which x3 is devoted either to parallelism or locality

    CS267 Lecture 1: Intro

  • Microprocessor Transistors per ChipGrowth in transistors per chipIncrease in clock rate

    CS267 Lecture 1: Intro

    Chart1

    1680000

    1680000

    1680000

    1680000

    1680000

    1680000

    1680000

    2800000

    2800000

    1750000

    9300000

    9300000

    360000

    540000

    1991.3

    1991.3

    850000

    800000

    850000

    1260000

    2300

    3500

    130000

    275000

    320000

    1200000

    1200000

    1200000

    1200000

    5000

    6500

    20000

    1989

    5000

    68000

    69000

    180000

    300000

    1200000

    5900000

    75000

    125000

    125000

    1300000

    1400000

    230000

Recommended

View more >