Survey
* Your assessment is very important for improving the workof artificial intelligence, which forms the content of this project
* Your assessment is very important for improving the workof artificial intelligence, which forms the content of this project
CS267/E233 Applications of Parallel Computers Lecture 1: Introduction Kathy Yelick [email protected] www.cs.berkeley.edu/~yelick/cs267_spr07 01/17/2007 CS267-Lecture 1 1 Why powerful computers are parallel circa 1991-2006 01/17/2007 CS267-Lecture 1 2 Tunnel Vision by Experts • “I 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 home” - Ken Olson, president and founder of Digital Equipment Corporation, 1977. • “640K [of memory] ought to be enough for anybody.” - Bill Gates, chairman of Microsoft,1981. • “On several recent occasions, I have been asked whether parallel computing will soon be relegated to the trash heap reserved for promising technologies that never quite make it.” - Ken Kennedy, CRPC Directory, 1994 01/17/2007 CS267-Lecture 1 Slide source: Warfield et al. 3 Technology Trends: Microprocessor Capacity Moore’s Law 2X transistors/Chip Every 1.5 years Called “Moore’s Law” Microprocessors 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 01/17/2007 CS267-Lecture 1 4 Microprocessor Transistors per Chip • Growth in transistors per chip • Increase in clock rate 100,000,000 1000 10,000,000 1,000,000 i80386 i80286 100,000 R3000 R2000 100 Clock Rate (MHz) Transistors R10000 Pentium 10 1 i8086 10,000 i8080 i4004 1,000 1970 1975 1980 1985 1990 1995 2000 2005 Year 01/17/2007 CS267-Lecture 1 0.1 1970 1980 1990 2000 Year 5 Impact of Device Shrinkage • What happens when the feature size (transistor size) shrinks by a factor of x ? • Clock rate goes up by x because wires are shorter - actually less than x, because of power consumption • Transistors per unit area goes up by x2 • Die size also tends to increase - typically another factor of ~x • Raw computing power of the chip goes up by ~ x4 ! - typically x3 is devoted either on-chip - parallelism: hidden parallelism such as ILP - locality: caches 01/17/2007 CS267-Lecture 1 6 But there are limiting forces Manufacturing costs and yield problems limit use of density • Moore’s 2nd law (Rock’s law): costs go up Demo of 0.06 micron CMOS Source: Forbes Magazine • Yield -What percentage of the chips are usable? -E.g., Cell processor (PS3) is sold with 7 out of 8 “on” to improve yield 01/17/2007 CS267-Lecture 1 7 Power Density Limits Serial Performance 01/17/2007 CS267-Lecture 1 8 More Limits: How fast can a serial computer be? 1 Tflop/s, 1 Tbyte sequential machine r = 0.3 mm • Consider the 1 Tflop/s sequential machine: - Data must travel some distance, r, to get from memory to CPU. - To get 1 data element per cycle, this means 1012 times per second at the speed of light, c = 3x108 m/s. Thus r < c/1012 = 0.3 mm. • Now put 1 Tbyte of storage in a 0.3 mm x 0.3 mm area: - Each bit occupies about 1 square Angstrom, or the size of a small atom. • No choice but parallelism 01/17/2007 CS267-Lecture 1 9 Revolution is Happening Now • Chip density is continuing increase ~2x every 2 years - Clock speed is not - Number of processor cores may double instead • There is little or no hidden parallelism (ILP) to be found • Parallelism must be exposed to and managed by software Source: Intel, Microsoft (Sutter) and Stanford (Olukotun, Hammond) 01/17/2007 CS267-Lecture 1 10 Why Parallelism (2007)? • These arguments are no long theoretical • All major processor vendors are producing multicore chips - Every machine will soon be a parallel machine - All programmers will be parallel programmers??? • New software model - Want a new feature? Hide the “cost” by speeding up the code first - All programmers will be performance programmers??? • Some may eventually be hidden in libraries, compilers, and high level languages - But a lot of work is needed to get there • Big open questions: - What will be the killer apps for multicore machines - How should the chips be designed, and how will they be programmed? 01/17/2007 CS267-Lecture 1 11 Outline all • Why powerful computers must be parallel processors Including your laptop • Large important problems require powerful computers Even computer games • Why writing (fast) parallel programs is hard • Principles of parallel computing performance • Structure of the course 01/17/2007 CS267-Lecture 1 12 Why we need powerful computers 01/17/2007 CS267-Lecture 1 13 Units of Measure in HPC • High Performance Computing (HPC) units are: - Flop: floating point operation - Flops/s: floating point operations per second - Bytes: size of data (a double precision floating point number is 8) • Typical sizes are millions, billions, trillions… Mega Mflop/s = 106 flop/sec Mbyte = 220 = 1048576 ~ 106 bytes Giga Tera Peta Exa Zetta Gflop/s = 109 flop/sec Tflop/s = 1012 flop/sec Pflop/s = 1015 flop/sec Eflop/s = 1018 flop/sec Zflop/s = 1021 flop/sec Gbyte = 230 ~ 109 bytes Tbyte = 240 ~ 1012 bytes Pbyte = 250 ~ 1015 bytes Ebyte = 260 ~ 1018 bytes Zbyte = 270 ~ 1021 bytes Yotta Yflop/s = 1024 flop/sec Ybyte = 280 ~ 1024 bytes • See www.top500.org for current list of fastest machines 01/17/2007 CS267-Lecture 1 14 Simulation: The Third Pillar of Science • Traditional scientific and engineering paradigm: 1) Do theory or paper design. 2) 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: 3) Use high performance computer systems to simulate the phenomenon - Base on known physical laws and efficient numerical methods. 01/17/2007 CS267-Lecture 1 15 Some Particularly Challenging Computations • Science - Global climate modeling Biology: genomics; protein folding; drug design Astrophysical modeling Computational Chemistry Computational Material Sciences and Nanosciences • Engineering - Semiconductor design Earthquake and structural modeling Computation fluid dynamics (airplane design) Combustion (engine design) Crash simulation • Business - Financial and economic modeling - Transaction processing, web services and search engines • Defense - Nuclear weapons -- test by simulations - Cryptography 01/17/2007 CS267-Lecture 1 16 Economic Impact of HPC • Airlines: - 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) for - device 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. 01/17/2007 CS267-Lecture 1 17 $5B World Market in Technical Computing 1998 1999 2000 2001 2002 2003 100% Other Technical Management and Support Simulation 90% Scientific Research and R&D 80% Mechanical Design/Engineering Analysis 70% Mechanical Design and Drafting 60% Imaging 50% Geoscience and Geoengineering 40% Electrical Design/Engineering Analysis Economics/Financial 30% Digital Content Creation and Distribution 20% Classified Defense 10% Chemical Engineering Biosciences 0% Source: IDC 2004, from NRC Future of Supercomputing Report 01/17/2007 CS267-Lecture 1 18 Global Climate Modeling • Problem Problem is to compute: f(latitude, longitude, elevation, time) temperature, pressure, humidity, wind velocity • Approach: - Discretize the domain, e.g., a measurement point every 10 km - Devise an algorithm to predict weather at time t+dt given t • Uses: - Predict major events, e.g., El Nino - Use in setting air emissions standards Source: http://www.epm.ornl.gov/chammp/chammp.html 01/17/2007 CS267-Lecture 1 19 Global Climate Modeling Computation • One piece is modeling the fluid flow in the atmosphere - Solve Navier-Stokes equations - Roughly 100 Flops per grid point with 1 minute timestep • Computational requirements: - To match real-time, need 5 x 1011 flops in 60 seconds = 8 Gflop/s - Weather prediction (7 days in 24 hours) 56 Gflop/s - Climate prediction (50 years in 30 days) 4.8 Tflop/s - To use in policy negotiations (50 years in 12 hours) 288 Tflop/s • To 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 more • Current models are coarser than this 01/17/2007 CS267-Lecture 1 20 High Resolution Climate Modeling on NERSC-3 – P. Duffy, et al., LLNL 01/17/2007 CS267-Lecture 1 21 A 1000 Year Climate Simulation • Demonstration of the Community Climate Model (CCSM2) • A 1000-year simulation shows long-term, stable representation of the earth’s climate. • 760,000 processor hours used • Temperature change shown • Warren 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.pdf 01/17/2007 CS267-Lecture 1 22 Climate Modeling on the Earth Simulator System Development of ES started in 1997 in order to make a comprehensive understanding of global environmental changes such as global warming. Its construction was completed at the end of February, 2002 and the practical operation started from March 1, 2002 35.86Tflops (87.5% of the peak performance) is achieved in the Linpack benchmark (world’s fastest machine from 2002-2004). 26.58Tflops was obtained by a global atmospheric circulation code. 01/17/2007 CS267-Lecture 1 23 Astrophysics: Binary Black Hole Dynamics • Massive 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 ! 01/17/2007 CS267-Lecture 1 24 01/17/2007 CS267-Lecture 1 25 Heart Simulation • Problem is to compute blood flow in the heart • Approach: - Modeled as an elastic structure in an incompressible fluid. - The “immersed boundary method” due to Peskin and McQueen. - 20 years of development in model - Many applications other than the heart: blood clotting, inner ear, paper making, embryo growth, and others - Use a regularly spaced mesh (set of points) for evaluating the fluid • Uses - Current model can be used to design artificial heart valves - Can help in understand effects of disease (leaky valves) - Related projects look at the behavior of the heart during a heart attack - Ultimately: real-time clinical work 01/17/2007 CS267-Lecture 1 26 Heart Simulation Calculation The involves solving Navier-Stokes equations - 64^3 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 memory - Until recently, limited to vector machines - Needs more features: - Electrical model of the heart, and details of muscles, E.g., - Chris Johnson - Andrew McCulloch - Lungs, circulatory systems 01/17/2007 CS267-Lecture 1 27 Heart Simulation Animation of lower portion of the heart 01/17/2007 CS267-Lecture 1 Source: www.psc.org 28 Parallel Computing in Data Analysis • Finding information amidst large quantities of data • General 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 satellite - telescope scanning the skies - microarrays generating gene expression data - scientific simulations generating terabytes of data - NSA analysis of telecommunications 01/17/2007 CS267-Lecture 1 29 Performance on Linpack Benchmark www.top500.org 01/17/2007 CS267-Lecture 1 30 01/17/2007 CS267-Lecture 1 31 Why writing (fast) parallel programs is hard 01/17/2007 CS267-Lecture 1 32 Principles of Parallel Computing • Finding enough parallelism (Amdahl’s Law) • Granularity • Locality • Load balance • Coordination and synchronization • Performance modeling All of these things makes parallel programming even harder than sequential programming. 01/17/2007 CS267-Lecture 1 33 “Automatic” Parallelism in Modern Machines • Bit level parallelism - within floating point operations, etc. • Instruction level parallelism (ILP) - multiple instructions execute per clock cycle • Memory system parallelism - overlap of memory operations with computation • OS parallelism - multiple jobs run in parallel on commodity SMPs Limits to all of these -- for very high performance, need user to identify, schedule and coordinate parallel tasks 01/17/2007 CS267-Lecture 1 34 Finding Enough Parallelism • Suppose only part of an application seems parallel • Amdahl’s law - let s be the fraction of work done sequentially, so (1-s) is fraction parallelizable - P = number of processors Speedup(P) = Time(1)/Time(P) <= 1/(s + (1-s)/P) <= 1/s • Even if the parallel part speeds up perfectly performance is limited by the sequential part 01/17/2007 CS267-Lecture 1 35 Overhead of Parallelism • Given enough parallel work, this is the biggest barrier to getting desired speedup • Parallelism overheads include: - cost of starting a thread or process - cost of communicating shared data - cost of synchronizing - extra (redundant) computation • Each of these can be in the range of milliseconds (=millions of flops) on some systems • Tradeoff: Algorithm needs sufficiently large units of work to run fast in parallel (I.e. large granularity), but not so large that there is not enough parallel work 01/17/2007 CS267-Lecture 1 36 Locality and Parallelism Conventional Storage Proc Hierarchy Cache L2 Cache Proc Cache L2 Cache Proc Cache L2 Cache L3 Cache L3 Cache Memory Memory Memory • Large memories are slow, fast memories are small • Storage hierarchies are large and fast on average • Parallel processors, collectively, have large, fast cache - the slow accesses to “remote” data we call “communication” • Algorithm should do most work on local data 01/17/2007 CS267-Lecture 1 potential interconnects L3 Cache Processor-DRAM Gap (latency) CPU “Moore’s Law” Processor-Memory Performance Gap: (grows 50% / year) DRAM DRAM 7%/yr. 100 10 1 µProc 60%/yr. 1980 1981 1982 1983 1984 1985 1986 1987 1988 1989 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 Performance 1000 Time 01/17/2007 CS267-Lecture 1 38 Load Imbalance • Load imbalance is the time that some processors in the system are idle due to - insufficient parallelism (during that phase) - unequal size tasks • Examples of the latter - adapting to “interesting parts of a domain” - tree-structured computations - fundamentally unstructured problems • Algorithm needs to balance load 01/17/2007 CS267-Lecture 1 39 Measuring Performance 01/17/2007 CS267-Lecture 1 40 Improving Real Performance Peak Performance grows exponentially, a la Moore’s Law In 1990’s, peak performance increased 100x; in 2000’s, it will increase 1000x 1,000 But efficiency (the performance relative to the hardware peak) has declined was 40-50% on the vector supercomputers of 1990s now as little as 5-10% on parallel supercomputers of today Close the gap through ... Mathematical methods and algorithms that achieve high performance on a single processor and scale to thousands of processors More efficient programming models and tools for massively parallel supercomputers 01/17/2007 CS267-Lecture 1 100 Teraflops Peak Performance Performance Gap 10 1 Real Performance 0.1 1996 2000 2004 41 Performance Levels • Peak advertised performance (PAP) - You can’t possibly compute faster than this speed • LINPACK - The “hello world” program for parallel computing - Solve Ax=b using Gaussian Elimination, highly tuned • Gordon Bell Prize winning applications performance - The right application/algorithm/platform combination plus years of work • Average sustained applications performance - What one reasonable can expect for standard applications When reporting performance results, these levels are often confused, even in reviewed publications 01/17/2007 CS267-Lecture 1 42 Performance on Linpack Benchmark www.top500.org 01/17/2007 CS267-Lecture 1 43 Performance Levels (for example on NERSC3) • Peak advertised performance (PAP): 5 Tflop/s • LINPACK (TPP): 3.05 Tflop/s • Gordon Bell Prize winning applications performance : 2.46 Tflop/s - Material Science application at SC01 • Average sustained applications performance: ~0.4 Tflop/s - Less than 10% peak! 01/17/2007 CS267-Lecture 1 44 Course Organization 01/17/2007 CS267-Lecture 1 45 Course Mechanics • This class is listed as both a CS and Engineering class • Normally a mix of CS, EE, and other engineering and science students • This class seems to be about: - 15 grads + 4 undergrads - X% CS, EE, Other (BioPhys, BioStat, Civil, Mechanical, Nuclear) • For final projects we encourage interdisciplinary teams - This is the way parallel scientific software is generally built 01/17/2007 CS267-Lecture 1 46 Rough Schedule of Topics • Parallel Programming Models and Machines - Shared Memory and Multithreading - Distributed Memory and Message Passing - Data parallelism • Parallel languages and libraries - Shared memory threads and OpenMP - MPI - Languages (UPC) • “Seven Dwarfs” of Scientific Computing - Dense Linear Algebra Structured grids Particle methods Sparse matrices Spectral methods Unstructured Grids Spectral methods (FFTs, etc) • Applications: biology, climate, combustion, astrophysics, … • Work: 3 homeworks + 1 final project 01/17/2007 CS267-Lecture 1 47 Lecture Scribes • Each student should scribe ~2 lectures • Next lecture is on single processor performance (and architecture) • Sign up on list to choose 1 topic in first part of the course 01/17/2007 CS267-Lecture 1 48 Reading Materials • Some on-line texts: - Demmel’s notes from CS267 Spring 1999, which are similar to 2000 and 2001. However, they contain links to html notes from 1996. - http://www.cs.berkeley.edu/~demmel/cs267_Spr99/ - Simon’s notes from Fall 2002 - http://www.nersc.gov/~simon/cs267/ - Ian Foster’s book, “Designing and Building Parallel Programming”. - http://www-unix.mcs.anl.gov/dbpp/ • Potentially useful texts: - “Sourcebook for Parallel Computing”, by Dongarra, Foster, Fox, .. - A general overview of parallel computing methods - “Performance Optimization of Numerically Intensive Codes” by Stefan Goedecker and Adolfy Hoisie - This is a practical guide to optimization, mostly for those of you who have never done any optimization • More pointers will be on the web page 01/17/2007 CS267-Lecture 1 49 What you should get out of the course In depth understanding of: • When is parallel computing useful? • Understanding of parallel computing hardware options. • Overview of programming models (software) and tools. • Some important parallel applications and the algorithms • Performance analysis and tuning 01/17/2007 CS267-Lecture 1 50 Administrative • Information Instructors: - Kathy Yelick, 777 Soda, [email protected] - TA: Marghoob Mohiyuddin ([email protected]) - Office hours TBD • Lecture notes are based on previous semester notes: - Jim Demmel, David Culler, David Bailey, Bob Lucas, Kathy Yelick and Horst Simon - Much material stolen from others (with sources noted) • Most class material and lecture notes are at: - http://www.cs.berkeley.edu/~yelick/cs267_spr07 01/17/2007 CS267-Lecture 1 51 Extra slides 01/17/2007 CS267-Lecture 1 52 Transaction Processing (mar. 15, 1996) 25000 other Throughput (tpmC) 20000 Tandem Himalaya IBM PowerPC 15000 DEC Alpha SGI PowerChallenge HP PA 10000 5000 0 0 20 40 60 80 100 120 Processors • Parallelism is natural in relational operators: select, join, etc. • Many difficult issues: data partitioning, locking, threading. 01/17/2007 CS267-Lecture 1 53 SIA Projections for Microprocessors 1000 100 Feature Size (microns) 10 Transistors per chip x 106 1 0.1 2010 2007 2004 2001 1998 0.01 1995 Feature Size (microns) & Million Transistors per chip Compute power ~1/(Feature Size)3 Year of Introduction based on F.S.Preston, 1997 01/17/2007 CS267-Lecture 1 54 Much of the Performance is from Parallelism Thread-Level Parallelism? Instruction-Level Parallelism Bit-Level Parallelism Name 01/17/2007 CS267-Lecture 1 55 Performance on Linpack Benchmark www.top500.org 100000 Earth Simulator 10000 ASCI White ASCI Red 1000 Rmax max Rmax mean Rmax min Rmax 100 System 10 04 03 Ju n 03 D ec 02 Ju n 02 D ec 01 Ju n 01 D ec 00 Ju n 00 D ec 99 Ju n 99 D ec 98 Ju n 98 D ec 97 Ju n 97 D ec 96 Ju n 96 D ec 95 Ju n 95 D ec 94 Ju n 94 D ec 93 Ju n D ec Ju n 93 1 0.1 Nov 2004: IBM Blue Gene L, 70.7 Tflops Rmax 01/17/2007 CS267-Lecture 1 56