Chapter 1. Computer Abstractions and Technology. Jiang Jiang

Transcription

1 Chapter 1 Computer Abstractions and Technology Jiang Jiang jiangjiang@ic.sjtu.edu.cn

2 [Adapted from Computer Organization and Design, 4 th Edition, Patterson & Hennessy, 2008, MK] Chapter 1 Computer Abstractions and Technology 2

3 Computer Technology Dramatic Change Processor Performance: 100-fold in the last decade Speed: Doubled every 1.5 year Memory DRAM capacity: 64-fold in the last decade Memory speed: about 10% per year Cost per bit: Improves about 25% every year Disk Capacity: Doubled every year 250-fold in the last decade Chapter 1 Computer Abstractions and Technology 3

4 Technology Trends Electronics technology continues to evolve Increased capacity and performance Reduced cost The cost-performance ratio DRAM capacity Year Technology Relative performance/cost 1951 Vacuum tube Transistor Integrated circuit (IC) Very large scale IC (VLSI) 2,400, Ultra large scale IC 6,200,000,000 Chapter 1 Computer Abstractions and Technology 4

5 The Computer Revolution Progress in computer technology Underpinned by Moore s Law: Doubling very 1.5 years (18 months) Moore's law describes a long-term trend in the history of computing hardware: The number of transistors on an IC has doubled approximately every two years. The law is named after Intel cofounder Gordon E. Moore, who described the trend in his 1965 paper. 1.1Introduction Chapter 1 Computer Abstractions and Technology 5

6 Moore's law The capabilities of many digital electronic devices are strongly linked to Moore's law: processing speed, memory capacity, sensors and even the number and size of pixels in digital cameras. Chapter 1 Computer Abstractions and Technology 6

7 Novel Applications Progress in computer technology makes novel applications feasible Computers in automobiles Cell phones Human genome project World Wide Web Search Engines AI (Artificial Intelligence) Natural Language Understanding Apple Siri Google Now Computers are pervasive Chapter 1 Computer Abstractions and Technology 7

8 The Processor Market PC(Personal Computer ) vs. Tablet Phablet = phone + tablet Chapter 1 Computer Abstractions and Technology 8

9 Post-PC Era A term coined by Apple CEO Tim Cook The trend The personal computer (PC) is replaced by other devices such as smartphones and tablet computers A greater shift towards the use of cloud-based mobile and cross-platform Web The popularity of smartphones and tablets have influenced the economy of the computer industry; sales of PCs have steadily fallen since surge in popularity for post-pc devices Chapter 1 Computer Abstractions and Technology 9

10 Why ARM edition? ARM: Advanced RISC Machine To highlight the importance of embedded system to the computing industry ARM is the most popular instruction set for embedded devices. As of 2009, ARM processors account for approximately 90% of all embedded 32-bit RISC processors. ARM processors are used extensively in consumer electronics A MIPS processor is used to present the fundamentals of hardware technologies, pipelining, memory hierarchy and I/O Chapter 1 Computer Abstractions and Technology 10

11 What You Will Learn How programs are translated into the machine language And how the hardware executes them The hardware/software interface What determines program performance And how it can be improved How hardware designers improve performance What is parallel processing Chapter 1 Computer Abstractions and Technology 11

12 Classes of Computers Desktop computers General purpose, variety of software Subject to cost/performance tradeoff Server computers Network based High capacity, performance, reliability Range from small servers to building sized Embedded computers Hidden as components of systems Stringent power/performance/cost constraints Chapter 1 Computer Abstractions and Technology 12

13 Below Your Program (Programmer) Layered structure Application software Written in high-level language System software Compiler: translates HLL code to machine code Operating System: service code 1.2 Below Your Program Hardware Handling input/output Managing memory and storage Scheduling tasks & sharing resources Processor, memory, I/O controllers Chapter 1 Computer Abstractions and Technology 13

14 Levels of Program Code High-level language Level of abstraction closer to problem domain Provides for productivity and portability Assembly language Textual representation of instructions Hardware representation Binary digits (bits) Encoded instructions and data Chapter 1 Computer Abstractions and Technology 14

15 Computer Organization The BIG Picture All computers consist of five components Processor (1) Datapath (2) Control (3) Memory (4) Input devices (5) Output devices 1.3 Under the Covers Chapter 1 Computer Abstractions and Technology 15

16 Computer Organization Not all memory are created equally Cache: fast (expensive) memory Placed closer to the processor Main memory: Less expensive memory Can have more Input/output includes User-interface devices Display, keyboard, mouse Storage devices Hard disk/ssd, CD/DVD/Blu-ray, flash Network adapters For communicating with other computers Chapter 1 Computer Abstractions and Technology 16

17 Anatomy of a Computer Output device Network Interface Input device Input device Chapter 1 Computer Abstractions and Technology 17

18 Anatomy of a Mouse Optical mouse LED illuminates desktop Small low-res camera Basic image processor Looks for x, y movement Buttons & wheel Supersedes roller-ball mechanical mouse 3D mouse Accelerometer Gyroscope Chapter 1 Computer Abstractions and Technology 18

19 Through the Looking Glass LCD screen: picture elements (pixels) Mirrors content of frame buffer memory Chapter 1 Computer Abstractions and Technology 19

20 Opening the Box The free repair manual: Chapter 1 Computer Abstractions and Technology 20

21 A Safe Place for Data Volatile main memory Loses instructions and data when power off Non-volatile secondary memory Magnetic disk Flash memory Optical disk (CDROM, DVD, BD (Blu-ray Disc)) SSD (Solid-state drive) Chapter 1 Computer Abstractions and Technology 21

22 Networks Communication and resource sharing Local area network (LAN): Ethernet Within a building Wide area network (WAN: the Internet) Wireless network: WiFi, Bluetooth ac: 1 gigabit per second Chapter 1 Computer Abstractions and Technology 22

23 Apple Inc. Most valuable company in history On August 20, 2012 Apple closed at a record share price of $ It had a market capitalization of $ billion. This is the highest nominal market capitalization and surpasses a record set by Microsoft in Key people Steve Jobs (Co-founder, Chairman, former CEO) Timothy D. Cook (CEO) Jonathan Ive (the Senior Vice President of Industrial Design) Apple s best-known products HW: Macintosh computers (Mac mini, imac, Mac Pro, Macbook / Air/Pro, xserver), ipod, iphone (4, gyroscope), ipad (sold 3 million in 80 days), ipad 2 SW: Mac OS X, ios Chapter 1 Computer Abstractions and Technology 23

24 Apple Ax processors Ax is a series of ARM-based System on Chips (SoC) designed by Apple Inc. Apple A4 A package on package (PoP) system-on-a-chip (SoC) designed by Apple and manufactured by Samsung. It combines an ARM Cortex-A8 CPU with a PowerVR GPU. First used by iphone 4. Apple A5 A dual-core ARM Cortex-A9 CPU with NEON SIMD accelerator and a dual core PowerVR SGX543MP2 GPU. First used by ipad 2. Apple A5x A dual-core ARM Cortex-A9 CPU. A quad-core PowerVR SGX543MP4 GPU as well as a quad-channel memory controller. First used by the new ipad. Chapter 1 Computer Abstractions and Technology 24

25 Apple Ax processors Apple A6 Custom Apple-designed dual-core CPU, called Swift, using modified ARM instruction set called ARMv7s, which is not a strict ARM licensed ISA. The support for VFPv4 extensions suggests that it is a Cortex-A15 class CPU. The GPU is a triple core PowerVR SGX543MP3. First used by iphone 5. Apple A6x Dual Swift cores, just as the A6, but has a new quad core GPU, quad channel memory and higher clock speed. First used by the fourth generation ipad. Apple A7 64-bit, Apple-designed ARMv8 based dual-core CPU Swift core, running at a 1.7 GHz and an integrated quad-core PowerVR GPU. First used by iphone 5S. Technology Trends ARM-based, modified ARM instruction set Multi-core 32-bit è 64-bit Chapter 1 Computer Abstractions and Technology 25

26 What Learn from Apple Forward-looking technologies Accelerometer, gyroscope, PoP, Siri Powerpc -> Intel -> ARM (?) Marketing Strategy Capturing the market quickly vs. New product updated almost annually The perfect combination of art and technology Perfect Industrial Design Close vs. Open ios vs. Android Apple Store Safe Place to download Distribution system Chapter 1 Fundamentals of Quantitative Design and Analysis 26

27 Inside the Processor (CPU) Datapath Performs operations on data Control Sequences datapath, memory,... Cache memory Small fast SRAM memory for immediate access to data Chapter 1 Computer Abstractions and Technology 27

28 Inside the Processor n AMD Barcelona: 4 processor cores Chapter 1 Computer Abstractions and Technology 28

29 Abstractions The BIG Picture Digging into the depths of HW/SW reveals more information Low-level details are hidden to offer a simple interface to higher levels The use of such layers, or abstraction, is a principal technique for design very sophisticated computer systems Instruction set architecture (ISA) One of the most important abstraction is the interface between hardware and the lowest-level software Chapter 1 Computer Abstractions and Technology 29

30 Instruction Set Architecture Anything programmers need to know to make a binary machine language program work correctly, including 7 dimensions: Registers Organization of programmable storage Data types and data structures: encoding and representations Instruction set Instruction formats Modes of addressing: accessing data items and instructions Exceptional conditions Chapter 1 Computer Abstractions and Technology 30

31 Instruction Set Architecture Main examples of ISA DEC Alpha (v1, v3) HP PA-RISC (v1.1, v2.0) SGI MIPS (MIPS I-V) Sun SPARC (v8, v9, Ultra versions) Intel (8086,80x Pentium MMX, II, III, IV...) Is binary compatibility is extraordinarily important? (True or False) Ecosystem Chapter 1 Computer Abstractions and Technology 31

32 Application Binary Interface OS will encapsulate the details of Doing I/O Allocating memory and Other low-level system function Application binary interface (ABI) The combination of the basic instruction set and the OS interface provided for application programmers, or the user portion of the instruction set plus the OS interface used by programmers Defines a standard for binary portability across computers Chapter 1 Computer Abstractions and Technology 32

33 Advantages of Abstraction Abstraction helps us deal with complexity Hide lower-level detail Hierarchical layers Interface between adjacent layers Abstraction enables different implementations of varying cost and performance An instruction set architecture(isa) allows computer designers to talk about functions independently from the actual HW Chapter 1 Computer Abstractions and Technology 33

34 Computer Architecture Computer Architecture = Instruction Set Architecture + Machine Organization + Hardware Computer architecture is described as... the attributes of a [computing] system as seen by the programmer, i.e. the conceptual structure and functional behavior, as distinct from the organization of the data flows and controls the logic design, and physical implementation. (Amdahl et al. 1964) Chapter 1 Computer Abstractions and Technology 34

35 Computer Architecture ISA is a subset of Computer Architecture Chapter 1 Computer Abstractions and Technology 35

36 Defining Performance Which airplane has the best performance? 1.4 Performance Boeing 777 Boeing 747 BAC/Sud Concorde Douglas DC-8-50 Boeing 777 Boeing 747 BAC/Sud Concorde Douglas DC Passenger Capacity E+ 04 Cruising Range (miles) Boeing 777 Boeing 747 BAC/Sud Concorde Douglas DC-8-50 Boeing 777 Boeing 747 BAC/Sud Concorde Douglas DC Cruising Speed (mph) 0 1E+0 5 2E+0 5 3E+0 5 Passengers x mph 4E+0 5 Chapter 1 Computer Abstractions and Technology 36

37 MIPS, FLOPS MIPS: Million Instructions Per Second Used to measure the integer performance of a computer VAX 11/780: a 1 MIPS machine FLOPS: FLoating point Operations Per Second A measure of a computer's performance, especially in fields of scientific calculations that make heavy use of floating point As of June 2013, the No. 1 supercomputers of Top500 is Tianhe-2 (NUDT), which is petaflops using 3,120,000 cores, 17.8 MW Computer Performance Name FLOPS yottaflops zettaflops exaflops petaflops teraflops gigaflops 10 9 megaflops 10 6 kiloflops 10 3 Chapter 1 Computer Abstractions and Technology 37

38 Response Time and Throughput Response time/ Execution time How long it takes to do a task Throughput Total work done per unit time E.g., tasks/transactions/ per hour How are response time and throughput affected by Replacing the processor with a faster version? Adding more processors? From different perspectives Response time: individual user Throughput: system We ll focus on response time for now Chapter 1 Computer Abstractions and Technology 38

39 Relative Performance Define Performance = 1/Execution Time X is n time faster than Y Performance Performance X Y = Execution time Y Execution time X = n Example: time taken to run a program 10s on A, 15s on B Execution Time B / Execution Time A = 15s / 10s = 1.5 So A is 1.5 times faster than B Chapter 1 Computer Abstractions and Technology 39

40 Measuring Execution Time Elapsed time Total response time, including all aspects Processing, OS overhead, I/O, idle time Determines system performance CPU time (CPU execution time) Time spent processing a given job Discounts I/O time, other jobs shares Comprises user CPU time and system CPU time User CPU time: the CPU time spent in a program itself System CPU time: the CPU time spent in the OS performing tasks on behalf of the program Different programs are affected differently by CPU and system performance Chapter 1 Computer Abstractions and Technology 40

41 CPU Clocking Operation of digital hardware governed by a constant-rate clock Clock (cycles) Data transfer and computation Update state Clock period Clock Period (Cycle Time): duration of a clock cycle e.g., 250ps = 0.25ns = s Clock Frequency (Rate): cycles per second e.g., 4.0GHz = 4000MHz = Hz Chapter 1 Computer Abstractions and Technology 41

42 CPU Time CPU Time = CPU Clock Cycles Clock Cycle Time = CPU Clock Cycles Clock Rate Performance improved by Reducing number of clock cycles Increasing clock rate Hardware designer must often trade off clock rate(frequency) against cycle count Why? Less work for higher frequency Chapter 1 Computer Abstractions and Technology 42

43 CPU Time Example Computer A: 2GHz clock, 10s CPU time Designing Computer B Aim for 6s CPU time Can do faster clock, but causes 1.2 clock cycles How fast (Clock rate) must Computer B clock be? Clock Rate B = Clock Cycles CPU Time B B = 1.2 Clock Cycles 6s A Clock Cycles A = CPU Time A Clock Rate A = 10s 2GHz = Clock Rate B = s 9 = s 9 = 4GHz Chapter 1 Computer Abstractions and Technology 43

44 Instruction Count and CPI Clock Cycles = Instruction Count Cycles per Instruction CPU Time = Instruction Count CPI Clock Cycle Time = Instruction Count Clock Rate CPI Instruction Count for a program Determined by program, ISA and compiler Average cycles per instruction Determined by CPU hardware If different instructions have different CPI Average CPI affected by instruction mix Chapter 1 Computer Abstractions and Technology 44

45 CPI Example Computer A: Cycle Time = 250ps, CPI = 2.0 Computer B: Cycle Time = 500ps, CPI = 1.2 Same ISA Which is faster, and by how much? CPU Time A CPU Time B CPU Time B CPU Time A = Instruction Count CPI A = I ps = I 500ps = Instruction Count CPI Cycle Time B B = I ps = I 600ps I 600ps = I 500ps = 1.2 Cycle Time A A is faster by this much Chapter 1 Computer Abstractions and Technology 45

46 CPI in More Detail If different instruction classes take different numbers of cycles Clock Cycles = n i= 1 (CPIi Instruction Count i) Weighted average CPI CPI = Clock Cycles Instruction Count = n i= 1 CPI i Instruction Count i Instruction Count Relative frequency Chapter 1 Computer Abstractions and Technology 46

47 CPI Example Alternative compiled code sequences using instructions in classes A, B, C Class A B C CPI for class IC in sequence IC in sequence Sequence 1: IC = 5 Clock Cycles = = 10 Avg. CPI = 10/5 = 2.0 Sequence 2: IC = 6 Clock Cycles = = 9 Avg. CPI = 9/6 = 1.5 Chapter 1 Computer Abstractions and Technology 47

48 Performance Summary The BIG Picture Instructions Clock cycles CPU Time = Program Instruction Seconds Clock cycle Performance depends on Algorithm: affects IC, possibly CPI Programming language: affects IC, CPI Compiler: affects IC, CPI Instruction set architecture: affects IC, CPI, Clock rate Chapter 1 Computer Abstractions and Technology 48

49 Understanding Performance Algorithm Determines number of operations executed Programming language, compiler, architecture Determine number of machine instructions executed per operation Processor and memory system Determine how fast instructions are executed I/O system (including OS) Determines how fast I/O operations are executed Chapter 1 Computer Abstractions and Technology 49

50 Power Trends 1.5 The Power Wall Dynamic power: Power that is consumed during switching Typically, the voltage was reduced about 15% per generation Power = Capacitive load Voltage 2 Frequency 30 5V 1V 1000 Chapter 1 Computer Abstractions and Technology 50

51 Reducing Power Suppose a new CPU has 85% of capacitive load of old CPU 15% voltage and 15% frequency reduction P 4 2 new Cold 0.85 (Vold 0.85) Fold 0.85 = = 0.85 = 2 P C V F old The power wall old We can t reduce voltage further We can t remove more heat old How else can we improve performance? old 0.52 Chapter 1 Computer Abstractions and Technology 51

52 Uniprocessor Performance Constrained by power, instruction-level parallelism, memory latency 1.6 The Sea Change: The Switch to Multiprocessors Chapter 1 Computer Abstractions and Technology 52

53 Multiprocessors Multicore microprocessors More than one processor per chip Requires explicitly parallel programming Instruction level parallelism (ILP) Hardware executes multiple instructions at once Hidden from the programmer: implicit Why programmer hard to explicitly parallel Performance oriented programming Scheduling for load balancing Optimizing communication and synchronization Chapter 1 Computer Abstractions and Technology 53

54 Manufacturing ICs 1.7 Real Stuff: The AMD Opteron X4 Yield: proportion of working dies per wafer Chapter 1 Computer Abstractions and Technology 54

55 AMD Opteron X2 Wafer n n X2: 300mm wafer, 117 chips, 90nm technology X4: 45nm technology Chapter 1 Computer Abstractions and Technology 55

56 Integrated Circuit Cost Cost per die = Cost per wafer Dies per wafer Yield Dies per wafer Wafer area Die area Yield = (1+ (Defects per 1 area Die area/2)) 2 Nonlinear relation to area and defect rate Wafer cost and area are fixed Defect rate determined by manufacturing process Die area determined by architecture and circuit design Chapter 1 Computer Abstractions and Technology 56

57 SPEC CPU Benchmark Benchmark Programs used to measure performance Supposedly typical of actual workload Standard Performance Evaluation Corp (SPEC) Develops benchmarks for CPU, I/O, Web, SPEC CPU2006: combined performance of CPU, memory and compiler SPECpower_ssj2008: evaluates the energy efficiency of server systems Chapter 1 Computer Abstractions and Technology 57

58 SPEC CPU2006 Consisting of a set of benchmarks CINT2006 (SPECint): 12 programs, testing integer arithmetic, with programs such as compilers, interpreters, word processors, chess programs etc. CFP2006 (SPECfp) : 17 programs, testing floating point performance, with physical simulations, 3D graphics, image processing, computational chemistry etc. Elapsed time to execute a selection of programs Negligible I/O, so focuses on CPU performance Normalize relative to reference machine A Sun Ultra Enterprise 2 workstation with a 296-MHz UltraSPARC II processor Summarize as geometric mean of performance ratios CINT2006 (integer) and CFP2006 (floating-point) n n i=1 Execution time ratio i Chapter 1 Computer Abstractions and Technology 58

59 CINT2006 for Opteron X Name Description IC 10 9 CPI Tc (ns) Exec time Ref time SPECratio perl Interpreted string processing 2, , bzip2 Block-sorting compression 2, , gcc GNU C Compiler 1, , mcf Combinatorial optimization ,345 9, go Go game (AI) 1, , hmmer Search gene sequence 2, , sjeng Chess game (AI) 2, , libquantum Quantum computer simulation 1, ,047 20, h264avc Video compression 3, , omnetpp Discrete event simulation , astar Games/path finding 1, , xalancbmk XML parsing 1, ,143 6, Geometric mean 11.7 High cache miss rates Chapter 1 Computer Abstractions and Technology 59

60 SPEC Power Benchmark Power consumption of server at different workload levels Performance: ssj_ops/sec Server side Java operations per second Power: Watts (Joules/sec) Overall ssj_ops per Watt ssj_ops i power i= 0 = i= 0 i Chapter 1 Computer Abstractions and Technology 60

61 SPECpower_ssj2008 for X4 Target Load % Performance (ssj_ops/sec) Average Power (Watts) 100% 231, % 211, % 185, % 163, % 140, % 118, % 920, % 70, % 47, % 23, % Overall sum 1,283,590 2,605 ssj_ops/ power 493 Chapter 1 Computer Abstractions and Technology 61

62 Other Benchmarks LINPACK A measure of a system's floating point computing power Makes use of the BLAS (Basic Linear Algebra Subprograms) libraries for performing basic vector and matrix operations Benchmark for TOP500 supercomputer EEMBC Embedded Microprocessor Benchmark Consortium Benchmarks for the hardware and software used in embedded systems Chapter 1 Computer Abstractions and Technology 62

63 Pitfall: Amdahl s Law Pitfall: Expecting the improvement of one aspect of a computer to increase overall performance by an amount proportional to the size of the improvement T improvement affected T improved = + factor unaffected Example: multiply accounts for 80s/100s How much improvement in multiply performance to get 5 overall? = + 20 Can t be done! n T 1.8 Fallacies and Pitfalls Corollary: make the common case fast Chapter 1 Computer Abstractions and Technology 63

64 Amdahl s Law Named after computer architect Gene Amdahl Often used in parallel computing to predict the theoretical maximum speedup using multiple processors maximum speedup In economics, law of diminishing marginal returns Chapter 1 Computer Abstractions and Technology 64

65 Fallacy: Low Power at Idle Fallacy: Computers at low utilization use little power Look back at X4 power benchmark At 100% load: 295W At 50% load: 246W (83%) At 10% load: 180W (61%) Google data center Mostly operates at 10% 50% load At 100% load less than 1% of the time Consider designing processors to make power proportional to load Chapter 1 Computer Abstractions and Technology 65

66 Pitfall: MIPS as a Performance Metric Pitfall: Using a subset of the performance equation as a performance metric Clock rate, instruction count, or CPI MIPS: Millions of Instructions Per Second Doesn t account for Differences in ISAs between computers Differences in complexity between instructions MIPS = Instruction count Execution time 10 Instruction count = Instruction count CPI 10 Clock rate 6 6 Clock rate = 6 CPI 10 CPI varies between programs on a given CPU Chapter 1 Computer Abstractions and Technology 66

67 Concluding Remarks Cost/performance is improving Due to underlying technology development Hierarchical layers of abstraction In both hardware and software Instruction set architecture The hardware/software interface Execution time: the best performance measure Power is a limiting factor Use parallelism to improve performance 1.9 Concluding Remarks Chapter 1 Computer Abstractions and Technology 67

68 Terminology CPU: Central Processing Unit Microprocessor Processor PE: Processing Element Microarchitecture Chapter 1 Computer Abstractions and Technology 68