Introduction

 This course is all about how computers work  But what do we mean by a computer?  Different types: desktop, servers, embedded devices  Different uses: automobiles, graphics, finance, genomics…  Different manufacturers: Intel, Apple, IBM, Microsoft, Sun…  Different underlying technologies and different costs!  Analogy: Consider a course on “automotive vehicles”  Many similarities from vehicle to vehicle (e.g., wheels)  Huge differences from vehicle to vehicle (e.g., gas vs. electric)  Best way to learn:  Focus on a specific instance and learn how it works  While learning general principles and historical perspectives 2

Why learn this stuff?  You want to call yourself a “computer scientist”  You want to build software people use (need performance)  You need to make a purchasing decision or offer “expert” advice  Both Hardware and Software affect performance:  Algorithm determines number of source-level statements  Language/Compiler/Architecture determine machine instructions (Chapter 2 and 3)  Processor/Memory determine how fast instructions are executed (Chapter 5, 6, and 7)  Assessing and Understanding Performance in Chapter 4 3

What is a computer?  Components:  Input (mouse, keyboard)  Output (display, printer)  Memory (disk drives, DRAM, SRAM, CD)  Network  Our primary focus: the processor (datapath and control)  Implemented using millions of transistors  Impossible to understand by looking at each transistor  We need to learn the logical design of each component 4

 Embedded processors prevail  Cell phones, car computers, digital TVs, videogame consoles, …  Designed to run dedicated applications  Annual growth rate of 40%  9% for desktops and servers Number of Distinct Processors Sold Millions of computers Embedded computer Desktops Servers 5

Uniprocessor Performance 6

Contributor 1: Technology  Processor  logic capacity:about 30% per year  clock rate:about 20% per year  Memory  DRAM capacity: about 60% per year (4x every 3 years)  Memory speed: about 10% per year  Cost per bit: improves about 25% per year  Disk  capacity: about 60% per year 7

Technology Improvement  Moore's law  The number of transistors per integrated circuit would double every 18 months Transistors i80x86 M68K MIPS Alpha Gordon Moore (co-founder of Intel)

Contributor 2: Computer Architecture  Exploiting Parallelism (Single processor)  Pipelining  Superscalar  VLIW (Very Long Instruction Word)  Multiprocessor  Media Instructions  Cache Memory 9

Advanced Architectural Features 10  Parallelism in processing  Instruction level parallelism (ILP)  Superscalar  Out of order execution  Branch prediction  VLIW (software approach)  Data level parallelism (DLP) & Task level parallelism (TLP)  SIMD instructions (media processing)  Multicore (multi-processor)  Latency and capacity in memory system  Low latency access using cache memory  Capacity increase in main memory

Superscalar  Multiple functional units  Multiple integer units  Multiple floating point units 11 ALPHA Pentium

How do computers work?  Need to understand abstractions such as:  Applications software  Systems software  Assembly Language  Machine Language  Architectural Issues: i.e., Caches, Virtual Memory, Pipelining  Sequential logic, finite state machines  Combinational logic, arithmetic circuits  Boolean logic, 1s and 0s  Transistors used to build logic gates (CMOS)  Semiconductors/Silicon used to build transistors  Properties of atoms, electrons, and quantum dynamics  So much to learn! 12

Levels of Abstraction  Delving into the depths reveals more information about machines  An abstraction omits unneeded detail, helps us cope with complexity 13 High level language program (in C) Assembly language program (for MIPS) Binary machine language program (for MIPS) swap (int v[], int k) { int temp; temp = v[k]; v[k] = v[k+1]; v[k+1] = temp; } swap (int v[], int k) { int temp; temp = v[k]; v[k] = v[k+1]; v[k+1] = temp; } swap: mull $2, $5, 4 add $2, $4, $2 lw $15, 0($2) lw $16, 4($2) sw $16, 0($2) sw $15, 4($2) jr $31 swap: mull $2, $5, 4 add $2, $4, $2 lw $15, 0($2) lw $16, 4($2) sw $16, 0($2) sw $15, 4($2) jr $ compiler assembler

Instruction Set Architecture (ISA)  A very important abstraction  Interface between hardware and low-level software  Standardizes instructions, machine language bit patterns, etc.  Advantage: different implementations of the same architecture  Disadvantage: sometimes prevents using new innovations  Design of instruction set  How to specify data location  Which instructions to include  Which data formats to support  How to encode instructions  Modern instruction set architectures:  IA-32, PowerPC, MIPS, SPARC, ARM, and others 14

 ENIAC built in World War II  The first general purpose computer  Used for computing artillery firing tables  80 feet long by 8.5 feet high and several feet wide  Each of the twenty 10 digit registers was 2 feet long  Used 18,000 vacuum tubes  Performed 1900 additions per second Moore’s Law: Transistor capacity doubles every months 15 Historical Perspective

Before ENIAC 16

Stored Program Computers  Instructions and data stored as binary numbers in memory  An instruction/data is referenced by its address  Advent of EDVAC by John von Neumann 17

Electronic Computers 2 nd Generation  Technologies  Processor: transistors  Memory: magnetic cores  General purposes  IBM System/360  Same architecture for a wide range of computers  Digital Equipment PDP-8  Supercomputer  Control Data

Electronic Computers 3 rd Generation  Technologies  Processor: IC  Memory: cores, SRAM and DRAM  IBM S/370  DEC PDP-11, VAX 11  CDC 7600  Cray-1 19

Electronic Computers 4 th Generation  Technologies  Processor: VLSI  Memory: SRAM and DRAM  IBM 3990, 4380  DEC VAX 8400  Vector supercomputers  Cray-2, Cray X-MP  Fujitsu, Hitachi, NEC 20

Electronic Computers 5 th Generation  Technologies  VLSI, SRAM, and DRAM with design tools  Read “Singularity is coming”  RISC processor  MIPS  PA-RISC  SPARC  Alpha  PowerPC  CISC processor  Intel Pentium  AMD 21

Lessons from Computer History  A new technology invents a new market  IBM S/360 triggers business applications  High density VLSI enables personal mobility  Architecture is resurrected  Simple one in ‘60 because of technology limit  Complex one in ‘80 for servicing many people  Simple one for mobility and low power  Now?  Mass market calls for standardization  Niche market is profitable but vulnerable to new technology  Cray, Apple, Sun, SGI 22