CDA 3101 Spring 2020 Introduction to Computer Organization

CDA 3101 Spring 2020 Introduction to Computer Organization Technology Trends Digital Logic 101 07 -09 Jan 2016 Mark Schmalz http: //www. cise. ufl. edu/~mssz/Comp. Org/Top-Level. html

Review (Last Class) • • Five components of the computer Principle of Abstraction to build systems as layers Pliable Data: a program determines what it is Stored program concept: instructions are just data Principle of Locality: memory hierarchy Greater performance by exploiting parallelism Compilation vs. interpretation Principles/Pitfalls of Performance Measurement

Overview (Today’s Class) • Computer generations • Technology applications synergism • Technology trends – Hardware – Software • Moore’s law • Basics of Digital Logic – Operations – Truth Tables

Computer Generations • Gen-0: Mechanical computers (BC to early 1940 s) • Gen-1: Vacuum Tubes (1943 -1959) • Gen-2: Transistors (1960 -1968) – John Bardeen, Walter Brattain, and William Shockley • Gen-3: Integrated Circuits (1969 -1977) – Jack Kilby (1958) • Gen-4: VLSI (1978 -present) • Gen-5: Optical? Quantum?

Digital Computer Milestones 1800 s Analytical Engine Babbage First digital computer 1936 Z 1 Zuse First relay machine 1943 COLOSSUS British gov’t First electronic computer 1944 Mark I Aiken First general-purpose computer 1946 ENIAC I Eckert/Mauchley Modern computer history starts 1949 EDSAC Wilkes First stored-program computer 1952 IAS Von Neumann Most computers use this design 1960 PDP-1 DEC First minicomputer 1964 360 IBM Computer family, architecture 1964 6600 CDC First scientific supercomputer 1974 8080 Intel First processor on a chip 1974 CRAY-1 Cray First vector supercomputer 1981 IBM PC IBM Personal computer era 1985 MIPS First commercial RISC machine 1990 RS 6000 IBM First superscalar microprocessor 2008 Blue Gene IBM Most powerful computer

Technology Trends • Technology application synergism (virtuous circle) – Intel’s nightmare: Fast CPUs, lack of application demands – Current application demands • • E-commerce servers Database servers Engineering workstations Ubiquitous mobile computing • Technologies – Compilers – Silicon ISA and computer organization • Silicon Valley or Iron Oxide Valley ? ?

IC Manufacturing Cost = f(area 4)

Hardware Technology Trends • Processor – 2 X in speed every 1. 5 years 100 X performance in last decade • Memory – DRAM capacity: 2 x / 2 years; 64 X size in last decade – Cost per bit: improves about 25% per year • Disk – capacity: > 2 X in size every 1. 0 years – Cost per bit: improves about 100% per year – 120 X size in last decade • New units! Mega (106) Giga (109) Tera (1012)

Memory Capacity (1970 -2000) Size (bits) Year Size(Mbit) 10000000 1000000 10000 1970 1975 1980 1985 Year 1990 1995 2000 1970 1983 1986 1989 1992 1996 2000 2010 2020 0. 001 (est. ) 0. 0625 0. 25 1 4 16 64 256 2 K 256 K = 32 GByte

Trend: Moore’s “Law” (1971 -2017) Moore’s Law : Number of Transistors on a Chip Doubles Every 2 Years

Reality: Moore’s “Law” (1971 -2018) Moore’s Law : Number of Transistors on a Chip Doubles Every 2 Years

Historical Memory Capacity 1980 -2010

Recent Memory Capacity Flash Memory 2013 -2020

Processor Capacity (1970 -2000) Moore’s Law (1965): 2 X transistors/Chip Every 1. 5 years All processors 10000 Alpha 21264: 15 million Pentium Pro: 5. 5 million Power. PC 620: 6. 9 million Alpha 21164: 9. 3 million Sparc Ultra: 5. 2 million 10000000 Transistors Moore’s Law Pentium i 80486 1000000 i 80386 i 80286 100000 i 8086 10000 i 8080 i 4004 1000 1975 1980 1985 Year 1990 1995 2000 After late 1990 s, spatial parallelism (multiple processors on chip) changed the quasi- linear appearance of this

Historical Capacity (1950 -2010) Moore’s Law (1965): 2 X transistors/Chip Every 2 years

Processor Performance (1970 -2019) Multico re Revoluti on

Historic Intel CPUs Pentium III – 800 MHz, 4 GB Memory Pentium 4 – 2+GHz, 4 GB Memory Itanium – 4+ GHz, > 4 GB Memory

Intel Processor Chip Layout 1990 s Pentium Pro • 306 mm 2 • 5. 5 M transistors Itanium (EPIC/IA-64) • ILP: 20 instructions • Compiler support • Massive hardware resources • 2 Floating Point Units • 4 Integer Units • 3 Branch Units • Internet Streaming SIMD • 128 FP registers • 128 integer registers

Intel Processors 2015 -2019 Intel Xeon Phi • Over 50 cores • > 100 M transistors Intel Ice Lake Hundreds of Cores Billions of transistors

Physical Limits on Moore’s Law • Limits imposed by insulator thickness (0. 2 nm) • Quantum tunneling effects => crosstalk • How much smaller? (0. 02 / 0. 2 nm = 100 x) • How much faster? Speed = k x Area -- 3 to 4 orders of magnitude faster (103 - 104) -- 3. 3 GHz => 5 THz to 10 THz effective • When? (~10 years from now…)

Physical Limits on Moore’s Law (Frank, 2002)

Will the Computer World End? • No, but things will get more interesting… • Opportunities -- Make faster processors, algorithms using current technology -- Increase bandwidth of buses that supply data to processors -- Exploit spatial parallelism (GPUs)

Solutions (? ) for Moore’s Law • Quantum Computing -- Different paradigm – all results at once -- How to find “correct” result? -- Implementation: Optics? Silicon? ? • Highly Experimental Technologies -- DNA Computing (Pattern Matching) -- Reversible Computing (Low Power) -- Compressive Computation ( FAST )

Tech Summary • Incredible improvements in processor, memory and communication • Technology application synergism • Technologies – Compiler – Silicon • Computer organization takes advantage of technology advances • Will Moore’s law last forever? /

New Topic – Digital Logic 101 • Digital logic – its place in CDA 3101 • Boolean Operations • Transistors and Digital Logic • Basic gates – and, or, not -- Transistor implementations -- Truth tables

Digital Logic in CDA 3101 Application (Browser) Software Compiler Assembler Operating System (Win, Linux) CDA 3101 Instruction Set Architecture Datapath & Control Hardware Memory Digital Logic Circuit Design Transistors I/O System

Boolean Operations • 0 & 1: the only values for variables and functions B = {0, 1} called Boolean numbers • The NOT function: f (A) = • Truth tables 1 if A is 0 0 if A is 1 – Completely define a Boolean function – n variables => 2 n entries in the truth table – Up to 16 Boolean functions of two variables – Shorthand: specify only entries with nonzero outputs

Transistors & Digital Logic Gate Symbol Truth Table (functional behavior) NOT gate (Inverter)

NAND Gate

NOR Gate

AND & OR Gates

Integrated Circuits

Conclusions • Technology development • Computer organization takes advantage of technology advances • Digital Logic & Boolean Numbers • Basic logic gates w/ Implementation • Concept of truth table • Next lecture: Boolean Algebra Complex logic & circuits