M1 Computers and Data

Transcription

1 M1 Computers and Data

2 Module Outline Architecture vs. Organization. Computer system and its submodules. Concept of frequency. Processor performance equation. Representation of information characters, signed and unsigned integers. IEEE 754 floating point standard.

3 Basic Computer Organization Processor Executes programs Main Memory Holds program and data I/O For communication and data Processor (CPU) ALU CONTROL REGISTERS MEMORY Interconnection Network I/O I/O I/O I/O

4 Inside the Processor Control Hardware: Hardware to manage instruction execution ALU: Arithmetic and Logical Unit (hardware to do arithmetic and logic operations) Registers: Small units of memory to hold data/instructions temporarily during execution Memory: Stores information being processed by the CPU Input: Allows the user to supply information to the computer Output: Allows the user to receive information from the computer

5 Concept of Time and Speed Frequency: Number of occurrences of a repeating event per unit time. SI unit: Hertz (Hz) The period is the duration of one cycle in a repeating event Period = Cycle time Cycle Time = 1 Frequency

6 Clock Cycle Clock is a special signal to hardware A well defined indication for event start and complete. Time (ns) Clock Rising Edge Falling edge Levels

7 Clock Cycle Clock is a special signal to hardware A well defined indication for event start and complete. Time (ns) Clock Addition ADDITION COMPLETES IN THIS DURATION ANOTHER BLOCK Both operands are ready. The result is ready. The result is consumed by another block.

8 Clock Cycle Clock is a special signal to hardware A well defined indication for event start and complete. Time (ns) Clock Memory Access Memory access happens here. ANOTHER BLOCK Address is ready Data is ready. The data is consumed

9 On Processor Performance How is frequency related to performance? Program ExecutionTime = Execution Time per Instruction Total Program Instructions CPU Time =Execution Time per Instruction Instruction Count Execution Time per Instruction= Cycles spent per Instruction Cycle Time CPU Time = IC Cycles per Instruction Cycle Time Example What is is the execution time of of a program containing a million instructions each occupying 4 cycles in in a 2 GHz processor?

10 Iron Law of Processor Performance CPU Time = IC Cycles per Instruction CycleTime Cycle Time= 1 Frequency CPU Time= IC CPI Frequency CPU time= Instructions Program cycles Clock Instruction Seconds Clock cycle

11 Iron Law of Processor Performance CPU time= Instructions cycles Clock Program Instruction Seconds Clock cycle COMPILER ARCHITECTURE AND ORGANIZATION

12 Data Representation

13 Objective How are program data (integers, strings, floats, addresses) represented (encoded) in an object file and during program execution?

14 ASCII 7-bit format Character Data American Standard Code for Information Interchange Numbers 0-9, letters a-z and A-Z, basic punctuation symbols, control codes, blank space. Range 0x0 0x1F 0x20 0x2F 0x3A 0x40 0x5B 0x60 0x7B - 0x7F Encoded Character Special Characters Non alphabet characters 0x30 0x39 0, 1,..., 8, 9 0x41 0x5A 0x61 0x7A A, B,..., Y, Z a, b,..., y, z

15 Character Data Unicode 8/16/32-bit formats (UTF-8, UTF-16, UTF-32) For most of the world's writing systems Contains a repertoire of more than 110,000 characters covering 100 scripts and various symbols All ASCII characters included in UTF-8.

16 Signed and Unsigned int x; unsigned int y; Integer Data Overflow: The number to be represented is larger than the range possible.

17 Signed Integer Data Sign Magnitude Representation Sign bit Magnitude bits x n 1 x n 2... x 2 x 1 x 0 MSB LSB ( 1) x n 1 [ xn 2 2 n x x ] Binary to/from Decimal conversion: (a) 0x4C (b) 27

18 Signed Integer Data Problems Sign bit uses up one bit space. Arithmetic circuits need extra steps to identify sign of the result. +0 and -0 exist.

19 Two's Complement Representation Leading zeros mean positive and leading ones mean negative

20 Two's Complement Representation x =x x... x x x ten n 1 n x = x 2 n 1 +x 2 n 2 n +...+x x 2 0 ten ten n 1 n Example Convert 17, -21, 32 into their 2's complement notation. Observe: Pick an 8-bit, 2's complement binary number. Invert its bits: 0 1 and 1 0. Add both these numbers in binary. What is the result?

21 Two's Complement Representation x x x + x = -1 (x + 1) = -x x x = -4 Binary to/from Decimal conversion: 0xFFFFFBD5

22 Sign Extension and Zero Extension Binary to to Decimal conversion: (a) (b) (c) (d) Replicate the Most Significant Bit to the left. Sign extension 1001, 11001, = = - 7 Zero extension 0111 = 7, = 7, = 7,...

23 Unsigned Integer Data Interpret the 8-bit quantity 0x81 as (a) a signed integer, and (b) an unsigned integer (a) (b) 129 What are the max and min unsigned values of 16bit unsigned data?

24 One's Complement Similar to two's complement. The negative is found by inverting bits. Two zeros: 0000 and In the 8-bit one's complement representation, which are the most negative and most positive numbers?

25 Byte Ordering A How does one read the 4 bytes at address 100? A 0 vs. A is the MSB (at the big end). 100 is the LSB (at the little end) vs Big Endian notation vs. Little Endian notation IBM, Network data Intel processors

26 Biased Representation The most negative value is represented by 0. Two's Complement Biased Notation (Biased by 128) (Excess 128)

27 Real Data (pi), (e) = 1.0 x ,155,760,000 = x 10-9 Scientific form of a real number Single digit before the decimal point 1.0 x 10-5, 0.1 x 10-4, 0.01 x 10-3,... Normalized representation No leading zeros 1.0 x 10-5 Not 0.1 x 10-4, 0.01 x 10-3,...

28 Real Data Binary Notation = ( ) = ( ) = ( ) = ( ) = ( ) = ( ) = ( ) 2

29 Real Data Binary Notation = x 2 0, x 2-1, x 2 1,... Scientific form x 2 2, x 2 3, x 2 4,... Normalized representation x 2 2. Computers use Normalized representation to store floats. 1.f x 2 e f: fraction, e: exponent.

30 Floating Point Representation s exponent fraction ( 1) s 1. f 2 e 127 IEEE 754 single precision format Exponent is biased by 127

31 IEEE 754 FP Representation Convert 0.5 into binary. Represent in IEEE 754 single precision floating point representation. Binary = 0.1 = 1.0 x 2-1. s=?, f =?, e =? s=0, f=0, e = (-1) = x3F000000

32 IEEE 754 FP Representation Convert 0.75 into IEEE 754 SP FP. Convert 0xBDCCCCCC into decimal. Four special values Exponent Fraction Value 0 0 Zero (± 0) Infinity (± ) 0 0 Denormal numbers (± 0.M x ) NaN (0/0 or -1)

33 IEEE 754 Double Precision s exponent fraction ( 1) s 1. f 2 e 1023

34 Real Data Representation Recap What are the place values of 7 positions after the decimal point in a binary real number? n 2 n-1 2 n decimal point Binary to to Decimal conversion: (a) 1.10 (b) (c) (d) (e)

35 Real Data Representation Recap Convert into IEEE 754 SP FP: 13.1, 12.2, 11.3, 10.4, 9.5, 8.6, 7.7, 6.8, 5.9

36 Module Outline Architecture vs. Organization. Computer system and its submodules. Concept of frequency. Processor performance equation. Representation of information characters, signed and unsigned integers. IEEE 754 floating point standard.

37

38 Word size = 32 bits Word Alignment Words can be stored linearly in addresses 0x00, 0x04, 0x08, 0x0C, 0x10,... These addresses are 'word boundaries' C A A B A 0 D 0 0 E 4 7 A A 0 D 0 0 E Word Aligned Word Not Aligned