Embed Size (px)
DESCRIPTION
Duke Computer Architecture
Compsci 250 / ECE 250: “C Programming”
Alvin R. Lebeck Some slides based on those from Daniel Sorin,
Andrew Hilton, Amir Roth, Gershon Kedem
2 © Alvin R. Lebeck From Sorin, Hilton, Roth
Administrivia
• Homework #1 • No class Monday
• Today ! Finish Data Representations ! C programming & Memory
• Updated Academic Honesty Policy ! Do not provide solutions to any archive (digital, paper, etc.) ! This is a simple extension of existing policy of not providing
solutions to another student.
Compsci / ECE 250
3 © Alvin R. Lebeck From Sorin, Hilton, Roth Compsci / ECE 250
Review: Answer
What floating-point number is 0xC1580000? 1100 0001 0101 1000 0000 0000 0000 0000
1 1000 0010 101 1000 0000 0000 0000 0000 0 22 23 30 31
S N M
• Sign = 1 means this is a negative number • Exponent = (128+2)-127 = 3 • Mantissa = 1.1011 • -1.1011x23 = -1101.1 = -13.5
4 © Alvin R. Lebeck From Sorin, Hilton, Roth
Special values in IEEE Floating Point
• How do you represent 0.0 in IEEE Floating Point? ! 0.0 = 000000000 ! But need 1.XXXXX representation?
• Exponent = 0 is denormalized ! Implicit 0. instead of 1. in mantissa ! Allows 0000….0000 to be 0 ! Helps with very small numbers near 0
• Results in +/- 0 in FP (but they are �equal�)
Compsci / ECE 250
5 © Alvin R. Lebeck From Sorin, Hilton, Roth
Other Weird FP numbers
• Exponent = 1111 1111 also not standard ! If all 0 mantissa: +/- ∞
1/0 = +∞ -1/0 = -∞
! If non zero mantissa: Not a Number (NaN)
sqrt(-42) = NaN
Compsci / ECE 250
6 © Alvin R. Lebeck From Sorin, Hilton, Roth
FP Addition/Substraction
• Addition/Subtraction ! Align the exponents ! Perform operation ! Renormalize the result (put into 1.M x 2E-127 form)
• What can possibly go wrong? • Well a lot • Search for disasters caused by numerical errors
Compsci / ECE 250
7 © Alvin R. Lebeck From Sorin, Hilton, Roth
Software Implication! Rounding Errors
• We only have 32-bits to represent floats ! Must approximate some values
• Limited bits for mantissa • Does (x+y)*z = (x*z+y*z)? • Mathematically yes, but assumes infinite precision • Assume base 10, four digits available (two to left, two to
right of decimal point) ! x = 99.96 x 103
! x = x + 0.07 ! x = 100.03 x 103 ! x = 10.00 x 104
• Numerical Analysis (CS 220) studies some of these issues
Compsci / ECE 250
8 © Alvin R. Lebeck From Sorin, Hilton, Roth Compsci / ECE 250
Floating Point Representation
• Double Precision Floating point: 64-bit representation: ! 1-bit sign ! 11-bit (biased) exponent ! 52-bit fraction (with implicit 1).
• �double� in Java, C, C++, …
1 11-bit 52 - bit Exp S Mantissa
9 © Alvin R. Lebeck From Sorin, Hilton, Roth Compsci / ECE 250
ASCII Character Representation 000 nul 001 soh 002 stx 003 etx 004 eot 005 enq 006 ack 007 bel 010 bs 011 ht 012 nl 013 vt 014 np 015 cr 016 so 017 si 020 dle 021 dc1 022 dc2 023 dc3 024 dc4 025 nak 026 syn 027 etb 030 can 031 em 032 sub 033 esc 034 fs 035 gs 036 rs 037 us 040 sp 041 ! 042 " 043 # 044 $ 045 % 046 & 047 ' 050 ( 051 ) 052 * 053 + 054 , 055 - 056 . 057 / 060 0 061 1 062 2 063 3 064 4 065 5 066 6 067 7 070 8 071 9 072 : 073 ; 074 < 075 = 076 > 077 ? 100 @ 101 A 102 B 103 C 104 D 105 E 106 F 107 G 110 H 111 I 112 J 113 K 114 L 115 M 116 N 117 O 120 P 121 Q 122 R 123 S 124 T 125 U 126 V 127 W 130 X 131 Y 132 Z 133 [ 134 \ 135 ] 136 ^ 137 _ 140 ` 141 a 142 b 143 c 144 d 145 e 146 f 147 g 150 h 151 i 250 j 153 k 154 l 155 m 156 n 157 o 160 p 161 q 162 r 163 s 164 t 165 u 166 v 167 w 170 x 171 y 172 z 173 { 174 | 175 } 176 ~ 177 del
Oct. Char
• Each character represented by 7-bit ASCII code (packed into 8-bits) • Convert upper to lower case ‘A’ + 3210 = ‘a’ • Other tables with decimal and HEX values.
10 © Alvin R. Lebeck From Sorin, Hilton, Roth
Unicode
• Many types • UTF-8: variable length encoding backward compatible
with ASCII ! Linux
• UTF-16: variable length ! Windows, Java
• UTF-32: fixed length
Compsci / ECE 250
11 © Alvin R. Lebeck From Sorin, Hilton, Roth
Data Representation Summary
• Need efficient representations • 2’s complement for signed integers • IEEE Floating Point for Real numbers • Bits are bits are bits, same bits can be interpreted as
different types (unsigned, 2’s complement, float, etc.) • Next up: C programming & memory
Compsci / ECE 250
12 © Alvin R. Lebeck From Sorin, Hilton, Roth
What you know today
Compsci / ECE 250
• JAVA ... System.out.println("Please Enter In Your First Name: "); String firstName = bufRead.readLine(); System.out.println("Please Enter In The Year You Were Born: "); String bornYear = bufRead.readLine(); System.out.println("Please Enter In The Current Year: "); String thisYear = bufRead.readLine(); int bYear = Integer.parseInt(bornYear); int tYear = Integer.parseInt(thisYear); int age = tYear – bYear ; System.out.println("Hello " + firstName + ". You are " + age + " years old");
13 © Alvin R. Lebeck From Sorin, Hilton, Roth
How does a Java program execute?
• Compile Java Source to Java Byte codes • Java Virtual Machine (JVM) interprets/translates Byte
codes • JVM is a program executing on the hardware… • Java has lots of things that make it easier to program
without making mistakes • JVM handles memory for you
! What do you do when you remove an entry from a hash table, binary tree, etc.?
Compsci / ECE 250
14 © Alvin R. Lebeck From Sorin, Hilton, Roth
The C Programming Language
• No virtual machine ! No dynamic type checking, array bounds, garbage collection, etc. ! Compile source file directly to machine (this is a little simplified)
• Closer to hardware ! Easier to make mistakes ! Can often result in faster code
• Generally used for ‘systems programming’ ! operating systems, embedded systems, database implementation ! There is object oriented C++ (C is a strict subset of C++)
Compsci / ECE 250
15 © Alvin R. Lebeck From Sorin, Hilton, Roth
The C Programming Language (Continued)
• No objects • Procedural, not object oriented
! No objects with methods
• Structures, unions like objects ! Member variables (no methods)
• Pointers – memory, arrays • External standard library – I/O, other facilities • Macro preprocessor (#<directive>) • Additional Resources
! Kernighan & Richie book The C Programming Language ! MIT open course Practical Programming in C link in ‘docs’ of web
site ! Drew Hilton Video Snippets
Compsci / ECE 250
16 © Alvin R. Lebeck From Sorin, Hilton, Roth
Compiling and Running a C Program
• Use the gcc program to create an executable file • gcc –o <outputname> <source file name> • gcc –o hello hello.c (must be in same directory as hello.c) • If no –o option, then default output name is a.out (e.g., gcc hello.c) • Type the program name on the command line
! ./ before “hello” means look in current directory for hello program
Compsci / ECE 250
17 © Alvin R. Lebeck From Sorin, Hilton, Roth
Debugging
• Print debugging… ! Just output information at different points in the program ! Not the most efficient, but often works.
• gdb <executable filename> • Good for stopping at set points in program and inspecting
variable values. ! If you get good at using a debugger it is easier/better than printf
debugging… ! See GDB Essentials video off https://www.cs.duke.edu/courses/
spring15/compsci250/docs.html
Compsci / ECE 250
18 © Alvin R. Lebeck From Sorin, Hilton, Roth
Variables, operators, expressions
• C variables are similar to Java ! Data types: int, float, double, char, void ! Signed and unsigned int ! char, short, int, long, long long can all be integer types
» These specify how many bits to represent an integer
• Constants ! Use #define C preprocessor ! E.g.,: #define MAX_SCORE 100
• Operators: ! Mathematical +, -, *, /, %, ! Logical !, &&, ||, ==, !=, <, >, <=, >= ! Bitwise &, |, ~, ^ , <<, >> (we’ll get to these next)
• Expressions: var1 = var2 + var3;
Compsci / ECE 250
19 © Alvin R. Lebeck From Sorin, Hilton, Roth Compsci / ECE 250
Bitwise AND / OR / XOR
• & operator performs bitwise AND • | operator performs bitwise OR • ^ operator performs bitwise Exclusive OR (XOR) • ~ operator performs bitwise NOT • Per bit
0 & 0 = 0 0 | 0 = 0 0 ^ 0 = 0 0 & 1 = 0 0 | 1 = 1 0 ^ 1 = 1 1 & 0 = 0 1 | 0 = 1 1 ^ 0 = 1 1 & 1 = 1 1 | 1 = 1 1 ^ 1 = 0
• For multiple bits, apply operation to individual bits in same position
011010 101110 001010
011010 101110 111110
AND OR 011010 101110 110100
XOR
20 © Alvin R. Lebeck From Sorin, Hilton, Roth Compsci / ECE 250
The SHIFT operator
• >> is shift right, << is shift left, operands are unsigned int and number of positions to shift
• (1 << 3) is …000001 -> …0001000 (it’s 23) • 0xff00 is 0xff << 8, and 0xff is 0xff00 >> 8
! Unsigned ints, be careful shifting signed integers (more later this semester)
• 0xAB & 0x0F = 0x0B • 0xAB & 0xF0 = 0xA0
• 0x0F and 0xF0 are called a bit mask
21 © Alvin R. Lebeck From Sorin, Hilton, Roth Compsci / ECE 250
Sakai In Class Quiz 2.1
• 32-bit word with x,y and button fields ! bits 0-7 contain x position ! bits 8-15 contain y position ! bits 16-17 contain button (0 = left, 1 = middle, 2 = right) ! Bits 18-31 don’t matter
• Which C Statements correctly set xpos, ypos & button?
button y x val = 01 1010 0011 0100 1100
A. xpos = val && 0x000ff; ypos = (val && 0x0ff00) >> 8; button = (val && 0x30000) >> 16;
B. xpos = val & 0x000ff; ypos = (val & 0x0ff00) >> 8; button = (val & 0x30000) >> 16;
C. xpos = val & 0x000ff; ypos = (val & 0x0ff00) << 8; button = (val & 0x30000) << 16;
D. xpos = val & 0x000ff; ypos = val & 0x0ff00 >> 1; button = val & 0x30000 >> 2;
22 © Alvin R. Lebeck From Sorin, Hilton, Roth Compsci / ECE 250
Sakai In Class Quiz 2.1: Answer
• 32-bit word with x,y and button fields ! bits 0-7 contain x position ! bits 8-15 contain y position ! bits 16-17 contain button (0 = left, 1 = middle, 2 = right) ! Bits 18-31 don’t matter
• Which C Statements correctly set xpos, ypos & button?
button y x val = 01 1010 0011 0100 1100
A. xpos = val && 0x000ff; ypos = (val && 0x0ff00) >> 8; button = (val && 0x30000) >> 16;
Logical AND
B. xpos = val & 0x000ff; ypos = (val & 0x0ff00) >> 8; button = (val & 0x30000) >> 16;
C. xpos = val & 0x000ff; ypos = (val & 0x0ff00) << 8; button = (val & 0x30000) << 16; Shift wrong direction; value not in least significant bits
D. xpos = val & 0x000ff; ypos = val & 0x0ff00 >> 1; button = val & 0x30000 >> 2; Precedence…shifts mask first; shift amount wrong
23 © Alvin R. Lebeck From Sorin, Hilton, Roth
Functions
• Encapsulate computation ! Reuse or clarity of code ! Cannot define functions within functions
• Must be declared before use! int div2(int x,int y); /* declaration here */ main() {
int a; a = div2(10,2);
} int div2(int x, int y) { /* implementation here */
return (x/y); }
• Or put functions at top of file (doesn’t always work)
Compsci / ECE 250
24 © Alvin R. Lebeck From Sorin, Hilton, Roth
Global Variables
• Global variables: Accessible from any function #include <stdio.h> float f = 0; void bar () { f = 0.5; } int main() { f =0.31234; bar(); printf(“The value is %f \n”, f); } Output is: The value is 0.5
Compsci / ECE 250
25 © Alvin R. Lebeck From Sorin, Hilton, Roth
Memory Partitions
• Text for instructions ! add res, src1, src2 ! mem[res] = mem[src1] +
mem[src2]
• Data ! static (constants, global variables) ! dynamic (heap, malloc / new
allocated) ! grows up
• Stack ! local variables ! grows down
• Variables are names for memory locations ! int x;
Stack
Data
Text Reserved 0
2n-1
Typical Address Space
Heap
Compsci / ECE 250
26 © Alvin R. Lebeck From Sorin, Hilton, Roth
A Simple Program’s Memory Layout
... int result; // global var int main() { int x; ... result = x + result; ...
} For now, think of this as performing: mem[0x208] = mem[0x400] + mem[0x208]
Stack
Text
Data
Reserved 0
2n-1
Heap
x 0x400 5
result 0x208 3
Compsci / ECE 250
27 © Alvin R. Lebeck From Sorin, Hilton, Roth
Sakai In Class Quiz 2.2
#include <stdio.h> int a = 0; float f = 0.5; void setvals() { float f = 0.5; a = 78; } int main() { int a = 23; f =0.31234; setvals(); printf(“The values are %d, %f \n”,a,f); }
• What is the result? a) a = 78, f = 0.5 b) a = 78, f = 0.31234 c) a = 23, f = 0.5 d) a = 23, f = 0.31234
Compsci / ECE 250
28 © Alvin R. Lebeck From Sorin, Hilton, Roth
Sakai In Class Quiz 2.2
#include <stdio.h> int a = 0; float f = 0.5; void setvals() { float f = 0.5; a = 78; } int main() { int a = 23; f =0.31234; setvals(); printf(“The values are %d, %f \n”,a,f); }
Compsci / ECE 250
Stack
Text
Data
Reserved 0
2n-1
Heap
a 0xF0020 f 0x F0000
f 0x20004 a 0x 20000
29 © Alvin R. Lebeck From Sorin, Hilton, Roth
Reference vs. Pointer
Java ! “The value of a reference type variable, in contrast to that of a
primitive type, is a reference to (an address of) the value or set of values represented by the variable” http://java.sun.com/docs/books/tutorial/java/nutsandbolts/datatypes.html
! Cannot manipulate the value of the reference
C ! A pointer is a variable that contains the location of another
variable ! A pointer is a memory location that contains the address of
another memory location ! Can manipulate the value of pointer (double edge sword)
Compsci / ECE 250
30 © Alvin R. Lebeck From Sorin, Hilton, Roth
Pointers
• Declaration of pointer variables ! int *x_ptr; char *c_ptr; void *ptr;
• How do we get the location (address) of a variable? Use one of the following: 1. Use the & ‘address of’ operator
! x_ptr = &intvar; 2. From another pointer (yes we can do arithmetic on them)
! x_ptr = y_ptr + 18; 3. Return from memory allocator
! x_ptr = (int *) malloc(sizeof(int)); ! Similar to java new, but must explicitly free memory (i.e., free(x_ptr));
Compsci / ECE 250
31 © Alvin R. Lebeck From Sorin, Hilton, Roth Compsci / ECE 250
Pointers
• �address of� operator & ! don�t confuse with bitwise AND operator
Given int x; int* p; // p points to an int p = &x;
Then *p = 2; and x = 2; produce the same result
Note: p is a pointer, *p is an int
• What happens for p = 2?;
0x26cf0
x 0x26cf0
p 0x26d00 ... On 32-bit machine, p is 32-bits
32 © Alvin R. Lebeck From Sorin, Hilton, Roth
Pointers
• De-reference using *ptr to get what is pointed at
Compsci / ECE 250
statement x x_ptr int x; ?? ?? int *x_ptr; ?? ?? x = 2 2 ?? x_ptr = &x; 2 &x *x_ptr = 68; 68 &x x_ptr = 200; 68 200 *x_ptr = 42 68 200
• Be careful with assignment to a pointer variable ! You can make it point anywhere…can be very bad ! You will, this semester, likely experience a “segmentation fault” ! What is 200?
33 © Alvin R. Lebeck From Sorin, Hilton, Roth
Sakai In Class Quiz 2.3 Which Version Correctly Swaps Values?
void swap (int x, int y){
int temp = x;
x = y;
y = temp;
}
main() {
int a = 3;
int b = 4;
swap(a, b);
printf(“a = %d, b= %d\n”, a, b);
}
A.
void swap (int *x, int *y){
int temp = *x;
*x = *y;
*y = temp;
}
main() {
int a = 3;
int b = 4;
swap(&a, &b);
printf(“a = %d, b= %d\n”, a, b);
}
B.
Compsci / ECE 250
34 © Alvin R. Lebeck From Sorin, Hilton, Roth Compsci / ECE 250
Strings as Arrays
• A string is an array of characters with �\0� at the end • Each element is one byte, ASCII code • �\0� is null (ASCII code 0) • char str1[256] is similar to str2 = (char *) malloc(256);
0 1 43 15 s t �\0� r i g
16 42
35 © Alvin R. Lebeck From Sorin, Hilton, Roth
Structures
• Loosely like objects ! Have member variables ! Do not have methods!
• Structure definition with struct keyword struct student_record {
int id; float grade;
} rec1, rec2;
• Declare a variable of the structure type with struct keyword struct student_record onerec, *tworec;
• Access the structure member fields with ! ‘.’ structvar.member (e.g., onerec.id = 12;) ! ‘->’ structprt->member (e.g., tworec->id;)
Compsci / ECE 250
36 © Alvin R. Lebeck From Sorin, Hilton, Roth
Summary
• C is different from Java ! Procedural ! Structures not objects… ! Pointers!
• Next time ! Explicit memory deallocation ! Pointer manipulation ! Linked Structures
Compsci / ECE 250
