System Overview
Download
Report
Transcript System Overview
CSE 5311
Fundamentals of
Computer
Science
Computer System Overview
Computer Systems
Computer systems have two major aspects - hardware
and software.
Hardware - The parts of the computer you can actually
see and touch.
The hardware (in a programmable computer) is
directed by the software to carry out tasks.
Software - Instructions and data stored in memory (a
type of hardware) in the form of magnetic or electrical
impressions.
The instructions are used to control the hardware.
Hardware can be divided into three groups:
Processing
• Central Processing Unit (CPU)
– Arithmetic Logic Unit (ALU)
– Control Unit (CU)
– Floating-Point Unit (FU)
– Memory Manager (MM)
• Direct Memory Access (DMA) controller
• Various other device controllers.
Input/Output Devices
•
•
•
•
•
Keyboard
Mouse
CRT
Printer
Disk Drive
Hardware
Hardware can be divided into
three groups (continued):
Memory
• Random Access Memory (RAM)
• Read Only Memory (ROM)
• Disks
– Floppy
– Hard
– CD
– DVD
• Tape
• Registers
• Cache (associative memory)
Hardware
Computer-System Architecture
Memory
Memory
Memory can be viewed in a hierarchical structure:
REGISTERS
volatile
CACHE
RAM
non-volatile
DISKS
Memory
A program must exist in memory in order to
execute.
When a program is executed, it begins a
process of migration from disk memory to
RAM, cache, and register memory.
Depending on a programs size, as it executes
parts of it may move back and forth between
disk memory and register memory many times.
The operating system is chiefly responsible
for data migration.
32-Bit Intel Register
Set
There is also a set of floating point registers
Disk Surface Physical Organization
Every disk is
organized into
tracks, which in
turn are divided
into sectors
Within a sector
the bit is the
smallest data unit
Sectors are
logically grouped
by the OS into
clusters
Logical Disk
Organization
Every formatted disk contains one boot record,
a root directory, and two file allocation tables.
The remainder of the disk space is “data area”
Physical Organization of Hard Drive
Each disk surface has a head which is fixed to the end
of an access arm. All arms and heads are moved
simultaneously by a single actuator.
Disk Drive Actuators
The actuators is responsible for the movement
of the read/write heads.
Physical Organization
of Disks
On hard drives the read/write “floats” above the
disk surface. Disk constantly rotates.
On floppy drives the read/write head touches the
disk surface. Disk only rotates when a read or write
is in process.
Software
Software can be divided into two groups:
Systems programs
• Operating system
• Program translators
– assemblers
– compilers
– interpreters
• Linkers
• Loaders
• Device drivers (may be considered as part of the OS)
• Note: Some persons consider operating systems to be the
only true system program. Everything else is considered
an application program.
Software
Software can be divided into two groups:
Application programs
•
•
•
•
•
payroll programs
games
spreadsheets
tax preparation software
text editors
Operating Systems
The job of the operating system is to manage system
resources by maximizing the throughput of the
computer system.
This is accomplished by keeping all hardware
resources as fully occupied as possible.
The operating system provides a buffer between
human users and the hardware and other programs
and the hardware.
User
OS
Programs
HW
OS in Relation to Other System Components
Program Translators
Compilers
Assemblers
Interpreters
The Translation
Process
Lexical
Analysis
source program
Parsing
Code
Generation
object program
All translators follow this general process. However, for
high-level languages the process is much more complex
than for low-level languages.
Compilers
Compilers convert programs written in a high-level language
into machine language.
Machine language is used to communicate with the CPU.
Examples of high-level languages that depend on
compilers: C, C++, PL/I, COBOL, FORTRAN, Pascal.
High-level languages have a one-to-many translation ratio.
One high-level instruction is translated by the compiler
into one or more machine instructions.
Mov b to register ax
a = b + c;
Add c to register ax
Compiler
Mov register ax to a
High-level Source
Code
Compiler
Object Library
Object File
Linker
Object Library
Executable File
The Compilation and
Linking Process
Compilers
All programming languages must be formally defined.
A language’s formal definition is known as its grammar.
The grammar is used to indicate is a sequence of source
code symbols (tokens) are syntactically correct.
There are different notations available for writing
grammars
The parser is the part of the compiler that, based on the
grammar, determines if the program is syntactically
correct.
Compilers
Grammar for a simplified version of Pascal expressed using
Backus-Naur Form (BNF) notation.
• Upper-case tokens represent reserve words.
• ‘id’ represents identifiers
• ‘int’ represents an integer constant
Compilers
Pascal program that conforms to the proceeding grammar.
Compilers
Lexical analysis is the
process of breaking the
source code into tokens
Tokens may optionally be
assigned a numeric code
The scanner is the part of the
compiler that performs lexical
analysis
Tokens requested by the
parser one at a time
Compilers
assign ::= id := <exp>
Recursive-descent parse of an assignment statement
TOKEN represents the most recent token returned by the scanner
‘advance to next token’ represents a call to the scanner to provide the
next token in the source file input stream
Compilers
exp ::= <term> { + <term> | - <term> }
Recursive-descent parse of an expression statement
Assemblers
Assemblers convert programs written in a low-level
language into machine language.
Examples of low-level languages that depend on
assemblers: Intel 8086-based assembler, Motorola
6800-based assembler, IBM System 360/70 assembler.
Low-level languages have a one-to-one translation ratio.
One assembly instruction is translated by the assembler
into one machine instruction.
mov ax,b
Mov b to register ax
add ax,c
Add c to register ax
mov a,ax
Assembler
Mov register ax to a
Machine Code
Object files and executable files are mostly composed of
machine instructions. Compilers and Assemblers produce
machine code; Interpreters produce p-code (byte code).
The first byte of a machine instruction is known as the
opcode. The opcode indicates what type of operation (add,
subtract, etc.) the instruction is to carry out.
The length of the machine instruction varies according to
the number of operands and addressing modes used.
Mod R/M
opcode
7
0
mod
7 6
reg
r/m
543 210
immed-low
7 -- 0
immed-high
7 -- 0
disp-low
7 -- 0
disp-high
7 -- 0
(The opcode indicates whether or not the immediate v alue f ield is present, as well as its size.)
Intel Instruction Format (8086/8088)
Low-level Source
Code
Assembler
Object Library
Object File
Linker
Object Library
Executable File
The Assembly and
Linking Process
Interpreters
Interpreters “execute” programs written in a high-level
language without converting it into machine language.
Examples of languages that depend on interpreters:
Quick BASIC, Java, Lisp, Smalltalk, Visual Basic.
Interpreters do not product object files containing
machine code, but sometimes produce pseudo
code.
Interpreters allow for features such as automatic
garbage collection
Assign (temp, Add (b,c))
a = b + c;
Interpreter
Assign (a, temp)
Interpreters
The interpreter provides a virtual run-time environment
A Java interpreter is called a “virtual machine”
software
p-code
interpreter
operating system
hardware
High-level Source
Code
Interpreter
(translation
mode)
P-code Library
P-code Library
The Interpretation
Process
P-code File
Interpreter
(execution
mode)
Basic Hardware
Architecture
Computer-System Architecture
Basic Architectural Components
The central processor communicates with main
memory and I/O devices via buses
Data bus for transferring data
Address bus for the address of a memory location
or an I/O port
Control bus for control signals (Interrupt request,
memory read/write …)
Each operation must be synchronized by the system
clock
Registers are high-speed storage within the
processor
Simplified CPU Design
Central Processing Unit
Typically a general purpose CPU will
contain:
Arithmetic Logic Unit
Floating Point Unit
• Similar to ALU but works with floats
Register File
• Contains all registers (not a true file)
Memory Management Unit
• Performs logical to physical address
translation
Control Unit
• Tells the ALU and FPU which operations to
carry out.
The Fetch-Decode-Execute Cycle
Is the basic cycle for instruction execution
Fetch the next instruction
• place it in queue
• update program counter register
Decode the instruction
• perform address translation
• fetch operands from memory
Execute the instruction
• store result in memory or registers
• set status flags according to result
Before fetching next instruction, the CPU checks if an
interrupt is pending (more on that later)
Microcode
The circuits in most general-purpose processors
(Intel, Motorola, IBM) are directly controlled by
software known as microcode (aka firmware).
Microcode resides in a type of read-only memory
known as control memory.
When a machine instruction is fetched from RAM by
the processor to be executed, the machine
instruction’s op code serves as an address of a microcoded subroutine in control memory that “executes”
the machine instruction.
Simplification of
a Micro CodeDriven
Architecture
The Function Unit is
the same as the ALU
Memory M is RAM
and ROM