03 Evolution

Early Computers¶

Something compiler

Aim was to optimize CPU time

Batch Processing Systems¶

Shared computer systems

Steps¶

• Operator hired
• Jobs with similar needs were batched together and run though computer, to reduce setup time

Problems¶

• Stopped jobs (normal/abnormal)
• Dump (Log file) Status of the memory is stored into a text file
• Load device with next job
• Restart computer

Issues¶

(i missed this)

CPU Burst

I/O Burst

Solution¶

• Perform CPU execution and I/O concurrently
• Multi-programming and Time Sharing
  • Multi-programming \(\ne\) Multi-Processing
  • Multi-Programming means using 1 CPU and performing multiple programs concurrently
  • Multi-Processing means using multiple CPUs

Time Sharing/Multi-Tasking Systems¶

Extension of Multiprogramming

Time Multiplexing¶

CPU switch between the programs kept in PM

This is done by dividing CPU time into Time Quanta(fixed intervals)

Involves timer = counters + clock

Space Multiplexing¶

After partitioning primary memory into OS and User area, the User Area is then further partitioned for multiple users

Advantages¶

• Supported user interaction
• Improved response time

Difference from Multi-Programming¶

A single CPU switches between multiple users, in a way that every user feels as if they are using the CPU.

Steps¶

1. P1 runs to completion
2. P1 requests I/O
3. P1 has not completed the execution, but the time quanta expired
4. Timer interrupt occurs (as the down-counting timer is over), the CPU switches to another user

Time Delay¶

Time between switching between 2 users

= Count value * Time Period

Features¶

• Job, CPU Scheduling
• Memory Management
• Resonable response time
  • Using Virtual memory
  • Swap/Roll
• File system and disk management
• job sync and communication
• Deadlocks handling
• protection and security

Swap/Roll¶

Multi-Tasking¶

Each program of the same user is a ‘task’. The processor switches between each task.

Computer System¶

Core = CPU, Register, Local Cache

Multi-Processor System¶

• i3 has dual-core
• i5, i7 have quad-core

Multi-Processing¶

Concurrent execution using multiple processors

Increased Throughput¶

Execution of programs is distributed to multiple processors, hence more output

Expected increase is the number of additional processors, but this is not real. Check Amdahl's law in CA

Improved Reliability¶

Even if one processor fails, complete failure is avoided. However there will be obvious delay.

Useful for testing fault Tolerance¶

2 systems perform the same task and the results will be checked.

However, increased redundancy and consumption of resources.

Desktop OS¶

WindowsOS, MacOS, Linux

Mobile OS¶

Android, iOS

Embedded Systems¶

Washing machine, dishwasher

Realtime OS¶

More complex, Dynamic and have time constraints

eg: Industrial control systems, Weapon systems

Distributed Systems¶

Multiple hardware devices are networked together.

Each device runs a subset of the ‘distributed OS’

When a process executes, the process is split into subprocesses which is sent to different nodes.

Advantages¶

Some more points are there

• Data access
• Special h/w requirements
• Load balancing

Definitions¶

Throughput¶

No of tasks completed

Turn around time¶

Time between starting of execution of a program and its completion

Questions¶

1. Steps when a user double clicks MS Powerpoint icon in a Windows Desktop
   • CPU recognizes user click (Hardware Interrupt), after execution of current instruction
   • Mode changes to kernel mode
   • Identifies .exe file in secondary memory
   • OS loads program into primary memory
   • Allocate the CPU
   • Hardware interrupt to display on screen
2. Important activities/functions done by OS
   • Functions of OS
3. Suppose a computer system has 10000 bytes of memory available. Out of this, the OS occupies 5000bytes. Now it is required to run a program whose size is 7000bytes. Is this possible?
   • No, it is not possible in basic computers
   • Every program has to be loaded into primary memory for its execution
   • The memory is split into 2 parts
   • OS segment
   • Program segment (for loading the program)
   • As only 5000bytes is available for the program segment, we cannot load this program, and hence we cannot run it
   • However, we can sort this issue using virtual memory
4. What is meant by dual mode of operation in intel CPUs
   • To ensure better security and stricter access priveleges, the CPU has 2 modes of operations
   • Instructions by the user are done in user mode, with lower priveleges
   • Instructions by the user are done in kernel mode, with higher priveleges
5. Can you guess in what mode OS program is run and in what mode user program is run?
   • Kernel mode
   • User mode
6. What is the difference between hardware and software interrupt in a computer system? Can you give examples of each type.
   • Hardware interrupt are interrupts involving I/O devices. eg: keyboard input
   • Software interrupts do not include I/O devices. eg: Mathematical errors, rejected priviledged instructions, Exceptions
7. Why do you require a bootstrap loader?
   • POST Diagnostics
   • Detect and initialize devices
   • Initialize registers and primary memory
   • Load the kernel into primary memory
8. Using the Interrupt Vector Table (of an 8086 processor operating in real mode) shown below, determine the address of the ISR of a device with interrupt vector 42H.

Answer \(= \text{4D6EA_H}\) 9. Differentiate between multi-programming & multi-processing - Multi-programming means running multiple programs, using only 1 processor - Multi-processing means running one or more programs, using multiple processors

System Call¶

It is the way for user instructions to perform priviledged instructions.

Virtual Memory¶

Combines primary and secondary memory into a single ‘logical memory’

Load only the required part of a program into primary memory.

Comments