Threads in Operating System | Multithreading | Interprocess Communication
Download as PPTX, PDF0 likes931 views
This document provides an introduction to threads. It discusses the differences between processes and threads, how threads are implemented in Linux, and challenges with multithreading like race conditions. Interprocess communication methods like shared memory and message passing are also covered. The benefits of multithreading include improved responsiveness and resource sharing. Multiprocessing uses multiple CPU cores to run programs in parallel while multithreading shares memory between threads.
Threads in Operating System | Multithreading | Interprocess Communication
- 2. PREVIOUS SESSION • Program vs Process • Process states • Process control block • Process creation through fork() • Exec(), wait and exit() system calls • Zombie and orphan process
- 3. AGENDA • Interprocess communication • What is a thread • Thread vs process • Multithreading and multiprocessing • Multithreading models • Challenges with multithreading
- 4. INTERVIEW QUESTIONS • What are similarities and differences between a process and a thread? • What is IPC? Give some examples of IPC • Why process coordination is necessary in OS? • What are advantages and disadvantages of using shared memory and IPC? • What is multi-tasking, multi-programming, multithreading and multiprocessing? • What are the challenges faced in multithreading programs?
- 5. • Difference between chrome and firefox in terms of multiprocessing and multithreading? • What is the difference between named pipe and unnamed pipe? • What are the benefits of multithreading?
- 6. INTERPROCESS COMMUNICATION • Independent vs cooperating process • Does a process shares data with other process? • Debugging becomes tough in cooperating processes. • Why process cooperation is necessary?
- 7. Need for process cooperation • Information sharing • Computational speedup • Modularity • Convenience • Detailed explanation: https://youtu.be/7jN8tgLqbFc
- 8. Models/Types of interprocess communication • Shared memory • A region of shared memory is first established • Processes then share information by reading and writing to this shared space. • Threads • Message passing • A link is established between two or more processes • Messages are then shared on this link • Detailed explanation: https://youtu.be/FCPUTFmAlIg
- 10. Shared memory vs Message Passing
- 11. Examples • Shared memory is achieved using multithreading. • Message passing examples • Sockets • Pipes • Remote procedure calls • Pipes • Ordinary pipes • Named pipes
- 13. PIPES • Unnamed pipe • ls | grep “hello” • cat filename | grep “hello” • Named pipe • mkfifo <file name> • Unnamed pipe can only be used to communicate between parent and child processes. • A named pipe exists on the file system. • Named pipe can be used for communication over a network. • Named pipe are bidirectional.
- 14. WHAT IS A THREAD • A thread is a basic unit of CPU utlilization. • A thread consists of: • Thread ID • Program counter • Register sets • Stack segment memory • What does CPU needs to execute instructions? • Registers • Program counter • Process is also a unit of CPU utilization, but not a basic unit. • Detailed explanation - https://youtu.be/pYbBgA76wDg
- 15. What a thread shares with other threads? • Code section memory • Data segment memory • Heap segment memory • Open files • Signals
- 16. • Linux has a unique implementation of threads. To the Linux kernel, there is no concept of a thread. Linux implements all threads as standard processes. • Each thread has a unique process control block and appears to the kernel as a normal process—threads just happen to share resources, such as an address space, with other processes. • Windows – explicit support for threads (lightweight process)
- 17. Threads implementation in Linux • Normal fork() • clone(SIGCHLD, 0); • Thread implementation • clone(CLONE_VM | CLONE_FS | CLONE_FILES | CLONE_SIGHAND, 0);
- 19. THREAD VS PROCESS PROCESS THREAD
- 20. Differences continued Detailed explanation - https://youtu.be/MwDccB9tvp0
- 21. MULTITHREADING • Multiple threads within a process • Should be supported by the hardware • Concurrency vs parallelism • Most applications are multithreaded today • Browser – rendering + fetching • Web server having multiple clients • Word processor – displaying graphics + waiting for keystroke + spell check
- 22. Benefits of multithreading • Responsiveness – browser, word-processor • Resource sharing – threads share the memory and resources by default • Economy – creating multiple processes is very costly. • Scalability – increases the parallelism, web servers
- 23. Challenges • Concurrency issues • Multiple threads trying to update the resource at the same time. • Leads to race condition • How i++ is actually executed?
- 24. Initially i = 7, final value = 9 and two threads incrementing i
- 25. Initial value = 7, final value = 8
- 26. MULTIPROCESSING • Multiple processes within a process • Using fork() to create multiple processes • Process creation is time-consuming than creating threads. • Process address space is not shared among processes.
- 27. CPUs, cores and hyperthreading • A core is the basic computational unit of CPU. It has registers and all other resources to execute a program. • Processes -> threads • CPU are the processor chips. • Initially, a CPU had only one core. • Today, dual-core or quad-core processors • Hyper threading – allows a single core to run multiple threads. Some hardware are duplicated.
- 29. When to use thread vs process • Threads are faster • Crashing one thread crashes other threads too. • No security between threads – lead to race condition • A tradeoff between speed vs reliability • Chrome vs firefox – multiprocessing vs multithreading • Web servers • Word processor • Computer games
- 30. Multiprogramming, multithreading, multiprocessing • Multiprogramming – ability of the system to run multiple programs concurrently. Time sharing OS. Can happen even on a single core. • Multitasking is similar to multiprogramming as we are able to do multiple tasks at the same time. • Multiprocessing – Using multiple cores for executing multiple programs in parallel. • Multithreading – Similar to multiprocessing but here the memory is shared between threads. Usually, very fast.
- 31. Multithreading models • How user threads are mapped to kernel threads? • Not asked in any interviews • 1:1 mapping • N:N mapping • M:N mapping • Videos: • https://youtu.be/lYNk1cNWCfc • https://youtu.be/1Jij_qY2kp0
- 32. NEXT SESSION: PROCESS SCHEDULING ALGORITHMS