Thread

A fundamental unit of CPU utilization

Those are have to be added to PCB

  1. It has to have own thread ID
  2. Keep track program counter
  3. Have to own register set
  4. Stack(Have to know where i was)

What has to be shared?

  • code section
  • data section
  • other OS resources
    • file handles
    • signal..
  • 2^N (Formula for the process created when N is indicating loop number)
    • for(i = 0 i<4; i++) means 2^4 = 16 processes are created
  • 2^N-1 (Formula for the # of child processes)

ThreadProcess

Benefits - Responseiveness - Resource sharing - Economy - Scalability

Multicore Programming

  • multicore(On same ship) / multi processor
  • each appears as separate processor to OS

Parallelism / Concurrency

  • Parallel system : More then 1 task at same time

  • Concurrent system : All task’s make “progress”

  • Hardware support increasing for multiple cores 2, 4, 8

Programming Challenges(multicore)

  • Ident: fying Tasks
    • How do we break up a process
  • Balance
    • How to distribute tasks across cores
  • Data splitting
    • How to place data[where physically]
  • Data dependency
    • Managing data shared by tasks
  • Testing and debugging
    • Concurrency add big increase in complexity

Types of Parallelism

  • Data
  • Task
  1. Data Parallelism
    • distribution of subsets of same data across multiple cores and then performing the same operation on each.
    • same function
  2. Task Parallelism
    • different function
    • distribution of tasks(threads) across multiple cores. Each task is performing a unique operation
    • same data
  • Most common to have a hybrid approach as location of data can have by impact on task performance

Example

  • 6 cores mean 6 processors
  • 6 threads mean could be one hardware threads or six hardware threads.

MultiThread Models

  • User Threads(User application simulating multithread, No support from kernel)
  • Kernel Threads(Management and support from the kernel)
  1. One-To-One Model - kernel threads
  2. Many-To-One Model - user threads (many user -> 1 kernel)
  3. Many-To-Many

Thread Libraries

  • User an API for management Thread

Amdahl

Thread Pool

Thread Issues

  1. Creation
  2. Signal Handling
    • Goes to specific thread
    • A signal is generated by the occurrence of a particular event
    • The signal is delivered to a process
    • Once delivered, the signal must be handled

Thread Cancellation

  1. Asynchronous cancellation
    • One thread immediately terminates the target thread
  2. Deferred cancellation
    • The target thread periodically checks whether it should terminate, allowing it an opportunity to terminate itself in an orderly fashion.