CHAPTER 4: THREAD AND CONCURRENCY

CHAPTER 4: THREAD AND CONCURRENCY

1. Multi-threading Models

   • Many to one
     • One thread blocking cause all to block
     • Many user level mapped to a single kernel thread
   • One to one
     • Each user-level maps to kernel thread
     • Create User-Level Thread=Create Kernel
     • Number of threads per process sometimes restricted due to overhead
   • Many to many
     • Many user-level mapped to many kernel
     • Allows OS to create sufficient number of kernel threads

2. Multicore Programming

   • Challenges
     • Dividing activities
     • Balance
     • Data splitting
     • Data dependency
     • Testing and debugging
   • Parallelism
     • Implies that a system can perform more than one task simultaneously
     • Type of Parallelism
       • Task Parallelism
         • Distributing threads across cores, each thread performing unique operation
       • Data Parallelism
         • Distribution subsets of the same data across multiple core, same operation on each
   • Concurrency
     • Supports more than one task making progress (single processor/core, scheduler providing concurrency

3. Benefits

   • Responsiveness
   • Scalability
   • Resource sharing
   • Economy

4. Threading issues

   • Semantics of forks() and exec() system calls
   • Signal Handling (synchronous & asynchronous)
     • Signals are used in UNIX to notify a process that a particular event has occurred
     • Every signal has default handler that kernel runs when handling signal
     • Single handler is used to process signals. Generated by particular event, delivered to a process, handled by one of two signal handler (default, user-defined)
   • Thread Cancellation
     • Terminating threads before it has finished
     • Target thread = Thread to be cancelled
     • 2 general approaches
       • Asynchronous Cancellation
       • Deferred Cancellation
   • Thread-Local Storage
     • Allows each thread to have copy of its data
     • Different from local variable, but similar to static data
   • Scheduler Activation
     • Many-to-Many and 2-level models require communication to maintain the appropriate number of kernel threads allocated to the application
     • Typically uses lightweight process
       • Appears to be virtual processor on which process can schedule user thread to run