04/09/23
Operating Systems
Web server application
Threads
Operating Systems Operating Systems
Definition Advantages of using threads
A basic unit of CPU utilization Responsiveness
o Private: Thread ID, program counter, register set, stack o Allowing a program to continue running even part of it is blocked or
o Shared: code section, data section, OS resources (IO & file) lengthy
Examples: Resource sharing
o Web browsers o Memory, resources
o Word processors
o Database engines Economy
o RPC!
o Fast
Versus Process: Scalability
o Time consuming o Threads may be running in parallel on processing cores
o Resource intensive
04/09/23
Operating Systems Operating Systems
Multicore programming Types of parallelism
Multicore or multiprocessor systems putting pressure on programmers, Types of parallelism
challenges include:
o Dividing activities o Data parallelism
o Balance o Task parallelism
o Data splitting
o Data dependency
o Testing and debugging
As # of threads grows, so does architectural support for threading
o CPUs have cores as well as hardware threads
o Consider Oracle SPARC T4 with 8 cores, and 8 hardware threads per core
Parallelism implies a system can perform more than one task
simultaneously
Concurrency supports more than one task making progress
o Single processor / core, scheduler providing concurrency
Operating Systems Operating Systems
Multicore programming RPC using threads
Concurrent execution on single-core system:
Parallelism on a multi-core system:
04/09/23
Operating Systems Operating Systems
User threads and kernel threads Multithreading models
User threads - management done by user-level threads library Many-to-One
Three primary thread libraries:
o POSIX Pthreads (kernel-level lib, user-level lib)
One-to-One
o Windows threads (kernel-level lib)
o Java threads (kernel-level lib)
Many-to-Many
Kernel threads - Supported by the Kernel
Asynchronous vs. synchronous threading
o Parent & child threads
Operating Systems Operating Systems
User level vs. kernel level threads Many-to-one
Many user-level threads mapped to single kernel
thread
One thread blocking causes all to block
Multiple threads may not run in parallel on multicore
system because only one may be in kernel at a time
Few systems currently use this model
Examples:
o Solaris Green Threads
o GNU Portable Threads
Used in very few systems.
04/09/23
Operating Systems Operating Systems
One-to-one Thread operations and states
Each user-level thread maps to kernel thread Spawn (sinh sản)
o When a new process is spawned, a thread for that process
Creating a user-level thread creates a kernel is also spawned
thread
Block
More concurrency than many-to-one o When a thread needs to wait for an event, it will block
Number of threads per process sometimes Unblock
restricted due to overhead o When the event for which a thread is blocked occurs, the thread is
moved to the Ready queue
Examples
o Windows
o Linux Finish
o Solaris 9 and later
o When a thread completes, its register context and stacks are deallocate
Operating Systems Operating Systems
Many-to-many model Pthread: POSIX thread
Allows many user level threads to be
mapped to many kernel threads
Allows the operating system to create a
sufficient number of kernel threads
Two-level Model:
o Similar to M:M, except that it allows a user thread
to be bound to kernel thread
04/09/23
Operating Systems Operating Systems
Pthreads code for joining 10 threads Java thread programming
Operating Systems Operating Systems
Windows multithread C program Implicit threading
Three methods explored
o Thread Pools (Win)
o OpenMP (C lib)
o Grand Central Dispatch (Mac OS, iOS)
Block is in “^{ }” - ˆ{ printf("I am a block"); } #pragma omp parallel for
for(i=0;i<N;i++) {
c[i] = a[i] + b[i];
}
04/09/23
Operating Systems Operating Systems
Thread-local storage Windows threads data structures
Thread-local storage (TLS) allows each thread to have its own copy of data Implements the one-to-one mapping, kernel-level
Each thread contains
Useful when you do not have control over the thread creation process (i.e., o A thread id
when using a thread pool) o Register set representing state of processor
o Separate user and kernel stacks for when thread runs in
user mode or kernel mode
Different from local variables o Private data storage area used by run-time libraries and
o Local variables visible only during single function invocation dynamic link libraries (DLLs)
o TLS visible across function invocations
The register set, stacks, and private storage area are
known as the context of the thread
Similar to static data
o TLS is unique to each thread
Data structures:
o Execution thread block, kernel thread block and thread
environment block
Operating Systems Operating Systems
Thread termination Linux threads
Thread cancelation Linux refers to them as tasks rather than threads
o Asynchronous cancellation
o Deferred cancellation Thread creation is done through clone() system call
clone() allows a child task to share the address space of the parent task (process)
Who is “target thread”? o Flags control behavior
struct task_struct points to process data structures (shared or unique)
04/09/23
Operating Systems
Questions?
