it allows a program to operate more efficiently by doing multiple things at the same time.

Threads

this creation is heavy-weight

Process creation

run within application

Threads

this creation is lightweight

Thread creation

puts pressure on programmers

Multicore or multiprocessor systems

it implies a system can perform more than one task simultaneously

Parallelism

supports more than one task making progress

Concurrency

it distributes subsets of the same data across multiple cores, same operation on each

Data parallelism

it is distributing threads across cores, each thread performing unique operation

Task parallelism

is a formula that specifies the potential speedup in latency of task execution at a specific workload

Amdahl’s Law

Many user-level threads mapped to single kernel thread

Many-to-One Model

Each user-level thread maps to kernel thread

One-to-One Model

Allows many user level threads to be mapped to many kernel threads

Many-to-Many Model

Similar to M:M, except that it allows a user thread to be bound to kernel thread

Two-level Model

it refers to the approach where the management of threads is abstracted away from the programmer.

Implicit threading

used to manage a pool of pre-created threads waiting for tasks to execute.

Thread Pools

Multiple threads (tasks) are forked and then joined.

Fork-Join Parallelism

Creation and management of threads done by compilers and run-time libraries rather than programmers

Implicit threading

also known as libdispatch, is a technology developed by Apple to optimize application support for systems with multi-core processors. It essentially simplifies writing code for iOS and macOS

Grand Central Dispatch (GCD)

is an open-source C++ template library developed by Intel for parallel programming on multi-core processors. Focus on tasks, not threads

Intel Threading Building Blocks (TBB)

makes it easy to write programs that run on multiple CPU cores simultaneously — using threads — without writing all the thread management and synchronization manually.

OpenMP

Terminating a thread before it has finished

Thread Cancellation

It is an application programming interface (API) that supports multi-platform shared-memory multiprocessing programming in C, C++, and Fortran

Open MultiProcessing (OpenMP)

Thread to be cancelled is

Target thread

one approach of thread cancellation that allows the target thread to periodically check if it should be cancelled

Deferred cancellation

primary API for Windows applications

Windows API

one approach of thread cancellation that terminates the target thread immediately

Asynchronous cancellation

primary data structures of a thread include:

ETHREAD, KTHREAD, TEB

Implements the one-to-one mapping

Windows Threads

thread id, user-mode stack, thread-local storage, in user space

TEB (thread environment block)

Linux refers to them as tasks rather than threads

Linux Threads

is the kernel mode representation of a thread within the Windows kernel.

KTHREAD (kernel thread block)

context of the thread

register set, stacks, and private storage area are known as the

is the user mode representation of a thread within a Windows process.

ETHREAD (executive thread block)

creation of this is done through clone() system call

Thread creation

split/divide tasks

Fork Parallelism

join all tasks

Join Paralellism