LQ1 - Operating Systems - Threads & Concurrency

heavy weight, light weight

process creation is ____ ____ while thread creation is ____ ____

parallelism

implies a system can perform more than one task simultaneously

concurrency

supports more than one task making progress

data parallelism

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

task parallelism

distributing threads across cores, each thread performing a unique operation

Amdahl’s Law

identifies performance gains from adding additional cores to an application that has both serial and parallel components

user threads

management done by user-level threads library

• POSIX Pthreads

• Windows threads

• Java threads

enumerate the three primary thread libraries

kernel threads

threads which are supported by the kernel

many-to-one

is a multithreading model where user-level threads are mapped to a single kernel thread

many-to-one

is a multithreading model where one thread blocking causes all to be blocked.

many-to-one

is a multithreading model where multiple threads may not run in parallel on muticore system because only one may be in kernel at a time.

one-to-one

is a multithreading model where each user-level thread maps to a kernel thread

many-to-many

is a multithreading model that allows many user level threads to be mapped to many kernel threads

many-to-many

is a multithreading model that allows the operating system to create a sufficient number of kernel threads

two-level model

is a multithreading model that is similar to M:M, except that it allows a user thread to be bound to

kernel thread

thread library

provides programmer with API for creating and managing threads

• library entirely in user space

• kernel-level library supported by the OS

two main ways of implementing thread libraries

Pthreads

A POSIX standard (IEEE 1003.1c) API for thread creation and synchronization. It be provided either as user-level or kernel-level.

implicit threading

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

• Thread Pools

• Fork-Join

• OpenMP

• Grand Central Dispatch

• Intel Threading Building Blocks

what are the five methods of implicit threading?

thread pools

An implicit threading method where a number of threads are created in a pool where they await work

Fork-Join Parallelism

An implicit threading method where multiple threads (task) are forked and then joined

Grand Central Dispatch

An implicit threading method that extends to C, C++ and Objective-C languages, API, and run-time

library.

It allows identification of parallel sections. It also manages most of the details of threading and puts blocks in “^{}”.

serial, main queue

____ is a type of dispatch queue where blocks are removed in FIFO order, queue is per process, called ___ ____

concurrent

is a type of dispatch queue where blocks are removed in FIFO order but several may be removed at a time

signals

are used in UNIX systems to notify a process that a particular event has occurred.

signal handler

is used to process signals

default handler

Every signal has ___ ____ that kernel runs when handling signal

user-defined

____ signal handler can override default

thread cancellation

Terminating a thread before it has finished

target thread

Thread to be canceled

Asynchronous cancellation

terminates the target thread immediately

Deferred cancellation

allows the target thread to periodically check if it should be cancelled. It is the default type of thread cancellation.

signals

On Linux systems, thread cancellation is handled through ___

interrupt()

Deferred cancellation uses the ____ method

Thread-local storage (TLS)

allows each thread to have its own copy of data

Thread-local storage (TLS)

Useful when you do not have control over the thread creation process

lightweight process (LWP)

An intermediate data structure between user and kernel threads

lightweight process (LWP)

Appears to be a virtual processor on which process can schedule user thread to run

upcalls

a communication mechanism, provided by scheduler activations, from the kernel to the upcall handler in the thread library

Windows API

primary API for Windows applications

context

The register set, stacks, and private storage area are known as the ___ of the thread

• thread id

• register set representing state of processor

• separate user and kernel stacks for when thread runs in user mode or kernel mode

• Private data storage area used by run-time libraries and dynamic link libraries (DLLs)

enumerate the components each windows thread has

ETHREAD (executive thread block)

a primary data structure of a Windows thread that includes pointer to process to which thread belongs and to KTHREAD, in kernel space

KTHREAD (kernel thread block)

a primary data structure of a Windows thread with scheduling and synchronization info, kernel-mode stack, pointer to TEB, in kernel space

TEB (thread environment block)

a primary data structure of a Windows thread with thread id, user-mode stack, thread-local storage, in user space

tasks

the Linux term for threads

clone()

system call to create a thread in Linux

struct task_struct

points to process data structures (shared or unique)

CLONE_FS

a Linux flag that indicates that File-system information is shared

CLONE_VM

a Linux flag that indicates that the same memory space is shared

CLONE_SIGHAND

a Linux flag that indicates that signal handlers are shared.

CLONE_FILES

a Linux flag that indicates that the set of open files is shared.