Introduction to Threads

00:14:06
https://www.youtube.com/watch?v=LOfGJcVnvAk

摘要

TLDRThis lecture delves into the topic of threads in operating systems, providing a clear distinction between processes and threads. It defines a thread as a basic unit of CPU utilization and execution, consisting of a thread ID, program counter, register set, and stack. Threads belonging to the same process share resources such as code, data sections, and system resources. The lecture compares single-threaded (heavyweight) processes, which handle single tasks, with multi-threaded processes that allow multiple tasks to be performed simultaneously. This approach enhances system efficiency by speeding up computation and improving concurrency in multi-processor systems. Additionally, the lecturer uses visuals to show the threading process in systems like Chromium, exemplifying a multi-threaded approach where threads perform tasks like displaying and downloading simultaneously. The lecture also outlines the benefits of multi-threaded programming, including improved responsiveness, resource sharing, economic efficiency, and enhanced use of multi-processor architectures.

心得

  • 💡 Threads are basic units of CPU utilization.
  • 🧩 Threads consist of a thread ID, program counter, register set, and stack.
  • 🔄 Threads within a process share code, data sections, and resources.
  • 🔀 Multi-threaded processes improve task concurrency.
  • ⚙️ Multi-threading boosts system efficiency.
  • 🌐 Chromium browser exemplifies multi-thread usage.
  • 🚀 Benefits include enhanced responsiveness and economy.
  • 🔧 Resource sharing reduces process creation costs.
  • 🔥 Utilizes multi-core processors for better concurrency.
  • 📈 Multi-threaded programming elevates computing power.

时间轴

  • 00:00:00 - 00:05:00

    In the lecture, the concept of threads is introduced, emphasizing their significance in operating systems. Threads are described as the basic units of CPU utilization, distinguishing them from processes. A process, described as a program in execution, can contain multiple threads, which share resources like code and data sections, while maintaining unique identifiers, program counters, and stacks. Single-threaded processes, known as heavyweight processes, perform one task at a time, whereas multi-threaded processes enhance efficiency by allowing simultaneous execution of tasks. Diagrams illustrate single versus multi-threaded processes, highlighting the shared and unique elements of threads, and demonstrate the advantage of multi-threading in improving system performance and speed.

  • 00:05:00 - 00:14:06

    The discussion continues on the multi-threaded processes by using a practical example of viewing threads in a system's browser process. By examining Chrome's running processes, users can observe multiple threads handling distinct tasks. Multi-threaded systems ensure responsiveness and resource sharing since threads can perform parallel tasks, like downloading and displaying webpages simultaneously. Benefits of multi-threading include enhanced user responsiveness, resource sharing without needing separate resources, economical context switching, and efficient use of multi-processor architectures, allowing concurrent execution across processors. The lecture concludes with a summary of threads' basic concepts—underlining their role in optimizing computing tasks and future learning directions.

思维导图

Mind Map

常见问题

  • What is a thread?

    A thread is a basic unit of CPU utilization and execution within a process.

  • What makes up a thread?

    A thread consists of a thread ID, a program counter, a register set, and a stack.

  • How do threads differ from processes?

    Threads are the basic units of execution within a process, which is a program under execution.

  • What resources do threads share within a process?

    Threads within the same process share the code section, data section, and other operating system resources like open files and signals.

  • What is a single-threaded process?

    A single-threaded or heavyweight process has only one thread of control and can perform only one task at a time.

  • What is a multi-threaded process?

    A multi-threaded process contains multiple threads, allowing it to perform several tasks simultaneously for greater efficiency.

  • Why is multi-threading beneficial for applications?

    Multi-threading increases responsiveness, allows resource sharing, is economical, and better utilizes multi-processor architectures.

  • How does multi-threading improve responsiveness?

    Multi-threading allows parts of a program to continue running even if another part is blocked, improving user responsiveness.

  • What is the economic benefit of multi-threading?

    Multi-threading is economical because it allows threads to share resources of a process, reducing the costly allocation needed for separate processes.

  • How does multi-threading utilize multi-processor architectures?

    Multi-threading enables different threads to run in parallel across multiple processors, improving concurrency.

查看更多视频摘要

即时访问由人工智能支持的免费 YouTube 视频摘要！
字幕
en
自动滚动:
  • 00:00:00
    in this lecture we'll be studying about
  • 00:00:02
    threats threats are an important topic
  • 00:00:04
    in operating system and in one of the
  • 00:00:06
    previous lectures when we discuss about
  • 00:00:08
    process management I have shown you the
  • 00:00:10
    difference between processes and threats
  • 00:00:12
    so I told you that a program under
  • 00:00:14
    execution is known as a process and
  • 00:00:17
    threat is a basic unit of execution so
  • 00:00:20
    each program may have a number of
  • 00:00:22
    processes associated with it and each
  • 00:00:25
    process can have a number of threats
  • 00:00:27
    executing in it so thread is a basic
  • 00:00:30
    unit of execution or a basic unit of CPU
  • 00:00:34
    utilization so that is how we can define
  • 00:00:36
    a thread it is a basic unit of CPU
  • 00:00:38
    utilization now let us see what
  • 00:00:41
    comprises of a thread so a thread
  • 00:00:43
    comprises of a thread ID a program
  • 00:00:46
    counter a register set and a stack so
  • 00:00:49
    each thread comprises of these four
  • 00:00:52
    unique items and apart from this it
  • 00:00:55
    shares with other threads belonging to
  • 00:00:57
    the same process it's Code section data
  • 00:01:00
    section and other operating system
  • 00:01:02
    resources such as open files and signals
  • 00:01:05
    so these are the things that comprises
  • 00:01:08
    of a thread and apart from that the
  • 00:01:11
    thread shares these things like the Code
  • 00:01:13
    section data section and other operating
  • 00:01:15
    system resources are shared between
  • 00:01:18
    threads belonging to the same process so
  • 00:01:20
    as I told you a process can contain
  • 00:01:23
    different threads so the threads
  • 00:01:25
    belonging to the same process will share
  • 00:01:27
    the resources slight open files signals
  • 00:01:30
    Code section data section and other
  • 00:01:32
    things so that is how a thread is
  • 00:01:35
    basically made up so a traditional or
  • 00:01:38
    heavyweight process has a single thread
  • 00:01:41
    of control but if a process has multiple
  • 00:01:44
    threads of control it can perform more
  • 00:01:47
    than one task at a time so a traditional
  • 00:01:50
    process also known as a heavyweight
  • 00:01:52
    process had only a single thread of
  • 00:01:55
    control that means it can have only one
  • 00:01:58
    thread but this is not a very efficient
  • 00:02:00
    system but if a process is having more
  • 00:02:02
    than one threads then each thread will
  • 00:02:05
    be assigned a different task and hence
  • 00:02:08
    that process can perform more than one
  • 00:02:10
    task at a time because it is having
  • 00:02:12
    multiple
  • 00:02:13
    and multiple threats will be performing
  • 00:02:15
    multiple actions so there we see how
  • 00:02:18
    treads becomes useful and how using
  • 00:02:21
    threads makes the system much more
  • 00:02:24
    efficient so here let us try to
  • 00:02:26
    understand the concept of threads using
  • 00:02:28
    these two diagrams so here in the first
  • 00:02:31
    diagram it is a single threaded process
  • 00:02:33
    which we refer to as a traditional or a
  • 00:02:35
    heavyweight process over here so here we
  • 00:02:38
    are having only a single thread so this
  • 00:02:41
    whole block is considered as a process
  • 00:02:44
    and this process contains only one
  • 00:02:46
    thread and this is the code section the
  • 00:02:49
    data section the files the registers and
  • 00:02:52
    the stack belonging to this process so
  • 00:02:55
    there is only one thread that means this
  • 00:02:57
    process can perform only one task at a
  • 00:02:59
    time but in contrast to this here we
  • 00:03:02
    have another diagram showing a
  • 00:03:04
    multi-threaded process so here this
  • 00:03:06
    whole block represents a multi-threaded
  • 00:03:09
    process so it is a single process and it
  • 00:03:12
    is having multiple threads over here so
  • 00:03:14
    if you see there are three threads in
  • 00:03:17
    this process and each of these threads
  • 00:03:19
    they have their own stack and registers
  • 00:03:22
    so all these three threads they have
  • 00:03:25
    their own stack and registers and if you
  • 00:03:27
    see on top the code section the data
  • 00:03:30
    section and the files belonging to this
  • 00:03:33
    process are shared by these three
  • 00:03:36
    threads and each thread they have their
  • 00:03:39
    own stacks and registers so this is a
  • 00:03:42
    diagram of a multi-threaded process so
  • 00:03:45
    in this case this process can perform
  • 00:03:48
    multiple tasks at a time because each
  • 00:03:50
    thread will be performing a different
  • 00:03:52
    task and hence multiple tasks can be
  • 00:03:55
    performed at a time by the help of this
  • 00:03:57
    multi-threaded process so from that
  • 00:04:00
    itself we must have understood that it
  • 00:04:02
    is much more efficient as compared to
  • 00:04:04
    the single threaded process and it will
  • 00:04:06
    make our computation faster and more
  • 00:04:09
    efficient so in most of the systems that
  • 00:04:11
    we use today we follow the system of
  • 00:04:14
    multi-threaded processes it would be
  • 00:04:16
    good if we can visually see how it works
  • 00:04:18
    in our system in order to understand how
  • 00:04:20
    processes are there and how threads are
  • 00:04:23
    associated with the processes so here I
  • 00:04:25
    have a small software
  • 00:04:26
    processed threats view which helps us to
  • 00:04:29
    see the processes and the threats
  • 00:04:31
    running in our system so if I click here
  • 00:04:34
    I can see the processes running in my
  • 00:04:36
    system so I have opened the chromium
  • 00:04:39
    browser in my system so here I can see
  • 00:04:41
    the chrome dot exe processes which are
  • 00:04:44
    running in my system so there are quite
  • 00:04:47
    a few number of chrome dot exe processes
  • 00:04:49
    running in the system now if I want to
  • 00:04:52
    see the thread associated with it I will
  • 00:04:54
    just click on it and here the details
  • 00:04:57
    about the thread is populated so there
  • 00:05:00
    are this many number of threads
  • 00:05:03
    associated with that single chrome dot
  • 00:05:05
    exe process that was for running so as I
  • 00:05:08
    told you the thread comprises of a
  • 00:05:10
    thread ID so these under IDs of the
  • 00:05:13
    thread which identifies those threads
  • 00:05:15
    uniquely and then these are the context
  • 00:05:17
    switches and the last context switches
  • 00:05:19
    that happened so we have discuss about
  • 00:05:21
    context switches in one of the previous
  • 00:05:23
    lectures and here there are so many
  • 00:05:24
    informations about the thread that we
  • 00:05:27
    have selected now if we come here we can
  • 00:05:30
    see that there is the address of the
  • 00:05:33
    thread and here we can see the details
  • 00:05:35
    of the stack associated with that
  • 00:05:38
    process so I told you that threads also
  • 00:05:40
    container stack so in this way we can
  • 00:05:43
    see what are the processes running in
  • 00:05:45
    our system and we can also see the
  • 00:05:47
    threads associated with those processes
  • 00:05:49
    so we clearly see that in this system it
  • 00:05:52
    is following a multi-threaded approach
  • 00:05:54
    so when I click that single process of
  • 00:05:57
    that chrome dot exe we see that this
  • 00:05:59
    many number of threads are running
  • 00:06:01
    inside that process so each of these
  • 00:06:03
    threads may be used for performing a
  • 00:06:05
    different task so in that way the
  • 00:06:07
    browser can do more than one task at a
  • 00:06:10
    time so let's say for example one of the
  • 00:06:12
    processes of this browser chrome load
  • 00:06:14
    exe is used for displaying the web page
  • 00:06:17
    on the browser window and let's say that
  • 00:06:19
    another thread is used for downloading
  • 00:06:22
    something from the internet so we know
  • 00:06:24
    that when we are downloading something
  • 00:06:25
    we can also view the page at the same
  • 00:06:28
    time so these two things are happening
  • 00:06:30
    simultaneously that is possible because
  • 00:06:32
    we are having multiple threads one
  • 00:06:34
    thread is taking care of displaying and
  • 00:06:36
    the other thread is taking care of
  • 00:06:37
    downloading so if we were having
  • 00:06:40
    a single-threaded process then while the
  • 00:06:43
    downloading was happening the page could
  • 00:06:45
    not be displayed or while the page is
  • 00:06:47
    being displayed the download will not
  • 00:06:49
    happen so that is this a simple example
  • 00:06:51
    that I want to take in order to make us
  • 00:06:53
    understand the usage of multi-threaded
  • 00:06:55
    processes so different threads perform
  • 00:06:57
    different tasks and hence multiple tasks
  • 00:06:59
    can be performed in the same time hence
  • 00:07:01
    improving our efficiency and speed of
  • 00:07:04
    computation so we have seen a visual
  • 00:07:06
    example of how multi thread processes
  • 00:07:08
    are there in our system now let us see
  • 00:07:10
    what are the benefits of having
  • 00:07:11
    multi-threaded processes so from the
  • 00:07:14
    example that I have taken you must have
  • 00:07:15
    already understood the benefits of using
  • 00:07:17
    multi-threaded processes but let us
  • 00:07:19
    understand these benefits in more detail
  • 00:07:22
    so the benefits of multi-threaded
  • 00:07:23
    programming can be broken down into four
  • 00:07:25
    major categories so let's see what they
  • 00:07:28
    are one-by-one so the first benefit is
  • 00:07:30
    responsiveness multi-threading an
  • 00:07:32
    interactive application may allow a
  • 00:07:35
    program to continue running even if part
  • 00:07:38
    of it is blocked or is performing a
  • 00:07:40
    lengthy operation thereby increasing
  • 00:07:43
    responsiveness to the user so we are
  • 00:07:46
    having a better responsiveness if we
  • 00:07:49
    have multi-threading so I can explain
  • 00:07:52
    this using the same example that I took
  • 00:07:53
    about the web browser since we said that
  • 00:07:56
    we are having different threads one for
  • 00:07:58
    displaying the web page and another one
  • 00:08:00
    for downloading and let's say that
  • 00:08:02
    another one for the user to interact
  • 00:08:04
    with the web page so if only one thread
  • 00:08:07
    was there then until and unless one of
  • 00:08:10
    the tasks completes the other task
  • 00:08:12
    cannot be done so the user have to wait
  • 00:08:14
    for one of the tasks to be complete in
  • 00:08:16
    order to go to the next task but in case
  • 00:08:19
    of multi-threading since different
  • 00:08:21
    threads are simultaneously performing
  • 00:08:22
    different tasks the responsiveness to
  • 00:08:25
    the user increases so that is the first
  • 00:08:27
    benefit of multi-threaded programming
  • 00:08:30
    let us see what is the next benefit so
  • 00:08:32
    the next benefit is a resource sharing
  • 00:08:34
    so by default thread share the memory
  • 00:08:36
    and resources of the process to which
  • 00:08:39
    they belong the benefits of sharing code
  • 00:08:41
    and data is that it allows an
  • 00:08:44
    application to have several different
  • 00:08:46
    threads of activity within the same
  • 00:08:48
    address space so the next benefit is
  • 00:08:51
    resource sharing so when we saw the
  • 00:08:53
    diagram of
  • 00:08:53
    different processes here we saw that
  • 00:08:56
    these multiple threats they are sharing
  • 00:08:58
    the resources of the same process so the
  • 00:09:01
    code data and files are shared between
  • 00:09:05
    these three threats so we said that it
  • 00:09:07
    shares with other threats belonging to
  • 00:09:09
    the same process it's Code section data
  • 00:09:12
    section and other operating system
  • 00:09:14
    resources suggest open files and signals
  • 00:09:17
    so we see that resource sharing is
  • 00:09:19
    happening in multi-threaded programming
  • 00:09:21
    so the benefit of sharing code and data
  • 00:09:24
    is that it will allow application to
  • 00:09:26
    have several different types of activity
  • 00:09:28
    within the same address space so we can
  • 00:09:30
    have several different threats in the
  • 00:09:32
    same address space of the same process
  • 00:09:34
    so by resource sharing we are making our
  • 00:09:37
    system more efficient because it does
  • 00:09:40
    not need to have separate or dedicated
  • 00:09:42
    resources for each and every threats so
  • 00:09:45
    sharing resource is another benefit of
  • 00:09:47
    multi-threaded programming so coming to
  • 00:09:50
    the next benefit we have economy so
  • 00:09:52
    allocating memory and resources for
  • 00:09:54
    process creation is costly because
  • 00:09:57
    threats share resource of the process to
  • 00:09:59
    which they belong it is more economical
  • 00:10:01
    to create and context switch threats so
  • 00:10:05
    we already saw that there is a resource
  • 00:10:07
    sharing happening in case of
  • 00:10:09
    multi-threading so if there was no
  • 00:10:11
    multi-threading then we will need to
  • 00:10:13
    have a separate process for each and
  • 00:10:16
    every task that has to be performed and
  • 00:10:18
    allocating memory and resources for the
  • 00:10:20
    processes is costly because each process
  • 00:10:24
    will need to have its dedicated
  • 00:10:25
    resources and memory which it may not be
  • 00:10:28
    able to share with other processes but
  • 00:10:31
    in threads as they are able to share the
  • 00:10:34
    resource of the processes to which they
  • 00:10:36
    belong it is more economical to create
  • 00:10:39
    and switch between threads because we
  • 00:10:42
    don't need to have those dedicated
  • 00:10:43
    resources they are able to share and
  • 00:10:45
    hence it is economy so this economy it
  • 00:10:48
    comes from this resource sharing now
  • 00:10:51
    moving on the next and the last benefit
  • 00:10:53
    that we have is utilization of multi
  • 00:10:56
    processor architectures the benefit of
  • 00:10:58
    multi-threading can be greatly increased
  • 00:11:00
    in a multiprocessor architecture where
  • 00:11:03
    threads may be running in parallel
  • 00:11:05
    different processors a single-threaded
  • 00:11:07
    process can only run on one CPU no
  • 00:11:11
    matter how many are available
  • 00:11:13
    multi-threading on a multi CPU machine
  • 00:11:15
    increases concurrency so the next
  • 00:11:18
    benefit we have is utilization of multi
  • 00:11:20
    processor architecture so we have
  • 00:11:22
    already studied about multi processor
  • 00:11:24
    architecture in this lecture series of
  • 00:11:26
    operating system so multi processor
  • 00:11:28
    architecture means there are several
  • 00:11:30
    number of processors in the world system
  • 00:11:32
    but if we are not having a
  • 00:11:34
    multi-threaded approach and instead
  • 00:11:36
    having a single threaded process then
  • 00:11:39
    what will happen is that if we are
  • 00:11:41
    having one process that one process can
  • 00:11:44
    run on only one processor or on one CPU
  • 00:11:47
    so you may have different CPUs so let's
  • 00:11:49
    say for example that we are having four
  • 00:11:51
    processors in your system but even
  • 00:11:54
    though you are having four processors
  • 00:11:56
    since your process is a single threaded
  • 00:11:59
    process that single process can run only
  • 00:12:02
    on one of the processor so no matter how
  • 00:12:05
    many number of processors you may have
  • 00:12:08
    having a single threaded process will
  • 00:12:10
    run in only one of the processor but if
  • 00:12:13
    we are having a multi-threaded approach
  • 00:12:16
    then what will happen is that we know
  • 00:12:18
    that in a multi thread approach each
  • 00:12:20
    process has different number of threads
  • 00:12:22
    associated with it so each of these
  • 00:12:25
    threads can run on these multiple
  • 00:12:28
    processors let's say that there are four
  • 00:12:30
    threads associated with a single process
  • 00:12:32
    and then there are four processors in
  • 00:12:35
    your system so each of those threads can
  • 00:12:38
    run on one of the processors so thread
  • 00:12:40
    one may run in processor one thread to
  • 00:12:42
    myrin processor two and so on so in this
  • 00:12:45
    way we see that the tasks are executed
  • 00:12:48
    concurrently at the same time and hence
  • 00:12:50
    the process will be completed quicker
  • 00:12:52
    and it is more efficient that is what we
  • 00:12:55
    mean by utilization of multi processor
  • 00:12:57
    architecture so if we don't have a
  • 00:12:59
    multi-threaded approach even though we
  • 00:13:01
    have multi processors it is not very
  • 00:13:03
    useful because we are not able to make
  • 00:13:06
    use of the multi processor architecture
  • 00:13:08
    but in case of multi-threading we are
  • 00:13:10
    able to make use of the multiprocessor
  • 00:13:12
    architecture hence improving our
  • 00:13:14
    efficiency of computing so these are the
  • 00:13:17
    main benefits of multi
  • 00:13:19
    programming so with this I hope you got
  • 00:13:21
    the basic idea of threads and what is
  • 00:13:24
    the meaning of single threaded process
  • 00:13:25
    and multi thread process and we also saw
  • 00:13:28
    that most of the systems now we follow a
  • 00:13:31
    multi-threaded approach and we see that
  • 00:13:33
    how a single process can have multiple
  • 00:13:36
    threads associated with it and how they
  • 00:13:38
    share the resources of that process to
  • 00:13:40
    which they belong and how these things
  • 00:13:43
    are beneficial for our computing so
  • 00:13:45
    we'll be studying more things about
  • 00:13:47
    threads in the coming lectures so for
  • 00:13:49
    now I hope the basics of thread is clear
  • 00:13:51
    to you
  • 00:13:52
    so thank you for watching and see you in
  • 00:13:53
    the next one
  • 00:13:55
    [Applause]
  • 00:13:57
    [Music]
标签
  • threads
  • operating systems
  • process management
  • multi-threading
  • CPU utilization
  • resource sharing
  • multi-processor architectures
  • responsiveness
  • computing efficiency