YT YouTube2Text

YouTube2Text → Transcripts → 03 04 Processes Part 4

03 04 Processes Part 4 — Transcript

by Santelmo · 1,726 words · 387 segments · language en · Watch on YouTube

Open in transcriber Summarize with AI Download SRT/VTT

Full transcript

0:08[Music]
0:19so the process within a system may be
0:21independent or cooperating
0:24okay
0:25so when you say cooperating
0:27these are processed okay so it can
0:29affect or be affected by
0:32other processes including sharing data
0:35so the reasons for cooperating processes
0:37are information sharing computation
0:39speed up
0:40modularity and
0:42convenience okay so cooperating process
0:46needs the inter-process communication or
0:49ipc
0:50which
0:51are represented using the two models
0:54also which are the shared memory and
0:56the message passing
0:59okay
1:00now
1:01when you say independent process
1:03so operating concurrently on the systems
1:06are those that can neither
1:08affect other processes or be affected by
1:10the other processes as mentioned here
1:14okay
1:14now
1:15under three zones
1:17what is information sharing in operating
1:19process so there may be several
1:21processes which need access to the same
1:24file for example okay example are
1:27pipelines
1:28okay
1:29so computation speed up
1:32so often
1:33a solution to a problem can be solved
1:36faster if the problem can be broken down
1:38into sub tasks
1:41okay
1:42to be solved simultaneously so
1:43particularly when multiple processors
1:46are involved here
1:48okay
1:48modularity
1:50the most efficient architecture
1:53may be to break the system down into
1:55cooperating modules so example databases
1:58with a client server architecture
2:00okay and the last one would be
2:02convenience
2:03okay what is convenience in operating
2:06process
2:07so even a single user may be
2:10multitasking
2:11such as editing compiling printing and
2:13running at the same code
2:15in different windows so that's
2:17convenience
2:19okay
2:20now let's see the model mentioned here
2:22the shared and the message passing
2:26okay
2:27now cooperating process requires some
2:31uh type of the inter-process
2:33communication which is mostly one of
2:36these two okay the shared memory
2:39or the message passing systems
2:42okay now is the difference between these
2:44two
2:45when you say shared memory this is
2:47faster once
2:49it is set up
2:50so because no system calls are required
2:53and access occurs at a normal memory
2:55speeds
2:56so however
2:58it is more complicated to set up and it
3:01doesn't work as well across multiple
3:04computers
3:05so shared memory is generally preferable
3:08when large amounts of information
3:10must be shared quickly
3:12on the same computer
3:14okay
3:16now for the message passing it requires
3:18system calls for every message transfer
3:21okay
3:23and it's therefore slower but it is
3:25simpler to set up and works well with
3:29multiple computers or across multiple
3:31computers
3:32so message passing is generally
3:34preferable
3:35when the amount and or frequency of data
3:38transfer is small
3:39or when multiple computers are involved
3:43okay so that is when the message passing
3:46is
3:48useful
3:50okay
3:52next
3:53let's talk about the producer consumer
3:55problem
3:56okay so the paradigm for operating
3:59process
4:01so the producer process
4:04produces information
4:05that is consumed by the consumer process
4:09so there are two variations here okay
4:11it could be the unbounded proper a
4:14buffer or the bounded buffer
4:17okay
4:18so for these two variations when you say
4:20unbounded buffer
4:22there is no practical limit on the size
4:24of the buffer
4:25so the producer never waits and the
4:28consumer waits if there is no buffer to
4:31consume
4:33okay
4:35on the bounded buffer
4:37it assumes that there is a fixed buffer
4:39size so the producer must wait if all
4:42buffers are full
4:44the consumer wage if there is no buffer
4:47to consume
4:48so this is a classic example
4:51in which one process is producing data
4:54and another process is consuming data
4:58okay
5:00so
5:01in this example in the order in which it
5:03is produced although
5:05that could vary okay so the data is
5:08passed via an intermediary buffer
5:11which may be either
5:13bounded
5:15or unbounded
5:16okay so with bounded buffer the producer
5:19may have to wait until there is a space
5:22available
5:24okay but when
5:26an unbounded buffer the producer the
5:29producer will never need to wait
5:32okay so the consumer may need to wait in
5:35either case until there is a data
5:37available
5:39okay so that is producer
5:41consumer problem
5:44okay so we will be using this
5:46producer consumer problem on the
5:49discussion of deadlocks
5:52okay
5:54now the inter-process communication for
5:55the shared memory okay so an area of
5:59memory shared among the process that
6:01wish to communicate so the communication
6:04is under the control of the users
6:06okay
6:07major issues to provide mechanisms that
6:10will allow the user processes
6:12to synchronize their action when they
6:14access shared memory
6:16so synchronization is discussed in great
6:19details in chapter six and seven
6:22right
6:24next
6:26how about the bounded buffer the shared
6:28memory solution so this example uses a
6:30shared memory and circular queue so note
6:34in the code okay
6:37that only the producer changes in
6:41okay and only the consumer changes out
6:46okay
6:47so and they can never be accessing the
6:49same array location at the same time so
6:52first the following data is set up
6:55in the shared memory area
6:58okay so the solution is correct but it
7:02can
7:03only use a buffer size -1 elements
7:07okay
7:10next
7:12then the producer process
7:15note that the buffer is full when in
7:18is one less than out in circular sense
7:23okay
7:27next
7:28the consumer process note that the
7:30buffer is empty
7:32when in is equal to out
7:35okay
7:36so
7:37[Music]
7:39next
7:40what about filling all the buffers would
7:42that be possible okay so so suppose that
7:45we wanted to provide a solution to the
7:47consumer producer problem that fills
7:51all the buffers
7:53what can we do so
7:55or we can do so by having an integer
7:57counter that keeps track
7:59of the number of full buffers so
8:01initially the counter is set to zero
8:03the integer counter is incremented by
8:05the producer after it produces a new
8:07buffer
8:09and the integer counter is and is
8:11decremented by the consumer after it
8:13consumes the buffer
8:16okay
8:18so you'll have this
8:20produce an item in the next
8:22produced
8:24this is for the producer
8:26okay
8:27and for the consumer
8:30you'll have this so if there is a
8:33resource available then it needs to be
8:35consumed
8:36okay
8:37so the next one would be the race
8:39condition so what is a race condition
8:42any time there is or there are two armor
8:44processors or threads operating on
8:46currently
8:48there is a potential for a particularly
8:50difficult class of problems
8:53we call it race condition
8:56okay
8:57so the identifying characteristics of
8:59phrase conditions is that the
9:01performance varies depending on which
9:04process or thread executes their
9:06instructions
9:07before
9:08the other one and this becomes a problem
9:12when the program runs correctly in some
9:14instance and incorrectly in others
9:17okay so race conditions are notoriously
9:20difficult to debug because they are
9:22unpredictable
9:24and repeatable and may not exhibit
9:26themselves for years
9:29okay
9:36okay so
9:37next
9:38note that the above solution also checks
9:40the return value from
9:42the read system call so to verify the
9:45number of characters
9:47read is equal to the number expected
9:50okay
9:51some of those checks were actually in
9:52the original code but
9:55they were omitted from the nodes for
9:57clarity okay
9:59and the real code also uses select
10:02before reading
10:04okay so to verify
10:06that there are characters present to
10:08read
10:10and to delay if not
10:12okay
10:13so note also that this problem could not
10:16be easily solved using synchronization
10:19tools
10:20that we're going to discuss in chapter
10:22six because the problem is not really
10:25one of two processes accessing the same
10:27data at the same time
10:30okay so these problems here about
10:33synchronization will be discussed in
10:35detail and thoroughly on the next
10:38few chapters the next two chapters okay
10:41specifically
10:44okay so next would be the
10:47inter-process communication the message
10:49passing
10:50message passing systems must support a
10:52minimum system calls for send
10:54message and receive message
10:57so you'll have this
10:58all right so these are the ipc facility
11:02operations
11:04okay
11:05so process communicate with each other
11:07without resorting to shared variables
11:10okay so the message size is either fixed
11:13or
11:14variable
11:18okay
11:19so a communication link okay so take
11:21note that we have your established
11:23communication link okay so a
11:25communication link must be established
11:27between the operating processes
11:30before messages can be sent
11:33so there are three key issues to be
11:35resolved in message passing systems as
11:38further explored in the next three
11:40subsections okay so we have the direct
11:43or indirect communications
11:45naming
11:46synchronous or asynchronous
11:48communication and automatic or
11:51explicit buffering
11:53okay so this summarizes this issues here
11:58now let's talk about the implementations
12:01of the communication link
12:02so it's either implemented in physical
12:05or logical
12:06okay in physical
12:08we have the shared memory the hardware
12:10bus and the network
12:12implementation
12:13logically we could have direct or
12:15indirect
12:16synchronous or asynchronous and
12:19automatic or explicit buffering
12:22okay now
12:24with direct communication
12:27okay the sender must know the name of
12:29the receiver to which it wishes to send
12:33message
12:34so for symmetric communication the
12:36receiver must also know the specific
12:39name of the sender
12:40from which it wishes to receive messages
12:44well for us asymmetric
12:47there is or this is not necessary
12:51okay
12:53now for indirect communications all
12:56right so indirect communication uses
12:59shared mailboxes
13:00or ports all right
13:04so
13:05multiple processes
13:07can share
13:08that
13:09mailbox or boxes
13:12so only one process can read any given
13:15message in a mailbox so initially
13:18process that creates the mailbox is the
13:20owner
13:21and is the only one allowed to read mail
13:24in that mailbox although
13:27this privilege may be transferred
13:29okay
13:31so of course
13:32the process that reads the message can
13:34immediately turn around and place an
13:36identical message back in the box for
13:38someone else to read
13:40but that may put it
13:42at the back end of the queue messages
13:46okay so the os provides system calls to
13:49create and delete mailboxes and to send
13:51and receive messages to or from
13:53the mailboxes
13:56all right so that's indirect
13:57communication
14:00okay so this is what i mentioned earlier
14:05now
14:08next is
14:10how about mailbox sharing
14:11as mentioned earlier for example you've
14:13got p1 p2 and p3 shear mailbox a so p1
14:18sends p2 and p3 receive who gets the
14:21message
14:23okay
14:24so solution would be
14:26allow a link to be associated with most
14:28or at most two processes
14:31allow only one process at the time to
14:33execute a receive operation
14:37allow the system to select arbitrarily
14:40the receiver
14:41sender is not or is notified who the
14:44receiver was
14:46okay that's indirect
14:48communication
14:58[Music]
15:06you

About this transcript

This page contains the full transcript of 03 04 Processes Part 4 by Santelmo, generated from the public captions YouTube serves with the video. The transcript has 1,726 words across 387 segments, with the original timestamps preserved so you can click any line to jump to that moment in the embedded player.

What you can do with it

Use the transcript to take notes, quote the speaker, build a study guide, generate a summary with ChatGPT or Claude via the YouTube Summary tool, or export it as a timed subtitle file with YouTube to SRT. You can also re-open it in the transcriber to translate the transcript into 100+ languages.

Free YouTube transcript tool

YouTube2Text is a free YouTube transcript generator — no signup, no daily limit. Paste any YouTube link and get the full transcript instantly, with timestamps, click-to-jump, translation to 100+ languages, AI prompts for ChatGPT, Claude, and Gemini, and exports to TXT, SRT, VTT, or Markdown.