In Assumptions on Timing - I don’t understand this concept of a book

Asked

Viewed 68 times

4

Cited in the book that a critical region has 4 key points:

  1. Progress:

Ensures that all threads are entering and leaving the critical region, avoiding deadlocks.

  1. Mutually Exclusive:

Only one thread may be acting in the critical region, the others must wait their turn.

  1. Bounded Waiting:

Ensures that all thread after a time limit is allowed within the critical region. Preventing a thread I stayed a long time without getting progress for lack of "luck" in the selection.

  1. No Assumptions on Timing:

It didn’t make sense to me. What does that mean?

Also, if any concept prior to the room is incorrect or incomplete, please correct me.

java multithreading competition thread-safety

  • There are good questions there, and although they are even related, they are not the same problem, edit this one for something more focused and keep the other questions separate can help well and get good answers (I see at least 3 different questions, although all of them are about competition).

  • 1

    edited! to so confused that I don’t know where to start

1 answer

5


This means that you cannot consider when things are going to happen. They will happen when they have to, in the order the application determines to be best and possible in the current processing flow and it will take as long as necessary to complete the task.

When you have competition it is common that the total execution or completion time is not deterministic (many cases can make the time not deterministic). You lose a little control of the execution order and how much processing time will be devoted to another task in the middle of the execution of yours. During execution your processing may be stopped temporarily (or indefinitely if a deadlock).

The start of your processing can take a long time to happen by joining some queue or because it depends on information that is not yet available. In this context the term probably mainly indicates these cases.

In distribution it complicates more because you don’t even know how long it will take the other machine to do something, if it has a queue, etc.

  • was clear, thank you. Regarding the other terms, my explanation was consistent?

  • From what I understand, yes, of course I didn’t see in the exact context that they were used, but it seems okay.

Browser other questions tagged java multithreading competition thread-safety

You are not signed in. Login or sign up in order to post.