Week 12 - Concurrency

Question 1

At how many levels can concurrency occur?

Answer 1

Concurrency can occur at four levels: machine instruction level, high-level language statement level, unit level, and program level.

Question 2

What is the primary goal of developing concurrent software?

Answer 2

The primary goal is to produce scalable and portable concurrent algorithms to make use of increasing numbers of processors.

Question 3

What are the categories of concurrency?

Answer 3

Physical concurrency (involving multiple independent processors), logical concurrency (achieved by time-sharing one processor), and coroutines (single-threaded concurrency).

Question 4

What role does a scheduler play in concurrent systems?

Answer 4

The scheduler delays task execution and maps task execution onto available processors, ensuring fair allocation of resources and efficient task execution.

Question 4

Describe the possible states of task execution.

Answer 4

The possible states include new (created but not started), ready (waiting for execution), running, blocked (waiting for an event), and dead (completed execution).

Question 4

Why is concurrency widely used in modern computing?

Answer 4

Concurrency allows for faster execution, even on single processors, and enables efficient utilization of multiprocessor systems. It also facilitates the design of software for real-world situations involving concurrency.

Question 5

How do tasks differ from ordinary subprograms in subprogram-level concurrency?

Answer 5

Tasks may be implicitly started, can execute concurrently with other program units, and may not return control to the caller immediately upon completion.

Question 6

Explain the need for competition synchronization with an example.

Answer 6

Competition synchronization ensures the correct outcome when multiple tasks compete for shared resources. For example, without proper synchronization, race conditions can lead to unpredictable results.

Question 6

What is task synchronization, and why is it necessary?

Answer 6

Task synchronization is a mechanism controlling the order of task execution, essential for ensuring correct and coordinated access to shared data among concurrent tasks.

Question 7

Define liveness and deadlock in the context of concurrency.

Answer 7

Liveness refers to the property of a program unit to eventually complete its execution. Deadlock occurs when all tasks lose liveness, resulting in a situation where no progress can be made.

Question 8

What are the key design issues to consider in concurrent systems?

Answer 8

Design issues include competition and cooperation synchronization, task scheduling control, task lifecycle management, and task creation and termination.

Question 9

Name three methods for providing synchronization in concurrent systems.

Answer 9

Semaphores, monitors, and message passing are three methods used for providing synchronization in concurrent systems.

Question 10

Describe the purpose of semaphores in concurrent systems.

Answer 10

Semaphores are used to manage access to shared resources and synchronize the execution of concurrent processes or threads.

Question 11

Explain the typical operations performed using semaphores in concurrent systems.

Answer 11

The DEPOSIT operation is used to indicate the availability of resources or space, while the FETCH operation is used to indicate the need for resources or data.

Question 12

What are the fundamental operations associated with semaphores?

Answer 12

The wait operation is used to decrement the semaphore’s value and potentially block a process, while the release operation is used to increment the semaphore’s value and potentially wake up a blocked process.

Question 13

How are producer and consumer tasks synchronized using semaphores?

Answer 13

Producers wait for available space before depositing data, while consumers wait for available data before fetching it. Semaphores are used to control access to shared buffers or resources.

Question 14

What is competition synchronization, and how is it achieved using semaphores?

Answer 14

Competition synchronization ensures that only one process or thread can access a shared resource at a time. This is typically achieved using binary semaphores, where access to the resource is controlled by acquiring and releasing the semaphore.

Question 15

What are the potential consequences of misusing semaphores?

Answer 15

Misuse of semaphores can lead to synchronization issues such as deadlock or livelock, where processes or threads become unable to make progress due to improper resource management or synchronization mechanisms.

Question 16

What asymmetry exists in task communication regarding message passing?

Answer 16

While a task sending a message needs to know the entry name in the receiving task, the receiving task’s entry name does not need to know the sender’s name.

Question 17

What happens when a task has multiple accept clauses with nonempty entry queues?

Answer 17

If more than one entry queue is nonempty, the system chooses one nondeterministically to accept a message.

Question 18

How can tasks have more than one entry point?

Answer 18

Tasks can define multiple entry points, each specified with its own entry clause.

Question 19

How does the system support cooperation synchronization with message passing using guarded accept clauses?

Answer 19

Guarded accept clauses allow specifying conditions under which a task can accept messages, enabling cooperation synchronization.

Question 20

What are protected objects, and how do they enhance concurrency?

Answer 20

Protected objects offer a more efficient way to implement shared data structures, providing mutually exclusive access to shared objects.

Question 20

How does the system model mutually exclusive access to shared data using message passing?

Answer 20

Shared data operations within tasks are encapsulated, and competition synchronization is implicitly ensured through the semantics of accept clauses.