->slos;

Distributed Synchronization and Message Passing¶

1. Explain why shared-memory synchronization primitives (semaphores, locks, monitors) are inapplicable on a distributed multiprocessor, and identify message passing as the correct alternative.

2. Distinguish between synchronous message passing (no buffer storage: sender and receiver must both be ready) and asynchronous message passing (buffered: sender can proceed before receiver is ready), and trace the timing of each scenario.

3. Describe the Actor model of message passing and identify at least two industry languages (Erlang, Scala) that support it, including real-world use cases.

Ada Tasks and Rendezvous¶

4. List the three characteristics of an Ada task: its own thread of control, its own execution state, and mutually exclusive entry procedures.

5. Explain Ada’s rendezvous mechanism: how a calling task and an accepting task synchronize at an entry procedure call, including which task waits when only one is ready.

6. Trace the sequence of events in a rendezvous — argument passing, suspension of the caller, execution of the entry body, return of out-parameters, and resumption of both tasks.

7. Explain why Ada’s rendezvous does not require shared memory and can therefore work on uniprocessors, tightly-coupled multiprocessors, and distributed systems alike.

Concurrency Patterns in Ada¶

8. Implement a divide-and-conquer parallel pattern in Ada using task type with entry procedures, and calculate the expected speedup relative to a sequential solution.

9. Describe the structure of Ada’s capacity-1 bounded buffer task type and explain how strict alternation between accept put and accept get enforces mutual exclusion without explicit locks.

10. Explain Ada’s select-when construct and demonstrate how it allows a capacity-N buffer task to conditionally accept put() or get() calls based on buffer state.

MPI¶

11. Describe MPI (Message Passing Interface) as an industry-standard distributed-memory parallel computing library, identify its origin (1994 consortium), and name at least two languages it supports natively.

Comparing Concurrency Constructs¶

12. Compare monitors, tasks/threads, and coroutines along the dimensions of “has its own thread” and “has its own execution state,” and give an example language for each.

13. Define a coroutine and explain how two coroutines share a single thread while maintaining separate execution states through explicit resume calls.

14. Classify the key concurrency entities across languages: processes (Smalltalk, MPI), tasks (Ada), and threads (C/C++, Java, Go, Python, Ruby, Scala), and explain what they have in common.

15. Summarize the progression of shared-memory synchronization constructs — semaphore → lock/condition variable → monitor — in terms of abstraction level and ease of correct use.