Important 10-Mark Questions (Module 1)
1. Explain the classification of parallel computers based on Flynn’s taxonomy.
2. Differentiate between SISD, SIMD, MISD, and MIMD systems with examples.
3. Explain the architecture and working principle of SIMD systems.
4. Discuss the characteristics, advantages, and limitations of MIMD systems.
5. Write short notes on Interconnection networks used in parallel computers (Bus,
Crossbar, Multistage).
6. Compare Shared-memory vs Distributed-memory systems with neat diagrams.
7. Explain Cache coherence problems and techniques used to maintain it.
8. Describe the role of processes and threads in parallel programming.
9. Explain synchronization issues in shared-memory systems with examples.
10. Discuss the need for coordinating processes/threads in parallel systems.
Module 2 – Important 10-Mark Questions
1. Explain the architecture of MIMD systems. Discuss how MIMD systems achieve
parallelism with suitable examples.
2. Describe the key differences between SIMD and MIMD architectures. Explain
how MIMD systems support multiple threads and processes.
3. What is GPU programming? Explain the structure of a GPU, CUDA cores, and thread
hierarchy in GPU programming.
4. Discuss the concept of hybrid parallel systems. How does combining CPU and GPU
processing enhance performance?
5. Explain performance metrics used in parallel programming. Derive expressions
for speedup and efficiency in MIMD systems.
6. State and derive Amdahl’s Law. Explain how it affects the performance improvement
of parallel programs.
7. What is scalability in parallel systems? Differentiate between strong scalability and
weak scalability with examples.
8. Describe the process of measuring execution time in MIMD programs. What factors
affect timing accuracy?
9. Explain GPU performance optimization techniques. Discuss how memory access
patterns and thread organization affect GPU performance.
10. . Compare CPU and GPU architectures. Highlight their differences in terms of
design, instruction execution, and performance characteristics
11. Discuss the challenges and advantages of programming hybrid CPU–GPU systems.
Provide real-world examples.
12. Explain the relationship between Amdahl’s Law and scalability. How can
parallel efficiency be maintained as system size increases?
13. What are the main performance bottlenecks in GPU-based systems? How can they be
reduced?
14. Explain the methods for evaluating the performance of MIMD programs.
Include timing models and experimental analysis.
15. Describe the importance of workload balancing in hybrid parallel systems. How does
imbalance affect performance?
Module -3 (till collective communication)
16. Explain the Message Passing Interface (MPI) model. Discuss its importance in
distributed memory programming and list the major features of MPI.
17. List and explain the different categories of MPI functions. Give examples of
functions used for process management, point-to-point communication, and collective
operations.
18. Explain the working of point-to-point communication in MPI using MPI_Send() and
MPI_Recv() functions. Illustrate with a simple example program.
19. Explain how numerical integration using the trapezoidal rule can be parallelized using
MPI. Write the steps involved and discuss how data is distributed among processes.
20. Design a parallel algorithm for computing definite integrals using the
Trapezoidal Rule in MPI.
Explain initialization, data distribution, computation, and result collection phases.
21. Discuss the challenges in input/output (I/O) operations in MPI programs.
Explain how distributed processes handle I/O and share data efficiently.
22. What is collective communication in MPI? Explain the working of any three
collective communication functions such as MPI_Bcast(), MPI_Reduce(), and
MPI_Gather().
23. Differentiate between collective communication and point-to-point communication in
MPI. Give examples of situations where each is preferable.
24. Explain how performance and scalability of an MPI program can be evaluated.
Discuss how communication overhead affects performance.
25. Describe the typical steps involved in writing and executing an MPI program.
Include initialization, communicator usage, data exchange, synchronization, and
finalization.
26. Explain the concept of derived data types in MPI.
How do they help in improving the efficiency of message passing?
27. Explain the use of reduction operations in MPI using MPI_Reduce() andMPI_Allreduce().
Provide an example to demonstrate their use.
28. Discuss the techniques for parallel file I/O in MPI.
How do functions like MPI_File_read and MPI_File_write help in handling distributed
data?
29. Explain how synchronization between MPI processes is achieved.
Discuss the role of barriers and collective calls in maintaining data consistency.