PARALLEL AND DISTRIBUTED

COMPUTING
22CS4106R

STUDENT ID: ACADEMIC YEAR: 2024-25

STUDENT NAME:
Table of Contents

1. Session 01: Introductory Session

2. Session 02: Demonstrate various programs in OpenMP, Parallel global 1
sum Parallel Matrix-Vector Multiplication with OpenMP
3. Session 03: Parallel Computation with OpenMP: Summation Techniques 7
and Matrix-Vector Multiplication
4. Session 04: Parallel implementation of PI, Fundamentals of MPI 13
Programming
5. Session 05: Parallel Computation and Sorting Techniques Using MPI and 21
OpenMP
6. Session 06: Exploring Parallel Sorting Algorithms: Implementations and 29
Analysis with MPI, OpenMP
7. Session 07: Advanced Parallel Computing in Mathematical Algorithms: 38
Harnessing OpenMP, and MPI for Matrix Operations, Integration, and
Sorting
8. 47
Session 08: Distributed Algorithms and Parallel Programming with MPI
9. Session 09: Distributed Mutual Exclusion and Parallel Programming 57
10. Session 10: Leader Election in Distributed Systems: Implementing Bully 67
and Ring Algorithms Using MPI/C Library
11. Session 11: Exploring Parallel Computing: Data Structures, Algorithmic 75
Patterns, and VisualizationTechniques
12. Session 12: Simulation and Distributed Programming: Consensus and 83
Matching Algorithms
13. Session 13: Distributed Systems and Cloud Computing: Web Services, 91
MPI, and HPC Workflows
 A.Y. 2024-25 LAB CONTINUOUS EVALUATION

S.No. Date Experiment Name Pre- In-Lab (25M) Post- Viva Total Faculty
Lab Program/ Data and Analysis & Lab Voce (50M) Signature
(10M) Procedure Results Inference (10M) (5M)
(5M) (10M) (10M)

Demonstrate various programs in

1.
OpenMP, Parallel global sum Session
Parallel Matrix-Vector Multiplication
with OpenMP
Parallel Computation with OpenMP:
2. Summation Techniques and Matrix-
Vector Multiplication
Parallel implementation of PI,
3. Fundamentals of MPI Programming

Parallel Computation and Sorting

Techniques Using MPI and OpenMP
4.
Exploring Parallel Sorting
5. Algorithms: Implementations and
Analysis with MPI, OpenMP.
Advanced Parallel Computing in
Mathematical Algorithms: Harnessing
6. OpenMP, and MPI for Matrix
Operations, Integration, and Sorting

Distributed Algorithms and Parallel

7. Programming with MPI
S.No. Date Experiment Name Pre- In-Lab (25M) Post- Viva Total Faculty
Lab Program/ Data and Analysis & Lab Voce (50M) Signature
(10M) Procedure Results Inference (10M) (5M)
(5M) (10M) (10M)
Distributed Mutual Exclusion and
8. Parallel Programming

Leader Election in Distributed

Systems: Implementing Bully and
9. Ring Algorithms Using MPI/C
Library
Exploring Parallel Computing: Data
10. Structures, Algorithmic Patterns, and
VisualizationTechniques
Simulation and Distributed
11. Programming: Consensus and
Matching Algorithms
Distributed Systems and Cloud
Computing: Web Services, MPI, and
12. HPC Workflows
 Experiment # <TO BE FILLED BY STUDENT> Student ID <TO BE FILLED BY STUDENT>
Date <TO BE FILLED BY STUDENT> Student Name <TO BE FILLED BY STUDENT>

Experiment 1
Title: Demonstrate various programs in OpenMP, Parallel global sum
Aim/Objective:
Demonstrate the versatility of OpenMP through multiple programs implementing parallel global sum
algorithms, emphasizing diverse parallelization approaches for optimal performance.
Description:
Explore OpenMP's capabilities by presenting distinct programs showcasing parallel global sum
computations, utilizing different OpenMP directives and strategies to achieve efficient parallelization.
Pre-Requisites:
Participants should have a solid understanding of C programming, familiarity with parallel computing
concepts, and access to a compiler with OpenMP support. A basic knowledge of multithreading and a
multiprocessor system for parallel execution are also recommended.

Pre-Lab:

1. Explain the purpose of the critical directive in OpenMP and provide a scenario where it is
beneficial?

2. Describe the role of the master directive in OpenMP. In what situations is it commonly
used?

Course Title PARALLEL & DISTRIBUTED COMPUTING ACADEMIC YEAR: 2024-25

Course Code(s) 22CS4106R Page 1 of 103
 Experiment # <TO BE FILLED BY STUDENT> Student ID <TO BE FILLED BY STUDENT>
Date <TO BE FILLED BY STUDENT> Student Name <TO BE FILLED BY STUDENT>

3. Discuss the significance of the default clause in OpenMP. How does it affect the sharing
of variables in a parallel region?

4. What is the purpose of the threadprivate directive in OpenMP? Provide an example of a

situation where using this directive is advantageous.

5. Explain the concept of the reduction clause in OpenMP. How does it facilitate parallel
computation?

Course Title PARALLEL & DISTRIBUTED COMPUTING ACADEMIC YEAR: 2024-25

Course Code(s) 22CS4106R Page 2 of 103
 Experiment # <TO BE FILLED BY STUDENT> Student ID <TO BE FILLED BY STUDENT>
Date <TO BE FILLED BY STUDENT> Student Name <TO BE FILLED BY STUDENT>

In-Lab:

1. Basic programs to demonstrate various #pragrams in OpenMP: Parallel • for • private

• shared • critical • MASTER Directive • THREADPRIVATE Directive • DEFAULT
Clause • FIRSTPRIVATE Clause •LASTPRIVATE Clause.
o /* Program to Demonstrate #pragrams in OpenMP: Parallel, for */
o /* Program to Demonstrate private clause */
o /* Program to Demonstrate shared clause */
o /* Program to Demonstrate Critical Directive */
o /* Program to Demonstrate Master Directive */
o /* Program to Demonstrate threadprivate Clause */
o /* Program to Demonstrate default Clause */
o /* Program to Demonstrate firstprivate clause */
o /* Program to Demonstrate lastprivate clause */

• Program:

Course Title PARALLEL & DISTRIBUTED COMPUTING ACADEMIC YEAR: 2024-25

Course Code(s) 22CS4106R Page 3 of 103
Experiment # <TO BE FILLED BY STUDENT> Student ID <TO BE FILLED BY STUDENT>
Date <TO BE FILLED BY STUDENT> Student Name <TO BE FILLED BY STUDENT>

2. Write programs to demonstrate SINGLE Directive BARRIER Directive ATOMIC

Directive
/* Program to Demonstrate single Directive */
/* Program to demonstrate barrier Directive */
/* Program to Demonstrate atomic Directive */

• Program:

Course Title PARALLEL & DISTRIBUTED COMPUTING ACADEMIC YEAR: 2024-25

Course Code(s) 22CS4106R Page 4 of 103
 Experiment # <TO BE FILLED BY STUDENT> Student ID <TO BE FILLED BY STUDENT>
Date <TO BE FILLED BY STUDENT> Student Name <TO BE FILLED BY STUDENT>

Sample VIVA-VOCE Questions (In-Lab):

 What does the SINGLE directive in OpenMP accomplish, and in what scenarios would
you use it?
 Explain the purpose of the BARRIER directive in OpenMP and why it is essential in
parallel programming.
 What is the role of the ATOMIC directive in OpenMP? Provide an example where
using ATOMIC is crucial?
 How does the BARRIER directive contribute to ensuring correct synchronization in
parallel programs?
 In what situations would you prefer using the SINGLE directive over the MASTER
directive, and vice versa?

Course Title PARALLEL & DISTRIBUTED COMPUTING ACADEMIC YEAR: 2024-25

Course Code(s) 22CS4106R Page 5 of 103
 Experiment # <TO BE FILLED BY STUDENT> Student ID <TO BE FILLED BY STUDENT>
Date <TO BE FILLED BY STUDENT> Student Name <TO BE FILLED BY STUDENT>

Post-Lab:
1. Implement a program to demonstrate Parallel global sum using OpenMP Reduction clause

• Program:

Evaluator Remark (if Any):

Marks Secured: out of 50

Signature of the Evaluator with Date

Course Title PARALLEL & DISTRIBUTED COMPUTING ACADEMIC YEAR: 2024-25

Course Code(s) 22CS4106R Page 6 of 103
 Experiment # <TO BE FILLED BY STUDENT> Student ID <TO BE FILLED BY STUDENT>
Date <TO BE FILLED BY STUDENT> Student Name <TO BE FILLED BY STUDENT>

Experiment 2
Title: Parallel Computation with OpenMP: Summation Techniques and Matrix-Vector
Multiplication
Aim/Objective:
Illustrate parallel computation techniques using OpenMP for two fundamental algorithms—
summation and matrix-vector multiplication—aiming to enhance performance through
concurrent processing.
Description:
Implement OpenMP-based programs for parallel summation and matrix-vector multiplication,
exploring various OpenMP directives and strategies to optimize computation speed and
scalability.

Pre-Requisites:
Proficiency in a language supporting OpenMP (e.g., C/C++), understanding of OpenMP
directives, familiarity with array and matrix manipulation, and a foundational grasp of parallel
programming concepts including thread management, synchronization, and parallelization
strategies.
Pre-Lab:
1. What is the primary objective of utilizing OpenMP in parallel summation techniques?

2. Briefly discuss one advantage of using OpenMP for parallel summation techniques
compared to sequential approaches.

Course Title PARALLEL & DISTRIBUTED COMPUTING ACADEMIC YEAR: 2024-25

Course Code(s) 22CS4106R Page 7 of 103
 Experiment # <TO BE FILLED BY STUDENT> Student ID <TO BE FILLED BY STUDENT>
Date <TO BE FILLED BY STUDENT> Student Name <TO BE FILLED BY STUDENT>

3 In the context of matrix-vector multiplication, define the term "data parallelism" and explain
how OpenMP leverages this concept for parallel computation.

4. Explain the role of load balancing in the context of parallel matrix-vector multiplication
using OpenMP, and why it is crucial for optimizing performance.

5. Name one key OpenMP directive used in the parallelization of computations. Provide a brief
description of its purpose?

Course Title PARALLEL & DISTRIBUTED COMPUTING ACADEMIC YEAR: 2024-25

Course Code(s) 22CS4106R Page 8 of 103
 Experiment # <TO BE FILLED BY STUDENT> Student ID <TO BE FILLED BY STUDENT>
Date <TO BE FILLED BY STUDENT> Student Name <TO BE FILLED BY STUDENT>

In-Lab:
1. Implement Parallel Summation using OMP - The Array Element Sum Problem, Tree
structure global sum - Parallel-tree-sum.c

• Program:

Course Title PARALLEL & DISTRIBUTED COMPUTING ACADEMIC YEAR: 2024-25

Course Code(s) 22CS4106R Page 9 of 103
 Experiment # <TO BE FILLED BY STUDENT> Student ID <TO BE FILLED BY STUDENT>
Date <TO BE FILLED BY STUDENT> Student Name <TO BE FILLED BY STUDENT>

2. Implement a program to demonstrate Parallel prefix sum using OpenMP - OMP Parallel
prefix sumfinal.c
• Program:

Course Title PARALLEL & DISTRIBUTED COMPUTING ACADEMIC YEAR: 2024-25

Course Code(s) 22CS4106R Page 10 of 103
 Experiment # <TO BE FILLED BY STUDENT> Student ID <TO BE FILLED BY STUDENT>
Date <TO BE FILLED BY STUDENT> Student Name <TO BE FILLED BY STUDENT>

Sample VIVA-VOCE Questions (In-Lab):

1. Explain the concept of parallel summation using OpenMP in the context of the array element
sum problem
2. How does the #pragma omp parallel directive contribute to the parallelization of the array
element sum problem?
3. Why is a critical section (#pragma omp critical) used in the parallel tree-sum implementation,
and what would happen if it were omitted?
4. How can the performance of the parallel tree-sum algorithm be influenced by the number of
threads used in the computation?
5. Explain the purpose of the reduction clause in OpenMP and discuss whether it could be
applied to improve the parallel tree-sum implementation.

Course Title PARALLEL & DISTRIBUTED COMPUTING ACADEMIC YEAR: 2024-25

Course Code(s) 22CS4106R Page 11 of 103
 Experiment # <TO BE FILLED BY STUDENT> Student ID <TO BE FILLED BY STUDENT>
Date <TO BE FILLED BY STUDENT> Student Name <TO BE FILLED BY STUDENT>

Post-Lab:
1. Implement Parallel matrix vector multiplication using OpenMP
• Program:

Evaluator Remark (if Any):

Marks Secured: out of 50

Signature of the Evaluator with Date

Course Title PARALLEL & DISTRIBUTED COMPUTING ACADEMIC YEAR: 2024-25

Course Code(s) 22CS4106R Page 12 of 103
 Experiment # <TO BE FILLED BY STUDENT> Student ID <TO BE FILLED BY STUDENT>
Date <TO BE FILLED BY STUDENT> Student Name <TO BE FILLED BY STUDENT>

Experiment 3
Title: Parallel implementation of PI, Fundamentals of MPI Programming
Aim/Objective:
To demonstrate the parallel computation of PI using the Monte Carlo method with OpenMP
for improved performance and to showcase the MPI_Bcast function for broadcasting data from
one process to all other processes in a communicator.
Description:
In this experiment, students will explore the program generates random points in a unit square
and estimates PI by calculating the ratio of points within a quarter circle to the total points.
OpenMP is utilized to parallelize the computation for efficient use of multiple threads
The program initializes a message on the root process, and then MPI_Bcast is used to broadcast
this message to all other processes in the communicator. This demonstrates the concept of one-
to-all communication.
Pre-Requisites:
• Basic understanding of the Monte Carlo method for PI estimation.
• Familiarity with OpenMP directives and parallel programming concepts.
• Understanding communication patterns in parallel computing.
Pre-Lab:
1. How does the MPI program partition the sequence of numbers in the example provided
for computing π?

2. Discuss the challenges associated with load balancing when partitioning a sequence of
numbers for parallel computation?

Course Title PARALLEL & DISTRIBUTED COMPUTING ACADEMIC YEAR: 2024-25

Course Code(s) 22CS4106R Page 13 of 103
 Experiment # <TO BE FILLED BY STUDENT> Student ID <TO BE FILLED BY STUDENT>
Date <TO BE FILLED BY STUDENT> Student Name <TO BE FILLED BY STUDENT>

3. Explain the differences between the MPI_Send and MPI_Recv functions in MPI. How are
these functions employed for point-to-point communication in the context of parallelizing the
computation of π?

4. Discuss the trade-offs between parallel efficiency and communication overhead in the MPI
implementation of the π computation?

5. How does the program attempt to address load balancing, and what factors can impact the
efficiency of the load balancing strategy in the provided MPI example?

Course Title PARALLEL & DISTRIBUTED COMPUTING ACADEMIC YEAR: 2024-25

Course Code(s) 22CS4106R Page 14 of 103
 Experiment # <TO BE FILLED BY STUDENT> Student ID <TO BE FILLED BY STUDENT>
Date <TO BE FILLED BY STUDENT> Student Name <TO BE FILLED BY STUDENT>

• Program:

Course Title PARALLEL & DISTRIBUTED COMPUTING ACADEMIC YEAR: 2024-25

Course Code(s) 22CS4106R Page 15 of 103
 Experiment # <TO BE FILLED BY STUDENT> Student ID <TO BE FILLED BY STUDENT>
Date <TO BE FILLED BY STUDENT> Student Name <TO BE FILLED BY STUDENT>

In-Lab:

1. Implement Basic programs to demonstrate Broad cast and Reduction using mpi-reduce.c ,
mpi-bcast.c
o /* Program to demonstrate MPI_Reduce for sum reduction */
o /* Program to demonstrate MPI_BCAST for sum reduction */

• Program:

Course Title PARALLEL & DISTRIBUTED COMPUTING ACADEMIC YEAR: 2024-25

Course Code(s) 22CS4106R Page 16 of 103
Experiment # <TO BE FILLED BY STUDENT> Student ID <TO BE FILLED BY STUDENT>
Date <TO BE FILLED BY STUDENT> Student Name <TO BE FILLED BY STUDENT>

2. Write programs to Partition a sequence ofnumbers into parts and adding them using
a. Point-to-Point communications
b. MPI_Bcast
c. MPI_Scatter

Course Title PARALLEL & DISTRIBUTED COMPUTING ACADEMIC YEAR: 2024-25

Course Code(s) 22CS4106R Page 17 of 103
 Experiment # <TO BE FILLED BY STUDENT> Student ID <TO BE FILLED BY STUDENT>
Date <TO BE FILLED BY STUDENT> Student Name <TO BE FILLED BY STUDENT>

Sample VIVA-VOCE Questions (In-Lab):

1. How does the Monte Carlo method work for estimating the value of PI in the provided
OpenMP program?
2. Explain the significance of the #pragma omp parallel for directive in the program.
3. What is the purpose of the reduction(+:count) clause in the parallel for loop?
4. How does the number of OpenMP threads affect the performance and accuracy of the
PI estimation?
5. Can you explain a scenario where the Monte Carlo method for PI estimation and
parallelization with OpenMP might be particularly advantageous?

Course Title PARALLEL & DISTRIBUTED COMPUTING ACADEMIC YEAR: 2024-25

Course Code(s) 22CS4106R Page 18 of 103
 Experiment # <TO BE FILLED BY STUDENT> Student ID <TO BE FILLED BY STUDENT>
Date <TO BE FILLED BY STUDENT> Student Name <TO BE FILLED BY STUDENT>

Post-Lab:
1. Explain MPI_Reduce, MPI_Send, MPI_Sendrecv, MPI_Barrier, MPI_Scatter, MPI_Gather,
MPI_Allgather,MPI_Allreduce
Ans:

Course Title PARALLEL & DISTRIBUTED COMPUTING ACADEMIC YEAR: 2024-25

Course Code(s) 22CS4106R Page 19 of 103
Experiment # <TO BE FILLED BY STUDENT> Student ID <TO BE FILLED BY STUDENT>
Date <TO BE FILLED BY STUDENT> Student Name <TO BE FILLED BY STUDENT>

Evaluator Remark (if Any):

Marks Secured: out of 50

Signature of the Evaluator with Date

Course Title PARALLEL & DISTRIBUTED COMPUTING ACADEMIC YEAR: 2024-25

Course Code(s) 22CS4106R Page 20 of 103
 Experiment # <TO BE FILLED BY STUDENT> Student ID <TO BE FILLED BY STUDENT>
Date <TO BE FILLED BY STUDENT> Student Name <TO BE FILLED BY STUDENT>

Experiment 4
Title: Parallel Computation and Sorting Techniques Using MPI and OpenMP

Aim/Objective:

Implement parallel computation techniques using MPI (Message Passing Interface) and
OpenMP. It focuses on distributed data processing, such as calculating averages and finding
maxima using MPI, and parallel sorting algorithms like odd-even transposition sort using
OpenMP.

Description:
This lab introduces the concepts of distributed data computation and parallel sorting using MPI
and OpenMP. In the Pre-Lab, MPI programs are developed to compute averages and maxima
by distributing and collecting data using collective communication functions like MPI_Scatter,
MPI_Gather, and MPI_Reduce. The In-Lab task involves implementing an odd-even
transposition sort, a parallel sorting algorithm, using OpenMP.

Pre-Requisites:
• Understanding of distributed systems and parallel computing concepts.
• Familiarity with MPI and OpenMP libraries.
• Basic knowledge of C/C++ programming.
• Tools: MPI (e.g., MPICH or OpenMPI) and GCC with OpenMP support.

Pre-Lab:
1. Develop an MPI program to compute the average numbers using the following functions and
compute the time required for the same.
• MPI_Scatter and MPI_Gather.
• MPI_Scatter and MPI_Reduce.

• Program:

Course Title PARALLEL & DISTRIBUTED COMPUTING ACADEMIC YEAR: 2024-25

Course Code(s) 22CS4106R Page 21 of 103
 Experiment # <TO BE FILLED BY STUDENT> Student ID <TO BE FILLED BY STUDENT>
Date <TO BE FILLED BY STUDENT> Student Name <TO BE FILLED BY STUDENT>

2. Write an MPI program to find the maximum number of each group, then master finds
maximum of results.
1. Program:

Course Title PARALLEL & DISTRIBUTED COMPUTING ACADEMIC YEAR: 2024-25

Course Code(s) 22CS4106R Page 22 of 103
 Experiment # <TO BE FILLED BY STUDENT> Student ID <TO BE FILLED BY STUDENT>
Date <TO BE FILLED BY STUDENT> Student Name <TO BE FILLED BY STUDENT>

In-Lab:
1. Program to implement odd-even transposition sort using OpenMP

• Program:

Course Title PARALLEL & DISTRIBUTED COMPUTING ACADEMIC YEAR: 2024-25

Course Code(s) 22CS4106R Page 23 of 103
 Experiment # <TO BE FILLED BY STUDENT> Student ID <TO BE FILLED BY STUDENT>
Date <TO BE FILLED BY STUDENT> Student Name <TO BE FILLED BY STUDENT>

VIVA-VOCE Questions (In-Lab):

1. What is the difference between MPI Gather and MPI_Reduce?
2. How does MPI_Scatter distribute data among processes?
3. Why is odd-even transposition sort suitable for parallel computation?
4. Explain the difference between shared memory (OpenMP) and distributed memory
(MPI) models.
5. What is the purpose of the schedule clause in OpenMP?

Course Title PARALLEL & DISTRIBUTED COMPUTING ACADEMIC YEAR: 2024-25

Course Code(s) 22CS4106R Page 24 of 103
 Experiment # <TO BE FILLED BY STUDENT> Student ID <TO BE FILLED BY STUDENT>
Date <TO BE FILLED BY STUDENT> Student Name <TO BE FILLED BY STUDENT>
Post-Lab:
1. Write an MPI program to perform matrix-vector multiplication using MPI_Scatter and
MPI_Gather.
• Program:

Course Title PARALLEL & DISTRIBUTED COMPUTING ACADEMIC YEAR: 2024-25

Course Code(s) 22CS4106R Page 25 of 103
 Experiment # <TO BE FILLED BY STUDENT> Student ID <TO BE FILLED BY STUDENT>
Date <TO BE FILLED BY STUDENT> Student Name <TO BE FILLED BY STUDENT>

2. Implement a parallel bubble sort using OpenMP and compare its performance with odd-
even transposition sort.
• Program:

Evaluator Remark (if Any):

Marks Secured: out of 50

Signature of the Evaluator with Date

Course Title PARALLEL & DISTRIBUTED COMPUTING ACADEMIC YEAR: 2024-25

Course Code(s) 22CS4106R Page 28 of 103
 Experiment # <TO BE FILLED BY STUDENT> Student ID <TO BE FILLED BY STUDENT>
Date <TO BE FILLED BY STUDENT> Student Name <TO BE FILLED BY STUDENT>

Experiment 5
Title: Exploring Parallel Sorting Algorithms: Implementations and Analysis with MPI,
OpenMP

Aim/Objective:
To Investigate parallel sorting algorithms through implementations and analysis using MPI,
OpenMP, aiming to compare and optimize their performance on various architectures.

Description:
Develop sorting algorithms in parallel using MPI for distributed memory systems, OpenMP
for shared-memory multi-core architectures. Conduct a comprehensive analysis to understand
the strengths and weaknesses of each parallelization approach.

Pre-Requisites:

• Understanding of sorting algorithms such as merge sort, quicksort, Shell sort, and
bitonic sort.
• Familiarity with OpenMP directives for multi-core CPU parallelism.

Pre-Lab:
1. Write a Program to implement Parallel Odd Even sort - MPI

• Program:

Course Title PARALLEL & DISTRIBUTED COMPUTING ACADEMIC YEAR: 2024-25

Course Code(s) 22CS4106R Page 29 of 103
 Experiment # <TO BE FILLED BY STUDENT> Student ID <TO BE FILLED BY STUDENT>
Date <TO BE FILLED BY STUDENT> Student Name <TO BE FILLED BY STUDENT>

2. Write a Program to implement Merge sort using MPI

• Program:

Course Title PARALLEL & DISTRIBUTED COMPUTING ACADEMIC YEAR: 2024-25

Course Code(s) 22CS4106R Page 30 of 103
 Experiment # <TO BE FILLED BY STUDENT> Student ID <TO BE FILLED BY STUDENT>
Date <TO BE FILLED BY STUDENT> Student Name <TO BE FILLED BY STUDENT>

3. Write a Program to implement Bitonic sort using OpenMP

• Program:

Course Title PARALLEL & DISTRIBUTED COMPUTING ACADEMIC YEAR: 2024-25

Course Code(s) 22CS4106R Page 31 of 103
 Experiment # <TO BE FILLED BY STUDENT> Student ID <TO BE FILLED BY STUDENT>
Date <TO BE FILLED BY STUDENT> Student Name <TO BE FILLED BY STUDENT>

In-Lab:
1. A communicator has 3 processes (ID: 0, 1 and 2) such that process 0 sends a binary value
001 to process 1, in-turn process 1 increment the value by 1 (i.e. 010) and send the same
to process 2, in-turn process 2 increment the value by 1 (i.e. 011) and send the same to
process 0. Continue the process until the binary value becomes 111. Develop a MPI
program to implement the above scenario (Rotating Binary Addition).
• Program:

Course Title PARALLEL & DISTRIBUTED COMPUTING ACADEMIC YEAR: 2024-25

Course Code(s) 22CS4106R Page 32 of 103
Experiment # <TO BE FILLED BY STUDENT> Student ID <TO BE FILLED BY STUDENT>
Date <TO BE FILLED BY STUDENT> Student Name <TO BE FILLED BY STUDENT>

Course Title PARALLEL & DISTRIBUTED COMPUTING ACADEMIC YEAR: 2024-25

Course Code(s) 22CS4106R Page 33 of 103
 Experiment # <TO BE FILLED BY STUDENT> Student ID <TO BE FILLED BY STUDENT>
Date <TO BE FILLED BY STUDENT> Student Name <TO BE FILLED BY STUDENT>
Sample VIVA-VOCE Questions (In-Lab):

1. How does the parallel odd-even sort algorithm work in the MPI implementation?
2. How is the array distributed among MPI processes, and why is it important for efficiency?
3. Explain how point-to-point communication is utilized in the odd-even sort algorithm for
parallelism.
4. What is the significance of the odd and even phases in the parallel odd-even sort?
5. How does the parallel odd-even sort handle the synchronization of processes after each
phase?

Course Title PARALLEL & DISTRIBUTED COMPUTING ACADEMIC YEAR: 2024-25

Course Code(s) 22CS4106R Page 34 of 103
 Experiment # <TO BE FILLED BY STUDENT> Student ID <TO BE FILLED BY STUDENT>
Date <TO BE FILLED BY STUDENT> Student Name <TO BE FILLED BY STUDENT>

Post-Lab:
1. Write a Program to implement Merge sort, Sample Sort using OpenMP

• Program:

Course Title PARALLEL & DISTRIBUTED COMPUTING ACADEMIC YEAR: 2024-25

Course Code(s) 22CS4106R Page 35 of 103
 Experiment # <TO BE FILLED BY STUDENT> Student ID <TO BE FILLED BY STUDENT>
Date <TO BE FILLED BY STUDENT> Student Name <TO BE FILLED BY STUDENT>

2. Analyze odd-Even Merge sort and Hyper cube quick sort algorithm with an example

• Program:

Evaluator Remark (if Any):

Marks Secured: out of 50

Signature of the Evaluator with Date

Course Title PARALLEL & DISTRIBUTED COMPUTING ACADEMIC YEAR: 2024-25

Course Code(s) 22CS4106R Page 37 of 103
 Experiment # <TO BE FILLED BY STUDENT> Student ID <TO BE FILLED BY STUDENT>
Date <TO BE FILLED BY STUDENT> Student Name <TO BE FILLED BY STUDENT>

Experiment 6
Title: Advanced Parallel Computing in Mathematical Algorithms: Harnessing OpenMP, and
MPI for Matrix Operations, Integration, and Sorting

Aim/Objective:
To explore the benefits of parallelizing numerical algorithms such as Gaussian Elimination,
Trapezoidal Integration, and Row Sums, focusing on enhancing computational efficiency
through parallel decomposition, load balancing, inter-process communication, and scalability
considerations.

Description:
Parallel Gaussian Elimination divides the matrix manipulation tasks among multiple
processing units. Parallel decomposition techniques distribute the workload, improving overall
performance by exploiting parallelism in matrix operations.

Pre-Requisites:

• Understanding of numerical algorithms, particularly Gaussian Elimination, Trapezoidal

Integration, and Row Sums.
• Familiarity with parallel decomposition techniques, load balancing strategies, and inter-
process communication mechanisms.
• Knowledge of scalability factors, considering problem size and communication
efficiency.

Pre-Lab:
1. How does parallelizing Gaussian Elimination enhance computational efficiency, and
what role do parallel decomposition techniques play in this context?

Course Title PARALLEL & DISTRIBUTED COMPUTING ACADEMIC YEAR: 2024-25

Course Code(s) 22CS4106R Page 38 of 103
Experiment # <TO BE FILLED BY STUDENT> Student ID <TO BE FILLED BY STUDENT>
Date <TO BE FILLED BY STUDENT> Student Name <TO BE FILLED BY STUDENT>

2. Briefly explain the concept of load balancing in the context of parallel Trapezoidal
Integration and identify one challenge associated with achieving optimal load
distribution.

3. In a parallel environment, describe a key consideration for efficiently distributing the

computation of row sums among multiple processing units.

Course Title PARALLEL & DISTRIBUTED COMPUTING ACADEMIC YEAR: 2024-25

Course Code(s) 22CS4106R Page 39 of 103
Experiment # <TO BE FILLED BY STUDENT> Student ID <TO BE FILLED BY STUDENT>
Date <TO BE FILLED BY STUDENT> Student Name <TO BE FILLED BY STUDENT>
4. How does inter-process communication contribute to the success of a parallel Gaussian
Elimination implementation, and what is its impact on overall performance?

5. Defining scalability in the context of parallel computing and identify two factors that
can affect the scalability of algorithms like Gaussian Elimination, Trapezoidal
Integration,and Row Sums.

Course Title PARALLEL & DISTRIBUTED COMPUTING ACADEMIC YEAR: 2024-25

Course Code(s) 22CS4106R Page 40 of 103
 Experiment # <TO BE FILLED BY STUDENT> Student ID <TO BE FILLED BY STUDENT>
Date <TO BE FILLED BY STUDENT> Student Name <TO BE FILLED BY STUDENT>
In-Lab:
1. Write a program to perform Parallel Gaussian Elimination Using OpenMP
• Program:

Course Title PARALLEL & DISTRIBUTED COMPUTING ACADEMIC YEAR: 2024-25

Course Code(s) 22CS4106R Page 41 of 103
 Experiment # <TO BE FILLED BY STUDENT> Student ID <TO BE FILLED BY STUDENT>
Date <TO BE FILLED BY STUDENT> Student Name <TO BE FILLED BY STUDENT>

2.Write a program to perform Trapezoidal Integration Using OpenMP

• Program:

Course Title PARALLEL & DISTRIBUTED COMPUTING ACADEMIC YEAR: 2024-25

Course Code(s) 22CS4106R Page 42 of 103
 Experiment # <TO BE FILLED BY STUDENT> Student ID <TO BE FILLED BY STUDENT>
Date <TO BE FILLED BY STUDENT> Student Name <TO BE FILLED BY STUDENT>

3.Implementation of a distributed algorithm in MPI to find the Minimum Spanning Tree of an

undirected graph.
• Program:

Course Title PARALLEL & DISTRIBUTED COMPUTING ACADEMIC YEAR: 2024-25

Course Code(s) 22CS4106R Page 43 of 103
 Experiment # <TO BE FILLED BY STUDENT> Student ID <TO BE FILLED BY STUDENT>
Date <TO BE FILLED BY STUDENT> Student Name <TO BE FILLED BY STUDENT>

Sample VIVA-VOCE Questions (In-Lab):

1. What is OpenMP, and why is it used in parallel programming?

2. What is Trapezoidal Integration, and in what scenarios is it used?
3. How does OpenMP divide the workload for computing integration among threads?
4. Why is it essential to use synchronization mechanisms in parallel Gaussian Elimination?
5. Explain the role of pivoting in Gaussian Elimination. How is it handled in a parallel
implementation?

Course Title PARALLEL & DISTRIBUTED COMPUTING ACADEMIC YEAR: 2024-25

Course Code(s) 22CS4106R Page 44 of 103
 Experiment # <TO BE FILLED BY STUDENT> Student ID <TO BE FILLED BY STUDENT>
Date <TO BE FILLED BY STUDENT> Student Name <TO BE FILLED BY STUDENT>

Post-Lab:
1. Write a program to implement a parallel version of Trapezoidal rule Using MPI.

• Program:

Course Title PARALLEL & DISTRIBUTED COMPUTING ACADEMIC YEAR: 2024-25

Course Code(s) 22CS4106R Page 45 of 103
 Experiment # <TO BE FILLED BY STUDENT> Student ID <TO BE FILLED BY STUDENT>
Date <TO BE FILLED BY STUDENT> Student Name <TO BE FILLED BY STUDENT>

2. Given a directed graph, write a MPI program to count all possible walks (both closed and
open) of length k from every source and destination pair using DFS.
• Program:

Evaluator Remark (if Any):

Marks Secured: out of 50

Signature of the Evaluator with Date

Course Title PARALLEL & DISTRIBUTED COMPUTING ACADEMIC YEAR: 2024-25

Course Code(s) 22CS4106R Page 46 of 103
 Experiment # <TO BE FILLED BY STUDENT> Student ID <TO BE FILLED BY STUDENT>
Date <TO BE FILLED BY STUDENT> Student Name <TO BE FILLED BY STUDENT>

Experiment 7
Title: Distributed Algorithms and Parallel Programming with MPI

Aim/Objective:
Implement the fundamental concepts of the Message Passing Interface (MPI), a standard for
parallel programming in distributed systems. It focuses on performing distributed computation
tasks using MPI, such as synchronization, communication, and computation among multiple
processes.

Description:
The lab explores the use of MPI to enable parallel computation and communication in
distributed systems. Key MPI functions such as MPI_Barrier() and MPI_Wtime() are
introduced, along with practical applications like tree-structured global sum, vector addition,
and topology-based communication. Distributed algorithms for clock synchronization
(Lamport’s clock and Vector clock) are simulated to reinforce concepts of time and order in
distributed systems. Additionally, advanced concepts such as non-blocking communication and
the Monte Carlo method for numerical computation are covered.

Pre-Requisites:
• Basic understanding of distributed systems.
• Familiarity with C/C++ programming.
• Knowledge of parallel processing concepts.
• MPI library installation and environment setup.
Pre-Lab:
1. Demonstrate the use of MPI_Barrier() and MPI_Wtime() with a program
• Program:

Course Title PARALLEL & DISTRIBUTED COMPUTING ACADEMIC YEAR: 2024-25

Course Code(s) 22CS4106R Page 47 of 103
 Experiment # <TO BE FILLED BY STUDENT> Student ID <TO BE FILLED BY STUDENT>
Date <TO BE FILLED BY STUDENT> Student Name <TO BE FILLED BY STUDENT>

In-Lab:
1. Lamports clock Algorithm Implementation in C – simulation.

• Program:

Course Title PARALLEL & DISTRIBUTED COMPUTING ACADEMIC YEAR: 2024-25

Course Code(s) 22CS4106R Page 48 of 103
 Experiment # <TO BE FILLED BY STUDENT> Student ID <TO BE FILLED BY STUDENT>
Date <TO BE FILLED BY STUDENT> Student Name <TO BE FILLED BY STUDENT>

2. Vector clock Algorithm Implementation in C - simulation

• Program:

Course Title PARALLEL & DISTRIBUTED COMPUTING ACADEMIC YEAR: 2024-25

Course Code(s) 22CS4106R Page 49 of 103
 Experiment # <TO BE FILLED BY STUDENT> Student ID <TO BE FILLED BY STUDENT>
Date <TO BE FILLED BY STUDENT> Student Name <TO BE FILLED BY STUDENT>
VIVA-VOCE Questions (In-Lab):
a. What is the purpose of MPI_Barrier()?
b. How does MPI_Wtime() differ from other time measurement functions?
c. Explain the significance of tree-structured communication in distributed
systems.
d. How do Lamport’s clocks differ from Vector clocks?
e. What is the advantage of non-blocking communication in MPI?

Course Title PARALLEL & DISTRIBUTED COMPUTING ACADEMIC YEAR: 2024-25

Course Code(s) 22CS4106R Page 50 of 103
 Experiment # <TO BE FILLED BY STUDENT> Student ID <TO BE FILLED BY STUDENT>
Date <TO BE FILLED BY STUDENT> Student Name <TO BE FILLED BY STUDENT>

Post-Lab:
1.Write an MPI Program that implements Ring Topology using Non-Blocking Send/Receive
• Program:

Course Title PARALLEL & DISTRIBUTED COMPUTING ACADEMIC YEAR: 2024-25

Course Code(s) 22CS4106R Page 51 of 103
 Experiment # <TO BE FILLED BY STUDENT> Student ID <TO BE FILLED BY STUDENT>
Date <TO BE FILLED BY STUDENT> Student Name <TO BE FILLED BY STUDENT>

2. Calculate PI using Monte Carlo method and MPI

• Program:

Evaluator Remark (if Any):

Marks Secured: out of 50

Signature of the Evaluator with Date

Course Title PARALLEL & DISTRIBUTED COMPUTING ACADEMIC YEAR: 2024-25

Course Code(s) 22CS4106R Page 56 of 103
 Experiment # <TO BE FILLED BY STUDENT> Student ID <TO BE FILLED BY STUDENT>
Date <TO BE FILLED BY STUDENT> Student Name <TO BE FILLED BY STUDENT>

Experiment 8
Title: Distributed Mutual Exclusion and Parallel Programming

Aim/Objective:
To implement various distributed mutual exclusion algorithms and parallel programming
techniques for efficient resource sharing and computation.
Description:
This lab explores key distributed algorithms for mutual exclusion and parallel computation.
These techniques ensure fair resource allocation and optimized execution in distributed and
parallel computing environments.

Pre-Requisites:
Familiarity with mutual exclusion algorithms, such as those used to manage access to shared
resources in a distributed environment, is also required. Proficiency in C programming is
necessary, along with knowledge of parallel programming libraries like OpenMP for shared-
memory systems and MPI for distributed-memory systems. Additionally, the lab requires a
properly configured environment with OpenMP and MPI support, such as GCC with OpenMP
enabled and MPI libraries like MPICH or OpenMPI, to execute and test the programs effectively.
Pre-Lab:
1. What is mutual exclusion, and why is it required in distributed systems?

Course Title PARALLEL & DISTRIBUTED COMPUTING ACADEMIC YEAR: 2024-25

Course Code(s) 22CS4106R Page 57 of 103
 Experiment # <TO BE FILLED BY STUDENT> Student ID <TO BE FILLED BY STUDENT>
Date <TO BE FILLED BY STUDENT> Student Name <TO BE FILLED BY STUDENT>
2. How does the Birman-Schiper-Stephenson protocol ensure causal ordering of events?

3. Explain how parallel computing improves computational performance.

Course Title PARALLEL & DISTRIBUTED COMPUTING ACADEMIC YEAR: 2024-25

Course Code(s) 22CS4106R Page 58 of 103
 Experiment # <TO BE FILLED BY STUDENT> Student ID <TO BE FILLED BY STUDENT>
Date <TO BE FILLED BY STUDENT> Student Name <TO BE FILLED BY STUDENT>
4. Describe the working of OpenMP and MPI in parallel programming.

Course Title PARALLEL & DISTRIBUTED COMPUTING ACADEMIC YEAR: 2024-25

Course Code(s) 22CS4106R Page 59 of 103
 Experiment # <TO BE FILLED BY STUDENT> Student ID <TO BE FILLED BY STUDENT>
Date <TO BE FILLED BY STUDENT> Student Name <TO BE FILLED BY STUDENT>
In-Lab:
1. C program to implement Birman-Schiper-Stephenson protocol - simulation
• Program:

Course Title PARALLEL & DISTRIBUTED COMPUTING ACADEMIC YEAR: 2024-25

Course Code(s) 22CS4106R Page 60 of 103
 Experiment # <TO BE FILLED BY STUDENT> Student ID <TO BE FILLED BY STUDENT>
Date <TO BE FILLED BY STUDENT> Student Name <TO BE FILLED BY STUDENT>

2. Ricart Agrawala distributed mutual exclusion algorithm

• Program:

Course Title PARALLEL & DISTRIBUTED COMPUTING ACADEMIC YEAR: 2024-25

Course Code(s) 22CS4106R Page 61 of 103
 Experiment # <TO BE FILLED BY STUDENT> Student ID <TO BE FILLED BY STUDENT>
Date <TO BE FILLED BY STUDENT> Student Name <TO BE FILLED BY STUDENT>

3. Implement token ring based mutual exclusion algorithm.

• Program:

Course Title PARALLEL & DISTRIBUTED COMPUTING ACADEMIC YEAR: 2024-25

Course Code(s) 22CS4106R Page 62 of 103
 Experiment # <TO BE FILLED BY STUDENT> Student ID <TO BE FILLED BY STUDENT>
Date <TO BE FILLED BY STUDENT> Student Name <TO BE FILLED BY STUDENT>

VIVA-VOCE Questions (In-Lab):

1. What is the time complexity of a conventional serial algorithm for multiplying two n x n
matrices?
2. Can you explain the step-by-step process of multiplying two matrices using the
conventional serial approach?
3. Compare and contrast row-wise 1D partitioning and 2D partitioning for matrix-vector
multiplication in terms of parallel efficiency and communication overhead.
4. How does the choice of partitioning strategy impact the parallel runtime of matrix-vector
multiplication?
5. Describe the Cannons algorithm for matrix-matrix multiplication on 16 processes. What
advantages does this algorithm offer in terms of parallelization?

Course Title PARALLEL & DISTRIBUTED COMPUTING ACADEMIC YEAR: 2024-25

Course Code(s) 22CS4106R Page 63 of 103
 Experiment # <TO BE FILLED BY STUDENT> Student ID <TO BE FILLED BY STUDENT>
Date <TO BE FILLED BY STUDENT> Student Name <TO BE FILLED BY STUDENT>

Post-Lab:
1. Implement a Program to implement Matrix-Matrix multiplication using OpenMP
• Program:

Course Title PARALLEL & DISTRIBUTED COMPUTING ACADEMIC YEAR: 2024-25

Course Code(s) 22CS4106R Page 64 of 103
 Experiment # <TO BE FILLED BY STUDENT> Student ID <TO BE FILLED BY STUDENT>
Date <TO BE FILLED BY STUDENT> Student Name <TO BE FILLED BY STUDENT>

2. Program to implement Matrix-Matrix multiplication using MPI

• Program:

Evaluator Remark (if Any):

Marks Secured: out of 50

Signature of the Evaluator with Date

Course Title PARALLEL & DISTRIBUTED COMPUTING ACADEMIC YEAR: 2024-25

Course Code(s) 22CS4106R Page 66 of 103
 Experiment # <TO BE FILLED BY STUDENT> Student ID <TO BE FILLED BY STUDENT>
Date <TO BE FILLED BY STUDENT> Student Name <TO BE FILLED BY STUDENT>

Experiment 9
Title: Leader Election in Distributed Systems: Implementing Bully and Ring Algorithms
Using MPI/C Library

Aim/Objective:
To explore distributed computation essentials through the implementation of Bully and
Ring algorithms using MPI and C library, focusing on demonstrating inter-process
communication and fault-tolerant leader selection.

Description:
Develop and analyze parallel applications using MPI for distributed computing. Emphasize
matrix operations, dot products, and numerical calculations, showcasing the versatility of
parallel computation in solving computationally intensive tasks.

Pre-Requisites:
Proficiency in C/C++ programming, understanding of parallel programming concepts,
familiarity with MPI programming models, and knowledge of matrix operations,dot
products, and numerical algorithms.

Pre-Lab:

1. How can parallel dot product be implemented using multi-threading or parallel processing?

2. How does parallel processing or GPU acceleration impact the performance of matrix
transposition operations?

Course Title PARALLEL & DISTRIBUTED COMPUTING ACADEMIC YEAR: 2024-25

Course Code(s) 22CS4106R Page 67 of 103
 Experiment # <TO BE FILLED BY STUDENT> Student ID <TO BE FILLED BY STUDENT>
Date <TO BE FILLED BY STUDENT> Student Name <TO BE FILLED BY STUDENT>

3.Write the algorithm of prefix sum and define the use of prefix sum reduce time complexity
in dynamic programming algorithms?

4. Write pseudocode for an in-place matrix transposition function, assuming the original matrix
is stored in row-major order in memory. Analyse its time and space complexity.

5. What are the advantages of using parallel computation for the dot product in terms of
performance?

Course Title PARALLEL & DISTRIBUTED COMPUTING ACADEMIC YEAR: 2024-25

Course Code(s) 22CS4106R Page 68 of 103
 Experiment # <TO BE FILLED BY STUDENT> Student ID <TO BE FILLED BY STUDENT>
Date <TO BE FILLED BY STUDENT> Student Name <TO BE FILLED BY STUDENT>

In-Lab:

1. Write a program to implement Bully algorithm using MPI

• Program:

Course Title PARALLEL & DISTRIBUTED COMPUTING ACADEMIC YEAR: 2024-25

Course Code(s) 22CS4106R Page 69 of 103
 Experiment # <TO BE FILLED BY STUDENT> Student ID <TO BE FILLED BY STUDENT>
Date <TO BE FILLED BY STUDENT> Student Name <TO BE FILLED BY STUDENT>

2. Write a program to implement ring algorithm using MPI

• Program:

Course Title PARALLEL & DISTRIBUTED COMPUTING ACADEMIC YEAR: 2024-25

Course Code(s) 22CS4106R Page 70 of 103
 Experiment # <TO BE FILLED BY STUDENT> Student ID <TO BE FILLED BY STUDENT>
Date <TO BE FILLED BY STUDENT> Student Name <TO BE FILLED BY STUDENT>

VIVA-VOCE Questions (In-Lab):

1. How does the concept of parallel prefix sum differ from the sequential approach, and why
is it beneficial in parallel computing?
2.Explain the role of MPI collective communications in the parallel prefix sum program. Which
specific MPI function is commonly used for such operations, and why?
3. Describe the process of matrix transposition and its significance in parallel computing. How
does MPI facilitate the parallelization of this operation across multiple processes?
4.What considerations should be taken into account when implementing matrix transposition
using MPI, especially in terms of data distribution and communication patterns?
5. What is the collective communication model in MPI, and how does it contribute to the
implementation of the dot product program?

Course Title PARALLEL & DISTRIBUTED COMPUTING ACADEMIC YEAR: 2024-25

Course Code(s) 22CS4106R Page 71 of 103
 Experiment # <TO BE FILLED BY STUDENT> Student ID <TO BE FILLED BY STUDENT>
Date <TO BE FILLED BY STUDENT> Student Name <TO BE FILLED BY STUDENT>

Post-Lab:
1. Implement a Fault-Tolerant Ring Algorithm with Failure Detection
• Program:

Course Title PARALLEL & DISTRIBUTED COMPUTING ACADEMIC YEAR: 2024-25

Course Code(s) 22CS4106R Page 72 of 103
 Experiment # <TO BE FILLED BY STUDENT> Student ID <TO BE FILLED BY STUDENT>
Date <TO BE FILLED BY STUDENT> Student Name <TO BE FILLED BY STUDENT>

2. Modify and Analyze the Bully Algorithm for Dynamic Node Changes.
• Program:

Evaluator Remark (if Any):

Marks Secured: out of 50

Signature of the Evaluator with Date

Course Title PARALLEL & DISTRIBUTED COMPUTING ACADEMIC YEAR: 2024-25

Course Code(s) 22CS4106R Page 74 of 103
 Experiment # <TO BE FILLED BY STUDENT> Student ID <TO BE FILLED BY STUDENT>
Date <TO BE FILLED BY STUDENT> Student Name <TO BE FILLED BY STUDENT>

Experiment 10
Title: Exploring Parallel Computing: Data Structures, Algorithmic Patterns, and Visualization
Techniques

Aim/Objective:
Explore parallel computing by investigating data structures, algorithmic patterns, and
visualization techniques, aiming to enhance understanding and proficiency in leveraging
parallelism for computational tasks.

Description:
Develop and analyze parallel algorithms, employ advanced data structures, and utilize
visualization techniques to understand and optimize the performance of parallel computations.

Pre-Requisites:
Proficiency in a programming language (e.g., C/C++), foundational knowledge of data
structures and algorithms, and a basic understanding of parallel computing concepts.

Pre-Lab:
1. What is a parallel prefix sum in the context of linked lists, and how does it differ from the
array-based approach?

2. Explain the significance of list ranking in parallel computing and its relevance to linked
lists.

Course Title PARALLEL & DISTRIBUTED COMPUTING ACADEMIC YEAR: 2024-25

Course Code(s) 22CS4106R Page 75 of 103
 Experiment # <TO BE FILLED BY STUDENT> Student ID <TO BE FILLED BY STUDENT>
Date <TO BE FILLED BY STUDENT> Student Name <TO BE FILLED BY STUDENT>

3. Briefly describe the Mandelbrot Set and how parallel computing accelerates its rendering
process.

4. Compare shared memory and distributed memory architectures in parallel computing,

highlighting their key differences.

5. How would you assess the performance of a parallel algorithm for linked lists? Can you
name one optimization technique for enhancing its efficiency?

Course Title PARALLEL & DISTRIBUTED COMPUTING ACADEMIC YEAR: 2024-25

Course Code(s) 22CS4106R Page 76 of 103
 Experiment # <TO BE FILLED BY STUDENT> Student ID <TO BE FILLED BY STUDENT>
Date <TO BE FILLED BY STUDENT> Student Name <TO BE FILLED BY STUDENT>

In-Lab:

1. Write a program to demonstrate Linked list Parallel prefix sum using MPI
• Program:

Course Title PARALLEL & DISTRIBUTED COMPUTING ACADEMIC YEAR: 2024-25

Course Code(s) 22CS4106R Page 77 of 103
 Experiment # <TO BE FILLED BY STUDENT> Student ID <TO BE FILLED BY STUDENT>
Date <TO BE FILLED BY STUDENT> Student Name <TO BE FILLED BY STUDENT>

2. Write a program to implement concurrent linked list

• Program:

Course Title PARALLEL & DISTRIBUTED COMPUTING ACADEMIC YEAR: 2024-25

Course Code(s) 22CS4106R Page 78 of 103
 Experiment # <TO BE FILLED BY STUDENT> Student ID <TO BE FILLED BY STUDENT>
Date <TO BE FILLED BY STUDENT> Student Name <TO BE FILLED BY STUDENT>

3.Parallel Program to render the Mandelbrot Set

• Program:

Course Title PARALLEL & DISTRIBUTED COMPUTING ACADEMIC YEAR: 2024-25

Course Code(s) 22CS4106R Page 79 of 103
 Experiment # <TO BE FILLED BY STUDENT> Student ID <TO BE FILLED BY STUDENT>
Date <TO BE FILLED BY STUDENT> Student Name <TO BE FILLED BY STUDENT>

VIVA-VOCE Questions (In-Lab):

1. Explain the concept of parallel prefix sum applied to a linked list. How does MPI facilitate
communication and coordination in the implementation of this algorithm?
2. Discuss the advantages and challenges of parallelizing prefix sum on a linked list data
structure. How does the linked list structure impact the parallelization approach compared
to arrays?

3. What is the significance of implementing a concurrent linked list? Describe the key
challenges and solutions involved in ensuring concurrent access and modification of a linked
list.
4. Explain the potential concurrency issues that may arise in a linked list and how the concurrent
linked list program addresses them. What synchronization mechanisms are employed?

5. Walk through the steps involved in rendering the Mandelbrot Set in parallel. How does
parallelization enhance the efficiency of generating the Mandelbrot Set?

Course Title PARALLEL & DISTRIBUTED COMPUTING ACADEMIC YEAR: 2024-25

Course Code(s) 22CS4106R Page 80 of 103
 Experiment # <TO BE FILLED BY STUDENT> Student ID <TO BE FILLED BY STUDENT>
Date <TO BE FILLED BY STUDENT> Student Name <TO BE FILLED BY STUDENT>

Post-Lab:
1. Parallel Program to implement List Ranking
• Program:

Course Title PARALLEL & DISTRIBUTED COMPUTING ACADEMIC YEAR: 2024-25

Course Code(s) 22CS4106R Page 81 of 103
 Experiment # <TO BE FILLED BY STUDENT> Student ID <TO BE FILLED BY STUDENT>
Date <TO BE FILLED BY STUDENT> Student Name <TO BE FILLED BY STUDENT>

2. Parallel Program to render the Mandelbrot Set.

• Program:

Course Title PARALLEL & DISTRIBUTED COMPUTING ACADEMIC YEAR: 2024-25

Course Code(s) 22CS4106R Page 82 of 103
 Experiment # <TO BE FILLED BY STUDENT> Student ID <TO BE FILLED BY STUDENT>
Date <TO BE FILLED BY STUDENT> Student Name <TO BE FILLED BY STUDENT>

3. OpenMP and MPI implementation of Travelling Salesman Problem using Branch and Bound
algorithm.
• Program:
Experiment # <TO BE FILLED BY STUDENT> Student ID <TO BE FILLED BY STUDENT>
Date <TO BE FILLED BY STUDENT> Student Name <TO BE FILLED BY STUDENT>

Evaluator Remark (if Any):

Marks Secured: out of 50

Signature of the Evaluator with Date

Course Title PARALLEL & DISTRIBUTED COMPUTING ACADEMIC YEAR: 2024-25

Course Code(s) 22CS4106R Page 83 of 103
 Experiment # <TO BE FILLED BY STUDENT> Student ID <TO BE FILLED BY STUDENT>
Date <TO BE FILLED BY STUDENT> Student Name <TO BE FILLED BY STUDENT>

Experiment 11
Title: Simulation and Distributed Programming: Consensus and Matching Algorithms

Aim/Objective:
To simulate the Lamport-Shostak-Pease Algorithm to achieve consensus in distributed systems
despite the presence of faulty nodes and to implement the Stable Marriage Problem using MPI,
utilizing asynchronous message passing for inter-process communication and logging events
such as pairings and rejections to a trace file.
Description:
The implementation of the Stable Marriage Problem using MPI demonstrates distributed
computing techniques, including asynchronous message passing for efficient inter-process
communication and event logging for monitoring and debugging. achieving consensus and
solving combinatorial problems using distributed programming. The Lamport-Shostak-Pease
Algorithm simulation addresses the challenge of reaching agreement in the presence of faulty
nodes, showcasing fault tolerance in distributed systems
Pre-Requisites:

Understanding of distributed systems concepts, particularly fault tolerance and consensus

mechanisms. Familiarity with the Lamport-Shostak-Pease Algorithm is essential, as it forms
the basis for achieving consensus in the presence of faulty nodes. Additionally, a foundational
knowledge of combinatorial optimization problems, such as the Stable Marriage Problem, is
necessary. Proficiency in C programming and experience with the Message Passing Interface
(MPI) are required to implement distributed solutions effectively.

Pre-Lab:
1. What is the significance of the Lamport-Shostak-Pease Algorithm in distributed
systems?

Course Title PARALLEL & DISTRIBUTED COMPUTING ACADEMIC YEAR: 2024-25

Course Code(s) 22CS4106R Page 84 of 103
Experiment # <TO BE FILLED BY STUDENT> Student ID <TO BE FILLED BY STUDENT>
Date <TO BE FILLED BY STUDENT> Student Name <TO BE FILLED BY STUDENT>
2. Explain the Byzantine Generals Problem and its relevance to achieving consensus
computation?

Course Title PARALLEL & DISTRIBUTED COMPUTING ACADEMIC YEAR: 2024-25

Course Code(s) 22CS4106R Page 85 of 103
Experiment # <TO BE FILLED BY STUDENT> Student ID <TO BE FILLED BY STUDENT>
Date <TO BE FILLED BY STUDENT> Student Name <TO BE FILLED BY STUDENT>

3. What is the Stable Marriage Problem, and how can it be solved using distributed
programming?

4. How does the asynchronous message passing differ from synchronous communication
in MPI?

Course Title PARALLEL & DISTRIBUTED COMPUTING ACADEMIC YEAR: 2024-25

Course Code(s) 22CS4106R Page 86 of 103
 Experiment # <TO BE FILLED BY STUDENT> Student ID <TO BE FILLED BY STUDENT>
Date <TO BE FILLED BY STUDENT> Student Name <TO BE FILLED BY STUDENT>

In-Lab:
1. Implement Lamport-Shostak-Pease Algorithm in C- simulation
• Program:

Course Title PARALLEL & DISTRIBUTED COMPUTING ACADEMIC YEAR: 2024-25

Course Code(s) 22CS4106R Page 87 of 103
 Experiment # <TO BE FILLED BY STUDENT> Student ID <TO BE FILLED BY STUDENT>
Date <TO BE FILLED BY STUDENT> Student Name <TO BE FILLED BY STUDENT>

2. MPI Distributed program to solve the Stable Marriage Problem. The processes should
communicate using asynchronous message passing. Log the trace of events (pairing,
breaking), as they happen, in a text file named as Log.txt.
• Program:

Course Title PARALLEL & DISTRIBUTED COMPUTING ACADEMIC YEAR: 2024-25

Course Code(s) 22CS4106R Page 88 of 103
 Experiment # <TO BE FILLED BY STUDENT> Student ID <TO BE FILLED BY STUDENT>
Date <TO BE FILLED BY STUDENT> Student Name <TO BE FILLED BY STUDENT>

VIVA-VOCE Questions (In-Lab):

1. How does the Lamport-Shostak-Pease Algorithm handle faulty nodes in a distributed

system?
2. What are the challenges of implementing the Stable Marriage Problem in a distributed
environment?
3. Explain the advantages of using asynchronous communication in MPI programs.
4. How can logging improve debugging and analysis in distributed systems?
5. What modifications would you make to the Lamport-Shostak-Pease Algorithm for non-
Byzantine faults?

Course Title PARALLEL & DISTRIBUTED COMPUTING ACADEMIC YEAR: 2024-25

Course Code(s) 22CS4106R Page 89 of 103
 Experiment # <TO BE FILLED BY STUDENT> Student ID <TO BE FILLED BY STUDENT>
Date <TO BE FILLED BY STUDENT> Student Name <TO BE FILLED BY STUDENT>

Post-Lab:

1. Modify the table Marriage MPI Program and include the dynamic addition.
• Program:

Course Title PARALLEL & DISTRIBUTED COMPUTING ACADEMIC YEAR: 2024-25

Course Code(s) 22CS4106R Page 90 of 103
 Experiment # <TO BE FILLED BY STUDENT> Student ID <TO BE FILLED BY STUDENT>
Date <TO BE FILLED BY STUDENT> Student Name <TO BE FILLED BY STUDENT>

• In the above task Log these changes to Log.txt

• Program:

Evaluator Remark (if Any):

Marks Secured: out of 50

Signature of the Evaluator with Date

Course Title PARALLEL & DISTRIBUTED COMPUTING ACADEMIC YEAR: 2024-25

Course Code(s) 22CS4106R Page 90 of 103
 Experiment # <TO BE FILLED BY STUDENT> Student ID <TO BE FILLED BY STUDENT>
Date <TO BE FILLED BY STUDENT> Student Name <TO BE FILLED BY STUDENT>

Experiment 12
Title: Distributed Systems and Cloud Computing: Web Services, MPI, and HPC Workflows

Aim/Objective:
To explore and implement distributed systems using web services, Message Passing Interface
(MPI), and High-Performance Computing (HPC) workflows by creating, consuming, and
running distributed applications and jobs on multi-node and single-node systems in HPC and
cloud environments.

Description:
It emphasizes hands-on implementation, effective resource management, and performance
analysis, providing a comprehensive understanding of real-world computing environments.
Developing and deploying RESTful web services, running distributed MPI applications on
high-performance computing (HPC) systems with both multi-node and single-node
configurations, and configuring and testing cloud-based MPI clusters on AWS

Pre-Requisites:
A basic understanding of distributed systems and web services, including REST and SOAP, is
essential for this lab. Familiarity with MPI programming and parallel computing concepts is
also required to successfully implement the tasks. Additionally, knowledge of cloud platforms
such as AWS and job schedulers like Slurm is necessary. Hands-on experience with
Linux/Unix commands and shell scripting will be beneficial, along with pre-configured access
to KLU HPC and AWS accounts to facilitate the practical exercises.

Pre-Lab:
1. What are web services, and how do they facilitate communication in distributed
systems?

Course Title PARALLEL & DISTRIBUTED COMPUTING ACADEMIC YEAR: 2024-25

Course Code(s) 22CS4106R Page 92 of 103
Experiment # <TO BE FILLED BY STUDENT> Student ID <TO BE FILLED BY STUDENT>
Date <TO BE FILLED BY STUDENT> Student Name <TO BE FILLED BY STUDENT>
2. What is the difference between multi-node and single-node jobs?

3. Describe the importance of workload managers like Slurm in HPC

Course Title PARALLEL & DISTRIBUTED COMPUTING ACADEMIC YEAR: 2024-25

Course Code(s) 22CS4106R Page 93 of 103
Experiment # <TO BE FILLED BY STUDENT> Student ID <TO BE FILLED BY STUDENT>
Date <TO BE FILLED BY STUDENT> Student Name <TO BE FILLED BY STUDENT>
4. How does AWS Parallel Cluster simplify running HPC workloads on the cloud?

Course Title PARALLEL & DISTRIBUTED COMPUTING ACADEMIC YEAR: 2024-25

Course Code(s) 22CS4106R Page 94 of 103
 Experiment # <TO BE FILLED BY STUDENT> Student ID <TO BE FILLED BY STUDENT>
Date <TO BE FILLED BY STUDENT> Student Name <TO BE FILLED BY STUDENT>

In-Lab:
1. Create a simple web service and write distributed application(calculator) to consume the
Web Service

Course Title PARALLEL & DISTRIBUTED COMPUTING ACADEMIC YEAR: 2024-25

Course Code(s) 22CS4106R Page 95 of 103
 Experiment # <TO BE FILLED BY STUDENT> Student ID <TO BE FILLED BY STUDENT>
Date <TO BE FILLED BY STUDENT> Student Name <TO BE FILLED BY STUDENT>

2. Run a multi-node MPI job with Slurm in KLU HPC

• Program:

Course Title PARALLEL & DISTRIBUTED COMPUTING ACADEMIC YEAR: 2024-25

Course Code(s) 22CS4106R Page 96 of 103
 Experiment # <TO BE FILLED BY STUDENT> Student ID <TO BE FILLED BY STUDENT>
Date <TO BE FILLED BY STUDENT> Student Name <TO BE FILLED BY STUDENT>

3. Run a single node job in AWS PCS

• Program:

Course Title PARALLEL & DISTRIBUTED COMPUTING ACADEMIC YEAR: 2024-25

Course Code(s) 22CS4106R Page 97 of 103
 Experiment # <TO BE FILLED BY STUDENT> Student ID <TO BE FILLED BY STUDENT>
Date <TO BE FILLED BY STUDENT> Student Name <TO BE FILLED BY STUDENT>

VIVA-VOCE Questions (In-Lab):

1. What are the key differences between REST and SOAP web services?
2. Explain the process of submitting and monitoring jobs in Slurm.
3. What is the significance of using AWS PCS for single-node jobs?
4. What factors affect the performance of multi-node jobs in HPC environments?

Course Title PARALLEL & DISTRIBUTED COMPUTING ACADEMIC YEAR: 2024-25

Course Code(s) 22CS4106R Page 98 of 103
 Experiment # <TO BE FILLED BY STUDENT> Student ID <TO BE FILLED BY STUDENT>
Date <TO BE FILLED BY STUDENT> Student Name <TO BE FILLED BY STUDENT>

Post-Lab:
1. Run a multi-node MPI job with Slurm in AWS PCS
• Program:

Course Title PARALLEL & DISTRIBUTED COMPUTING ACADEMIC YEAR: 2024-25

Course Code(s) 22CS4106R Page 99 of 103
 Experiment # <TO BE FILLED BY STUDENT> Student ID <TO BE FILLED BY STUDENT>
Date <TO BE FILLED BY STUDENT> Student Name <TO BE FILLED BY STUDENT>

2. Launching an Amazon EC2 MPI Cluster

• Program:

Evaluator Remark (if Any):

Marks Secured: out of 50

Signature of the Evaluator with Date

Course Title PARALLEL & DISTRIBUTED COMPUTING ACADEMIC YEAR: 2024-25

Course Code(s) 22CS4106R Page 99 of 103