Scribd
Upload
9 views

IT301 Lab Exercise 5

This document provides instructions for two programming assignments on parallel computing. 1) Students are asked to time the sequential and parallel execution of a pi calculation program to compare performance. 2) They must write a program to add elements of two arrays in parallel using OpenMP and analyze the execution times for various array sizes and schedule types, plotting the results in a graph. The analysis focuses on when parallel execution time surpasses sequential time with increasing problem size.

Uploaded by

Prerana
Copyright
© © All Rights Reserved
Available Formats
Download as PDF, TXT or read online on Scribd
9 views

IT301 Lab Exercise 5

This document provides instructions for two programming assignments on parallel computing. 1) Students are asked to time the sequential and parallel execution of a pi calculation program to compare performance. 2) They must write a program to add elements of two arrays in parallel using OpenMP and analyze the execution times for various array sizes and schedule types, plotting the results in a graph. The analysis focuses on when parallel execution time surpasses sequential time with increasing problem size.

Uploaded by

Prerana
Copyright
© © All Rights Reserved
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 3

NATIONAL INSTITUTE OF TECHNOLOGY KARNATAKA, SURATHKAL

DEPARTMENT OF INFORMATION TECHNOLOGY


IT 301 Parallel Computing
LAB 5 - 30th September 2022

Name :

Roll No. :

Execute following programs and put screen shots of the output. Write analysis of the result before
uploading in IRIS as a single pdf file. For programming exercises, write the code and also put a
screenshot of the results.

Total Marks =2+8=10 marks

1. How to compare sequential and parallel program execution times. ?

Include following header files in the program.

#include <sys/time.h>

#include <stdlib.h>

//Declare following variables.

struct timeval TimeValue_Start;

struct timezone TimeZone_Start;

struct timeval TimeValue_Final;

struct timezone TimeZone_Final;

long time_start, time_end;

double time_overhead;

Just before starting parallel region code , note down the time(start time)

gettimeofday(&TimeValue_Start, &TimeZone_Start);

After finishing the parallel region, get the end time.

gettimeofday(&TimeValue_Final, &TimeZone_Final);
Calculate the overhead time as follows:

time_start = TimeValue_Start.tv_sec * 1000000 + TimeValue_Start.tv_usec;

time_end = TimeValue_Final.tv_sec * 1000000 + TimeValue_Final.tv_usec;

time_overhead = (time_end - time_start)/1000000.0;

printf("\n\n\t\t Time in Seconds (T) : %lf",time_overhead);

Example: Execute the program and compare sequential and parallel executions. [sequential
result : 1Mark + Parallel Execution 1 Mark=2 Marks]

#include <stdio.h>
#include <sys/time.h>
#include <omp.h>
#include <stdlib.h>
int main(void)
{
struct timeval TimeValue_Start;
struct timezone TimeZone_Start;
struct timeval TimeValue_Final;
struct timezone TimeZone_Final;
long time_start, time_end;
double time_overhead;
double pi,x;
int i,N;
pi=0.0;
N=1000;
gettimeofday(&TimeValue_Start, &TimeZone_Start);
#pragma omp parallel for private(x) reduction(+:pi)
for(i=0;i<=N;i++)
{
x=(double)i/N;
pi+=4/(1+x*x);
}
gettimeofday(&TimeValue_Final, &TimeZone_Final);
time_start = TimeValue_Start.tv_sec * 1000000 + TimeValue_Start.tv_usec;
time_end = TimeValue_Final.tv_sec * 1000000 + TimeValue_Final.tv_usec;
time_overhead = (time_end - time_start)/1000000.0;
printf("\n\n\tTime in Seconds (T) : %lf\n",time_overhead);
pi=pi/N;
printf("\n \tPi is %f\n\n",pi);
}

2. Write a sequential program to add elements of two arrays (c[i]=a[i]*b[i]. Convert the same program for

parallel execution. Initialize arrays with random numbers. Consider an array size as 10k, 50k and 100k.

Analyse the result for maximum number of threads and various schedule() functions. Based on observation,

perform analysis of the total execution time and explain the result by plotting the graph. [increase array size

until parallel execution time is less than sequential execution.]

Schedule() Total Execution Total Execution Total Execution Total Execution time
time for number time for number time for number for number of
of iterations 5K of iterations 10K of iterations 50K iterations 100K

Sequential execution

static

Static, chunksize

Dynamic, chunksize

Guided

runtime

Plot the graph using any software and write your observation.

Results in each Row in the above table=1 mark=> 6 * 1= 6 Marks

Plotting graph=1 Mark

Observation =1 Marks

Total = 8 Marks

You might also like