Programming/C-Plusplus
From HPC
Contents
Programming Details
C++ is a general-purpose, imperative computer programming language. It is derived from the C language and shares many constructs with it, it also expands into object-orientated programming paradigms.
|
There is direct support for openMP and openMPI. |
Programming example
#include <omp.h>
#include <iostream>
int main ()
{
int thread_number;
#pragma omp parallel private(thread_number)
{
#pragma omp for schedule(static) nowait
for (int i = 0; i < 100; i++)
{
thread_number = omp_get_thread_num();
cout << "Thread " << thread_number << " says " << i << endl;
}
}
return 0;
}
#include <stdlib.h>
#include <stdio.h>
#include <omp.h>
#include <iostream>
#include <chrono>
#include <vector>
using namespace std;
#pragma optimize( "[t]", on)
int main(int argc, char *argv[]){
const int size = atoi(argv[1]);
const int maxVal = 255;
int row, col, ele;
double** m1 = new double*[size];
double** m2 = new double*[size];
double** res = new double*[size];
for (int i = 0; i < size; i++){
m1[i] = new double[size];
m2[i] = new double[size];
res[i] = new double[size];
}
int procs = omp_get_num_procs();
omp_set_num_threads(procs);
std::chrono::duration<double, std::micro> time;
for (int row = 0; row < size; row++){
for (int col = 0; col < size; col++){
m1[row][col] = 0 + (rand() % maxVal);
m2[row][col] = 0 + (rand() % maxVal);
}
}
auto start = std::chrono::high_resolution_clock::now();
#pragma omp parallel shared(m1, m2, res)
{
#pragma omp for nowait private(row, col, ele) schedule(dynamic)
for (int row = 0; row < size; row++) {
for (int ele = 0; ele < size; ele++){
double temp = 0;
for (int col = 0; col < size; col++){
temp += m1[row][ele] * m2[ele][col];
}
res[row][col] += temp;
}
}
}
auto end = std::chrono::high_resolution_clock::now();
time = end - start;
cout << (time.count() / 1000 / 1000) << endl;
return 0;
}
Compilation
The program would be compiled in the following way, optional Intel compiler available too:
[username@login01 ~]$ module add gcc/8.2.0 [username@login01 ~]$ g++ -o test2 -fopenmp test2.c (openMP support) [username@login01 ~]$ ./test2 Thread 22 says 32 Thread 11 says 39 .....
Modules Available
The following modules are available for C
- module add gcc/8.2.0 (GNU compiler)
- module add intel/2018 (Intel compiler)
Note: OpenMP is a library directive within the compiler and does not require any additional module to be loaded.
Usage Examples
Batch example
#!/bin/bash #SBATCH -J openmpi-single-node #SBATCH -N 1 #SBATCH --ntasks-per-node 28 #SBATCH -o %N.%j.%a.out #SBATCH -e %N.%j.%a.err #SBATCH -p compute #SBATCH --exclusive #SBATCH --mail-user= your email address here echo $SLURM_JOB_NODELIST module purge module add gcc/8.2.0 export I_MPI_DEBUG=5 export I_MPI_FABRICS=shm:tmi export I_MPI_FALLBACK=no /home/user/CODE_SAMPLES/OPENMP/demo
[username@login01 ~]$ sbatch demo.job Submitted batch job 234552