README.md

Exercise 5: Parallel reduction

Objective

The goal of this exercise is to use a Kokkos parallel reduction.

Step 1: create a 2D matrix

Open the file main.cpp and create a 2D matrix of size Nx x Ny of type double.

Initialize the matrix with the values of your choice using Kokkos::parallel_for and Kokkos::MDRangePolicy with the default execution space.

Step 2: compute the sum of the matrix

Use the Kokkos::parallel_reduce to compute the sum of the matrix and add the code to print the result.

Step 3: compute the maximum of the matrix

Use Kokkos::parallel_reduce and the reducer Kokkos::Max to compute the max element and add the code to print the result.

Step 4: timers

Add timers to measure the time spent to compute the sum and the max of the matrix.

Step 5: compile and run the program

Compile the program using the OpenMP backend and run it. Check the output and the timings.

If you have access to a GPU, compile the program using a GPU backend and run it. Check the output and compare them with the OpenMP backend.