OpenMP Sample for Matrix Multiplication

This sample is a multithreaded implementation of matrix multipication using OpenMP*. In the implementation, each thread can concurrently compute some submatrix of the product without needing OpenMP data or control synchronization. The algorithm uses OpenMP to parallelize the outer-most loop.

System Requirements

Hardware:
- Any Intel® processor like Intel® Core™ i3, i5, or i7 processors and Intel® Xeon® E3 or E5 processor family
For Microsoft* Windows*:
- Microsoft Visual Studio* 2013, 2015 or 2017 Professional Edition or above
- Intel® Parallel Studio XE 2019 Composer Edition for C++ Windows* or newer
For Linux*:
- GNU* GCC 4.5 or newer
- Intel® Parallel Studio XE 2019 Composer Edition for C++ Linux* or newer
For macOS*:
- macOS* 10.12 or above
- Xcode* 8.x or above
- Intel® Parallel Studio XE 2019 Composer Edition for C++ macOS* or newer

Build Instructions

For Microsoft Visual Studio users:

Open Visual Studio and load the solution file openmp_samples_20xx.sln which corresponds to the version of Visual Studio.
Choose a release configuration (either x64 or Win32). Build the solution. Run the application, which is an OpenMP parallel program, and see execution time.
- Debug and Release: uses Intel C++ compiler
Turn off OpenMP as follows, in order to generate a serial application. Rebuild the solution. Run the application again, and record execution time of serial program.
- Project Properties -> C/C++ -> Language [Intel C++] -> OpenMP Support: set to No
  The warning "warning #3180: unrecognized OpenMP #pragma" will be emitted a few times during compilation.
Setting execution time of serial program as baseline, calculate speedup of OpenMP parallel program.
- Speedup = execution time of OpenMP parallel program / execution time of serial program
(Optionally) To build the solution with Visual C++ compiler, follow the instructions below.
- Right-click on the openmp_samples project in Solution Explorer -> Intel compiler -> Use Visual C++
- Rebuild the solution. Then run the application.

For Microsoft Windows Command Line users:

Open one of the following Intel C++ compiler command prompt from Start menu.
- IA-32 Visual Studio 201X environment
- Intel 64 Visual Studio 201X environment
Build and run the application with OpenMP support:
build
build run
Build without OpenMP support, and run the application as baseline:
build serial
build run serial
Calculate speedup of OpenMP parallel program.
- Speedup = execution time of OpenMP parallel program / execution time of serial program
(Optionally) To build the application with Visual C++ compiler, follow the instructions below.
- Open an appropriate Visual C++ command prompt from Start menu, either x86 or x64.
- Build with command build or build serial. They will invoke Visual C++ compiler to build the application. Then run the application with command build run.

For Linux or macOS users:

From a terminal window, navigate to the project folder
Set the environment variables for Intel® C++ compiler:
source <icc-install-dir>/bin/compilervars.sh <arg>
where <icc-install-dir> is the directory structure containing the bin directory and <arg> is ia32 or intel64
To compile with OpenMP support and generate a parallelized application:
make
To run the application:
make run
To compile without OpenMP support and generate a serial application as baseline:
make serial
To run the application:
make run
Calculate speedup of OpenMP parallel program.
- Speedup = execution time of OpenMP parallel program / execution time of serial program