Interactive Guide: OpenMP in VS Code

?

Why is Parallel Programming Needed?

The End of the "Free Lunch"

For decades, computers got faster because individual processor cores got faster (Moore's Law). Old software would automatically speed up on new hardware. Around 2005, this stopped. It became too difficult to make single cores faster due to heat and power limits. The industry's solution was to add more cores to a single chip. Today, even a standard phone has multiple cores, but a normal program can only use one at a time, leaving the rest idle. Parallel programming is the key to unlocking the full power of modern hardware.

What is OpenMP?

OpenMP (Open Multi-Processing) is one of the simplest ways to write parallel code. It's a set of commands (called #pragma directives) that you add to your C/C++ code. These pragmas are simple instructions to the compiler, telling it which parts of your code (like a `for` loop) can be safely run on multiple cores at the same time. You don't have to manage threads manually OpenMP handles the complex work for you.

1

Prerequisites

Before writing code, you need to set up your development environment. This involves installing VS Code, the necessary extension, and a C/C++ compiler that supports OpenMP. Select your operating system below to see the specific compiler instructions.

1A. Install Visual Studio Code & C/C++ Extension

Download VS Code from code.visualstudio.com. Then, inside VS Code, go to the Extensions view (`Ctrl+Shift+X`) and install the `C/C++` extension from Microsoft.

1B. Install Compiler for Windows (MinGW-w64)

MinGW-w64 is a development toolchain that provides a version of the GCC compiler for Windows. This is necessary because Windows does not come with a built-in C/C++ compiler that supports OpenMP, so we install this to get the required tools.

Download and run the MSYS2 installer from msys2.org.
Open the MSYS2 terminal and run `pacman -Syu`.
Install the toolchain with: `pacman -S --needed base-devel mingw-w64-ucrt-x86_64-toolchain`.
Add the compiler to your Windows Path. This is typically `C:\msys64\ucrt64\bin`.

1B. Install Compiler for Linux (build-essential)

GCC is usually pre-installed. If not, open your terminal and run the appropriate command for your distribution:

sudo apt update && sudo apt install build-essential

1C. Install and Setup MPI

MPI (Message Passing Interface) is required if you want to run distributed-memory parallel programs. Follow these steps for Windows setup:

Install prerequisites - MS-MPI Redistributable + SDK.
Download and install both the MS-MPI Redistributable (runtime) and MS-MPI SDK (headers & libs). Official Microsoft Download Center.
Once both SDK and setup is installed, verify these paths exist:
- C:\Program Files\Microsoft MPI\Bin → Add this to your PATH variable.
- C:\Program Files (x86)\Microsoft SDKs\MPI\Include
- C:\Program Files (x86)\Microsoft SDKs\MPI\Lib\x64
Edit system environment variables and add the remaining two to User Variables:
- INCLUDE → C:\Program Files (x86)\Microsoft SDKs\MPI\Include
- LIB → C:\Program Files (x86)\Microsoft SDKs\MPI\Lib\x64
Search for x64 Native Tools Command Prompt for VS and run:
set INCLUDE
set LIB
You should see your MPI paths listed.
if you dont see x64 Native Tools download it using Visual Studio Community with the Desktop Development with C++ workload.
Once Installed - Under Workloads Tab - Select Desktop Development with C++ - Click Install and let it download the C++ compiler toolchain. After installation search in Start Menu for x64 Native Tools Command Prompt for VS
Open x64 Native Tools Command Prompt for VS and Navigate to the folder where your MPI code is saved (example if file is in `E:\PC_LAB`): cd /d "E:\PC_LAB"
Now you can compile the code using cl 5.c msmpi.lib
Run using mpiexec -n 2 5.exe

2

Project Setup

Create a new folder on your computer (e.g., `MyOpenMPProject`).
Open this folder in VS Code (`File > Open Folder...`).
Create a new source file (`File > New File`) and save it with a `.c` extension (e.g., `main.c`).

3

Configure Build Task

Create a `tasks.json` file in your `.vscode` folder to tell VS Code how to compile your program with the `-fopenmp` flag.

{
  "version": "2.0.0",
  "tasks": [
    {
      "label": "C/C++: Build with OpenMP",
      "type": "shell",
      "command": "gcc",
      "args": [ "-g", "-fopenmp", "${file}", "-o", "${fileDirname}/${fileBasenameNoExtension}" ],
      "group": { "kind": "build", "isDefault": true }
    }
  ]
}

4

Compile & Run

Compile the Code

The easiest method is to use the build task. With your `.c` file open, press `Ctrl+Shift+B`. Alternatively, you can compile manually from the terminal. For a file named `my_program.c`, the command is:

gcc -fopenmp my_program.c -o my_program

Run the Executable

After a successful compilation, an executable file is created. Run it from the terminal:

# On Windows
./my_program.exe

✓

Troubleshooting & FAQ

Cause: The compiler cannot find the OpenMP library. This usually means your build task is missing the `-fopenmp` flag.

Cause: The terminal cannot find your compiler. The folder containing `gcc.exe` is not in your system's PATH environment variable.

OpenMP is for shared-memory parallelism (multi-core CPU on one machine). It's easy to use with `#pragma` directives. MPI is for distributed-memory parallelism (across multiple machines in a cluster). It's more complex, involving explicit message passing.

Interactive Guide to OpenMP