Skip to content

README.md

Latest commit

 

History

History
194 lines (126 loc) · 4.84 KB

README.md

File metadata and controls

194 lines (126 loc) · 4.84 KB

2024_Research_MassifCheck

Qt desktop application for running and analyzing Valgrind Massif heap memory profiling output.

✨ Current Features

  • Select a C or C++ source file (.c or .cpp) to automatically compile it and analyze with Valgrind Massif
  • Select a compiled binary to analyze
  • Run Valgrind's Massif tool with custom options
  • Parse Massif output
  • (In Progress) Visualize heap memory usage

🚀 Getting Started

1. Prerequisites

  • Qt Creator (6.0+)
  • WSL (Windows Subsystem for Linux)
  • Ubuntu distribution inside WSL
  • g++ and gcc compilers installed inside WSL
  • Valgrind installed in WSL

2. Installing Qt Creator

Download and install from https://www.qt.io/download

✅ Recommended: Use MSVC or MinGW kits for Desktop development

3. Setting up WSL and Valgrind

a. Enable WSL

Open PowerShell as Administrator and run:

wsl --install

b. Set Ubuntu as the Default WSL Distribution if you have multiple

wsl -l -v

This will list all installed distros, for example:

  NAME            STATE           VERSION
* Ubuntu-22.04    Running         2
  Debian          Stopped         2

Then set Ubuntu as default:

wsl --set-default Ubuntu-22.04

c. Install build-essential tools in Ubuntu

Open powershell and launch wsl

wsl

Then inside wsl run

sudo apt update
sudo apt install build-essential

The build-essential package includes gcc, g++, and other tools required for compiling C/C++ programs.

d. Install Valgrind in Ubuntu

Open powershell and launch wsl

wsl

Then inside wsl run

sudo apt update
sudo apt install valgrind

Verify installation:

valgrind --version

🛠️ Building the Project

  • Open the project folder containing CMakeLists.txt in Qt Creator

  • Configure with a Desktop kit

  • Click the Run button ▶️

💡 Make sure your Valgrind-executed binaries are accessible from WSL paths if running through Windows UI.

🧪 Automated Massif Profiling for C and C++ Tests with run_massif_on_tests.sh (Optional)

The run_massif_on_tests.sh helper script is optional and automates the process of compiling and running C and C++ tests with Valgrind's Massif tool to generate profiling data. This script can be used as a faster way to generate Massif output files, but all the same functionality is available through the Qt application itself.

🔍 What it does:

  • Searches for .c and .cpp files listed in the test/ directory.

  • Compiles them using the appropriate compiler:

    • gcc for .c files
    • g++ for .cpp files

  • Automatically creates the massif_files/ directory if it doesn't exist.

  • Runs each compiled test through valgrind --tool=massif.

  • Saves the output as massif.out.<n> files inside the massif_files/ directory.

📂 Example file structure:

├── QT_Massif_App/
├── test/
│   ├── test1.c
│   ├── test2.c
│   └── test3.cpp
├── massif_files/
│   ├── massif.out.1
│   ├── massif.out.2
│   └── massif.out.3
└── run_massif_on_tests.sh

💡 This script is helpful for quickly generating Massif profiling data for test binaries outside the Qt application interface.

⚙️ How to use:

  1. (Optional) Predefined Test Files:

    • The script is currently hardcoded to run the following test files: test1.c, test2.c, test3.cpp.
    • To add your own test files, simply edit the TEST_FILES array in the script. For example:
    TEST_FILES=("test1.c" "test2.c" "test3.cpp" "my_new_test4.c")
    • The script will automatically detect the correct compiler based on the file extension (gcc for .c and g++ for .cpp).

    Note: This step is optional. If you don't need to modify the test files, you can skip it and use the default ones.

  2. Make sure run_massif_on_tests.sh is executable:

    chmod +x run_massif_on_tests.sh

  3. Run the script:

    ./run_massif_on_tests.sh

The script will automatically compile and run each test, saving the profiling results in the massif_files/ directory.

📌 Notes

  • Valgrind runs only on Linux-based systems (including WSL).

  • Visualization panel and full report view are currently in development.

  • The project is modular and open to contributions.

👤 Contributors