Maystern/MatrixGuard

MatrixGuard: Secure and Efficient Matrix Operations on ARM and x86

Note: Click here for report.

The primary objective of the project is to implement efficient matrix multiplication algorithms for both ARM and x86 architectures, and to compare their performance using different optimization techniques, such as multi-threading, instruction set optimization, and compilation optimization.

If you find our work useful, please consider citing our work using the bibtex:

@Article{MatrixMultiplicationOptimization,
	author  = {Jiacheng Luo},
	title   = {MatrixGuard: Secure and Efficient Matrix Operations on ARM and x86},
	year    = {2023},
}

Running the project on the Ubuntu operating system is recommended, and Git and CMake tools are used for code version control, compilation and building.

The detailed steps for building the project code using Git and CMake are as follows:

1. Use the command git clone https://github.com/Maystern/MatrixGuard.git in the current directory to download the project code.
2. Use cd MatrixGuard command to navigate to the root directory of the project.
3. Execute the command sh Run_Matmul.sh in the root directory of the project to automatically compile the project code using CMake, which generates binary executable matmulTest and moves it to the ./build/bin directory of the project root folder.
4. In the root directory of the project, run the command ./matmulTest matmulSizeLists to perform and test matrix multiplication of size $matmulSize \times matmulSize$ as an example.
5. To enable and execute the optimization techniques, modify the src/CMakeLists.txt file in the root directory of the project.