Convolution Kernal on 2D matrix

Details

• Kernal size: 3x3
• Matrix size: 80x300

Implementation steps

• Load distance_vector data from file
• Run average execution time
• Improve functionality with GPU
• Quantify performance

Running the software

$ ./run.sh 0

Flags

A flag can be added to specify the mode in the bash file Naive

$ ./run.sh 0

Optimized

$ ./run.sh 1

flags:

• -O3 : -Ofast | Speeds up the program signifanctly
• -g | Debug mode

Useful sources

• https://github.com/Xilinx/Vitis-HLS-Introductory-Examples/blob/master/

Contact

For more information about the project, you can contact Stijn Brugman (s.r.d.brugman@student.utwente.nl).