Capstone Project: Diffusion Simulation with Cuda, MPI

https://github.com/khanhtq2101/ParallelDiffusionSimulation.git
Implementing Gauss Seidel algorithm by parallel programming with cuda and demo visualization.

How to run?

1. CUDA:

• Create result folder:

   mkdir -p ./results/CUDA

• Compile and run the GaussSeidel.cu file:

   nvcc -o GaussSeidel_CUDA GaussSeidel.cu  
   ./GaussSeidel_CUDA

The output will be saved on /results/CUDA folder.

2. MPI:

• Create result folder:

  mkdir -p ./results/MPI

• Compile the GaussSeidel_MPI.c file:

   mpicc -o GaussSeidel_MPI GaussSeidel_MPI.c  

• Run the compiled file:

  mpirun -np 16 GaussSeidel_MPI

The output will be saved on /results/MPI folder.

3. Display the results:

• Run file text_to_imgage.py file to convert result in text to image.

   python text_to_image.py 

• Run file display.py if you want to display process in some iterations.

   python display.py

• Run file create_gif.py to create gif for diffusion process.

   python create_gif.py 

4. Results:

• Top down visulization: Initialization: 80 at centers and 25 around. Boundary condition: 100 at top right and 0 at bottom left.
  
  Gift display:
  
• Top right visualization: Initialization: 80 at centers and 25 around. Boundary condition: 100 at top down, and 0 at left and right side.