Parallel Programming Lab - MIT CSE - Sem 6

• Lab 1 --> MPI Point-to-Point Communication (Missed it)
• Lab 2 --> MPI Point-to-Point Communication
• Lab 3 --> MPI Collective Communication
• Lab 4 --> MPI Collective Communication with Error Handling
• Lab 5 --> CUDA programs on arrays
• Lab 6 --> CUDA programs on matrices
• Lab 7 --> CUDA programs on strings
• Lab 8 --> CUDA programs on matrices (continued)

Running MPI Examples

Uses OpenMPI's wrapper compiler for C.

To compile

mpicc my_app.c -o my_app

Execute the serial / parallel jobs

Note: mpirun, mpiexec, and orterun are synonyms. Specify number of MPI processes -np

mpirun -np 4 my_app

Running CUDA Examples

Everything CUDA related was done using Google Colab's free Nvidia GPUs.

Copy the required .cu file contents to the cell with double magic %%cu on top in cuda_c_template.ipynb