Dense_MatMul_OpenMP

Usage

clang -fopenmp -L/usr/local/opt/llvm/lib -I/usr/local/opt/llvm/include <filename.c> -o && ./<output>

Speeding up dense matrix multiplication using openMP and blocking (ensures less cache miss)

• Improves naive matrix multiply from 550ms to 44ms (6ms using -O3 flag)
• Hardware: 15 inch Macbook Pro 2019

Naive matrix multiplication

• Source- M-M_naive.c

Time for the loop = 550.06 milliseconds
Element 0,0 = 131.472137
Element 63,63 = 131.706528
Element 511,511 = 133.296707

Matrix multiplication with blocking

• Source- M-M_blocked.c

Time for the loop = 323.61 milliseconds
Element 0,0 = 131.472122
Element 63,63 = 131.706589
Element 511,511 = 133.296707

Matrix multiplication with OpenMP

• Source- M-M_omp.c

Time for the loop = 91 milliseconds
Element 0,0 = 131.472137
Element 63,63 = 131.706604
Element 511,511 = 133.296707

Matrix multiplication with blocking and OpenMP

• Source- M-M_blocked_omp.c

Time for the loop = 44 milliseconds
Element 0,0 = 131.472137
Element 63,63 = 131.706589
Element 511,511 = 133.296707