GitHub - ryanslam/mat_mult_speedup: Using parallel processing and various techniques, I was able to achieve a 41x speedup for matrix multiplication.

ryanslam / mat_mult_speedup Public

Notifications You must be signed in to change notification settings
Fork 0
Star 0

Using parallel processing and various techniques, I was able to achieve a 41x speedup for matrix multiplication.

Notifications

Name		Name	Last commit message	Last commit date
Latest commit History 1 Commit
log_files		log_files
.DS_Store		.DS_Store
Makefile		Makefile
README.txt		README.txt
matmult_omp.cpp		matmult_omp.cpp
report.pdf		report.pdf
run_mult.sh		run_mult.sh

Repository files navigation

Parallelized Matrix Multiplication:
    - Files/directories:
        * /log_files
        * matmult_omp.cpp
        * Makefile
        * run_mult.sh

    - How to run:
        * ./matmult_omp <nrows> <mcols> <mcols2> -t <num_threads>
            * Example: ./matmult_omp 9000 2500 3750 -t 20

    - Makefile:
        * Compiles and creates an executable for matmult_omp.cpp.
        * make clean: removes the executable.

    - run_mult.sh:
        * Script logged the results from my computations to a txt file.
    
    - /log_files:
        * Directory contains experiment results from runs on the different parameters.