README

Studying the speedup on matrix multiplication achieved by multithreading and multiprocessing

matmul.c implements matrix multiplication in various modes (single threaded, multi-threading and multi-processing) and shows the speedup achieved by using the above said methods

Compiling

gcc -pthread matmul.c -o matmul

Usage

./matmul --ar <rows_in_A>  --ac <cols_in_A> --br <rows_in_B>  --bc <cols_in_B> [--interactive]

--ar : Specifies number of rows in matrix A
--ac : Specifies number of columns in matrix A
--br : Specifies number of rows in matrix B
--bc : Specifies number of columns in matrix B
--interactive : Run the program in interactive, i.e., user can give the input for matrices A and B. In interactive mode the user is asked for input for each mode and product of the given inputs is also displayed.
matmul -h :shows te usage of options and parameters

By default the program generates random matrices of the given dimensions and displays the speedup and time taken only anad not the result of multiplication

Example Outputs

Interactive Mode

./matmul --ar 3 --ac 3 --br 3 --bc 3 --interactive
Enter A:
1 2 3
4 5 6
7 8 9
Enter B:
1 0 0
0 1 0
0 0 1
Result:
1 2 3 
4 5 6 
7 8 9 
Enter A:
1 2 3
4 5 6
7 8 9    
Enter B:
1 0 0
0 1 0
0 0 1
Result:
1 2 3 
4 5 6 
7 8 9 
Enter A:
1 2 3
4 5 6
7 8 9
Enter B:
1 0 0
0 1 0
0 0 1
Result:
1 2 3 
4 5 6 
7 8 9 
Time taken for single threaded: 1 us
Time taken for multi process: 1 us
Time taken for multi threaded: 1 us
Speedup for multi process : 1 x
Speedup for multi threaded : 1 x

Non-interactive Mode

./matmul --ar 100 --ac 100 --br 100 --bc 100
Time taken for single threaded: 10 us
Time taken for multi process: 14 us
Time taken for multi threaded: 14 us
Speedup for multi process : 0.71 x
Speedup for multi threaded : 0.71 x

./matmul --ar 300 --ac 300 --br 300 --bc 300
Time taken for single threaded: 329 us
Time taken for multi process: 54 us
Time taken for multi threaded: 15 us
Speedup for multi process : 6.09 x
Speedup for multi threaded : 21.93 x

./matmul --ar 500 --ac 500 --br 500 --bc 500
Time taken for single threaded: 2365 us
Time taken for multi process: 133 us
Time taken for multi threaded: 57 us
Speedup for multi process : 17.78 x
Speedup for multi threaded : 41.49 x

KushagraIndurkhya / matrix-multiplier-c