This is a simple method to campare cuda and a CPU based programming language (C++).
In this code, you can choose the size of the matrix and the values, and at the end you can see the performance and results by clicking on "PRESS TO SEE RESULT".
And after run the programm:
No LIMIT MM: You can use all sizes that you want as size of a matrix.
LIMIT MM: maximum size of matrix = 316;
I would be happy to provide any further information if necessary.