avimit
Joined: 16 Nov 2005
Posts: 415
Helped: 68
Location: Fleet, UK
|
 24 Aug 2008 23:37 Re: Design 16 bit CPU for image processing application |
|
|
|
First thing is to delcare an array type which can hold a matrix.
Then you may write a function, which takes in 2 such matrices as inputs, and then multipies the elements of array as in matrix multiplication and returns the resulting array of results.
But I must warn that a 'function' is always a combinational implementation. By using a function for mat multiplication, you may end up a massvely parallel hardware, which will be quite bulky.
The other way is to think of a sequential way of multiplying elements, and storing results in memory or registers one by one. Thats what I did when I implemented Motion Estimation Algorithm for MPEG-4.
Hope it helps,
Kr,
Avi
|
|
