How to code a function for matrix vector multiplication?
We code up and error-check our general matrix-vector multiplication function that accesses the entries of the matrix using column partitions.
00:00 -Think deeply about what we want our functions to do
00:19 -Start driver error checks: verify proper number of inputs
01:10 -To error check, create a library of benchmark problems
02:06 -Error check 1: verify proper number of inputs
02:53 -Teach yourself MATLAB
03:43 -Error check 2: verify proper value of input “Trans” variable
06:09 -Error check 3: verify alpha and beta are double-precision scalars
10:46 -Error check 4: verify that A is a double-precision 2D array
15:07 -Error check 5A: verify that vector y has proper dimensions
24:58 -Handle case where alpha equals zero
25:40 -Error check 5B: verify that vector x has proper dimensions
28:15 -Code general matrix-vector multiplication
28:45 -Code column axpy version matrix-column-vector multiplication
31:10 -Verify output for column axpy code works as we expect
35:01 -Code column dot version row-vector-matrix multiplication
38:02 -Verify output for column dot code works as we expect
41:12 -It takes a lot of work to get good code
42:32 -Challenge 1: write analogous function that processes matrix by rows
43:01 -Challenge 2: Share your reaction to this work
43:13 -Conclusion