[Day 1 of 100] Tuesday ,16 Sept 2014
Goals for the day :
1. Linear Algebra - cost of algorithms (dot product based ,axpy based) for
- y=Ax+y => 2mn flops for dot or axpy
- y=Ax+y => 2 * Sum :k=0 to (n-1) of (n+1 ) =2 n(n+1) for mod algo(dotp) 4.2.3 & 2n*n for axpy based 4.2.3
- y=Lx+y => n(n+1) dotp /axpy
- y=Ux+y => n(n+1) dot product based
- y=Ux+y =>n(n+1) axpy
- y=Atx+y =>n(2m+1) dotp based ,
- y=Atx+y => 2mn for axpy
- y=Ltx+y => Sum:k=0 to (n-1) of (2k+3) = n(n+2) for dotp
- y=Ltx+y => Sum:k=0 to (n-1) of (2k+2) = n(n+1) for axpy
- y=Utx+y => Sum:k=0 to (n-1) of (2k+3) = n(n+2) for dotp
- y=Utx+y => Sum:k=0 to (n-1) of ( 2n-2k) = n(n+1) for axpy
- x=Lx Sum:k=0 to (n-1) of 2(n-k) = n(n+1) for dotp
- x=Lx Sum:k=0 to (n-1) of (2k+1) for axpy
- x=Ltx Sum:k=0 to (n-1) of(2n-2k) for dotp
- x=Ltx Sum:k=0 to (n-1) of (2k+1) for axpy
- x=Ux => Sum:k=0 to (n-1) of (2n-2k) = n(n+1) for dotp
- x=Ux => Sum:k=0 to (n-1) of (2k+1) = n*n for axpy
- x=Utx => Sum:k=0 to (n-1) of (2k+2) for dotp
- x=Utx => Sum:k=0 to (n-1) of () for axpy unsure
Need to understand/correct the dotp cost discrepancy(psi := dotp+psi) in the pdf
Total time taken = 3.5 hrs
2. Python3 tutorial
immutability of strings, lists,slicing,
for--else usage
Total time taken = 1 hr
3. Bezier Curve fundamentals
parametric fns , control variable t ,varying t to get x,y coords
