Skip to content

dshimo/symm.lu

Repository files navigation

CS378 Project: symm.lu

Wesley Chung, David Shi, Hiep Vu
Symmetric matrix-matrix multiplication
C := AB + C
A symmetric, stored in upper-triangular part
Symm.lu( A, B, C )

Below, we have a generic loop invariant for a matrix A divided into four quadrants. Alt text

Food for thought, is it possible to do something with a patrix partitioned as such...? Alt text

Proof for Loop Invariant (where A_{BR} is 0) Professor says this is hardest Alt text