- #1

dinaharchery

- 24

- 0

I am trying to learn matrix compression algorithms for sparse matrices. I understand the Compressed Row Storage (CRS) format but I am having difficulty with the Block Compressed Row Storage (BCRS) format - where the size of the sub-block(s) are 2x2.

Can anyone please point me to an algorithm/code for implementing BCRS 2x2? I am looking to take a standard 2-D floating-point matrix in C++ and converting it to the BCRS 2x2 format as well as Matrix-Vector multiplication using the BCRS 2x2 format.

I know there are a lot of programs for this out there but I really would like know the algorithm myself, so that I can learn this rather than re-implement someone elses code - although at this point I will take what I can get.

Thanks.