[petsc-users] a cost question for sparse factorization

Umut Tabak u.tabak at tudelft.nl
Sat Apr 28 04:56:43 CDT 2012


> All you will get are asymptotics, not sharp operation counts. For 2D, 
> the bounds are O(n^{1.5} flops and O(n log n) space; for 3D they are 
> O(n^2) flops and O(n^{4/3}) space. This is covered in Alan George's 
> book, http://www.cse.illinois.edu/courses/cs598mh/george_liu.pdf
>
Dear Jed,
I am not good at these cost computations, is 'n' the number of nonzeros 
in the matrices?



More information about the petsc-users mailing list