Hello,

I am using the GMRES solver for a very large sparse system and trying to precondition it using the ILU0 preconditioner. I'm a little confused about how to achieve this with the setFillFactor function in the ILUT preconditioner. I'm not exactly sure how the fillFactor translates into the fill-in procedure. Can some give me a code snippet or clarify how to do this?


Thank you!

Siddharth

Looking at the code, ILUT keeps at most the m largest non-zeros per row of L (and U) , where:

m = ((mat.nonZeros()*m_fillfactor)/n + 1) / 2

and n is the number of rows.