5 Jan
2015
5 Jan
'15
9:53 a.m.
On 5 Jan 2015, at 09:48, Colin Cotter <colin.cotter@imperial.ac.uk> wrote:
Does it help if the matrices are columnwise-blocks? I could assemble the entire matrix in a "column kernel" and then perform surgery on it?
You can do whatever you like to the "element tensor" block before insertion. However, there's no way of getting the element tensor /out/ of a CSR matrix to do surgery on it in a parallel loop. Lawrence