Re: [deal.II] Trouble with vector valued matrix assembly optimizations

[Wolfgang Bangerth]

On 2/17/22 14:50, Kyle Schwiebert wrote:
> My confusion is amplified by the block which I have labeled "Just do
> some checking..." There is no output coming from those print statements,
> which seems to support my understanding of what is happening.

This makes sense to me, though :-) It could be because component_i and
component_j are never different, or because the complicated formula you
show is never nonzero. The latter looks to be the case: if i and j
correspond to different vector components, then one or the other factor
of each of the products are always zero, and consequently the whole term
is always zero if i and j correspond to different vector components.

So I think it is right that the block never produces output. Why do you
think that it should?

Best
W.

--
------------------------------------------------------------------------
Wolfgang Bangerth email: bang...@colostate.edu
www: http://www.math.colostate.edu/~bangerth/