*Apparently I messed up before and posted an unfinished question. I tried deleting it, don’t know how well that went. Anyway, here is the question:*

I have an assembled matrix M from a form `a`

, say

```
u = TrialFunction(V)
v = TestFunction(V)
a = f * u*v*dx
```

where `f`

(\equiv f) is some function. If I know f>\alpha > 0 , then M is positive definite and there is a unique Cholesky factorization M = LL^T .

Does fenics have a way for me to get the matrix L ? (or L^T for that matter?)

There seem to be preconditioners based on Cholesky factorization, but I can’t seem to find a method which allow me to extract the factor?