Open orenbenkiki opened 2 months ago
A
is assumed to be a linear operator on the object of b
. As a vector, b
is interpreted as its value vec(b)
. So it's not sized to work. What you're trying to do here is a batch of linear solves, which implicitly occurs via matrix \ matrix, but in a general sense with general linear operators it's making many more assumptions and we're not quite ready to handle the general case (think Krylov methods), so right now it avoids the batch case.
I thought it was a simple matter of inverting the T(A)A matrix once and applying it to the B matrix? And similarly to other solvers?
Describe the bug 🐞
I'm getting the error:
When solving a linear system MF * FG = MG.
Expected behavior
I expected
solve
to behave similarly to\
in this case.Minimal Reproducible Example 👇
Error & Stacktrace ⚠️
using Pkg; Pkg.status(; mode = PKGMODE_MANIFEST)
versioninfo()