SciPy LinearLUSolver has artificial limit of 10 iterations

[guyer]

Since iterations were introduced, the SciPy LinearLUSolver has been constrained to a maximum of 10 iterations. None of the other LinearLUSolver implementations do this.

[wd15]

It should never really require more that 2 or 3 iterations from my recollection. LU shouldn't really need to iterate at all, but for big systems or poorly condition iterating helps with numerical round off.

[guyer]

Got it. Still, seems like people should be able to set iterations= to whatever they want, even if it's pointless.