Poisson Solver does not actually run to convergence
MattFerraro opened this issue · comments
Matt Ferraro commented
Repeatably only the first iteration of the Poisson Solver will actually run to convergence of epsilon < 1e-5. Subsequent runs never converge and instead time out with epsilon ~ 0.05.
This probably isn't a huge problem but it indicates a weakness in the solver or a weakness in the conditioning. Perhaps the integral of the loss is no longer equal to zero and just needs to be re-centered?
Matt Ferraro commented
Yep, the fix was simply to shift the loss function up or down such that sum(loss) = 0 before handing it to the poisson solver. Same for the heightmap solver. It now reliably and quickly converges.