Similar to #316 , we are seeing some hard to reproduce test failures in the xpt sens test in test_trans_beam. I can't be 100% sure that this step size change fixes the error but it does reduce the largest relative error by roughly 2 orders of magnitude:
Similar to #316 , we are seeing some hard to reproduce test failures in the xpt sens test in
test_trans_beam
. I can't be 100% sure that this step size change fixes the error but it does reduce the largest relative error by roughly 2 orders of magnitude:Before:
After: