search

QuazarTech / DG_Maxwell

Solution of Maxwell equations using Discontinuous Galerkin method.
GNU General Public License v3.0
9 stars 5 forks source link

Large error in calculation of double integral. #61

Open amanabt opened 7 years ago

amanabt commented 7 years ago

Commit: 3b42c3fd2b0715a445de86f70a296ee7fdf733ef

Issue: The error in the calculation of double integral image is of the order of image in the test_wave_equation_2d.py::test_A_matrix unit test.

As of now, the cause of the issue is not known.

amanabt commented 7 years ago

The reference values were calculated using the LGL points with an accuracy of 1e-4, whereas, the unit test was using LGL points with an accuracy of machine error. The difference in accuracy of the LGL points lead to large errors b/w reference and test A matrix.

Recalculating the reference values using LGL points with higher accuracy reduced the error b/w reference and test A matrix to machine error.

Issue fixed in PR #44