search

Mayank447 / Pseudo-Spectral-Discretization

GNU General Public License v3.0
0 stars 1 forks source link

Fix dimensionful normalisation #19

Closed chillenzer closed 2 months ago

chillenzer commented 2 months ago

I stumbled across the need for normalisation in #12's test_real_basis.py and that shouldn't be necessary. Turns out I'm really not used to working with dimensionful quantities. Here's the fix. With this correct normalisation of the eigenfunctions, the manual normalisation in test_real_basis.py is no longer necessary.

Mayank447 commented 2 months ago

Hey, so the normalization there is now redundant. I added the normalization when we were dealing making the Fourier transforms normalization in FFT in numpy

Mayank447 commented 2 months ago

Okay, so the normalization should be by sqrt(num_lattice_points) since the vector length is num_lattice point and if you take the inner product of any eigenfunction with it's complex conjugate (e.g. zero mode) it will be equal to num_lattice_points.

Second this thing will already implemented in #12.