automatic mode rotation for bent waveguide #2028
Conversation
QimingFlex
changed the title
Thanks! I’ve validated the process locally with several cases involving differently oriented bends, and I’ll add those to the tests. This PR isn’t fully working yet, but it’s close. I’ll wrap up the code based on your feedback. |
QimingFlex
force-pushed
the
qiming/mode_rotation
branch
5 times, most recently
from
1353ea5
to
c1fd5c1
Compare
QimingFlex
force-pushed
the
qiming/mode_rotation
branch
from
e9eb835
to
00976a1
Compare
QimingFlex
force-pushed
the
qiming/mode_rotation
branch
2 times, most recently
from
6e51b86
to
4e6dbc7
Compare
tidy3d/plugins/mode/mode_solver.py
Outdated
| cmp_1, cmp_2 = source_names | ||
|
|
||
| for mode_idx in range(mode_index.size): | ||
| for f_idx, f_val in enumerate(f.values): |
There was a problem hiding this comment.
I would think it should be possible to vectorize if not along both of these dims, then at least over the frequency one. This could be left as an optimization for later though.
| } | ||
|
|
||
| # Transform fields from Cartesian to cylindrical coordinates | ||
| for i, _rho in enumerate(rho): |
There was a problem hiding this comment.
I suspect this could be slow in some cases, but we may have no choice, or do you have any ideas about vectorizing?
QimingFlex
force-pushed
the
qiming/mode_rotation
branch
5 times, most recently
from
b1c38bc
to
c5f79fc
Compare
QimingFlex
force-pushed
the
qiming/mode_rotation
branch
from
c5f79fc
to
2517468
Compare
| cmp_1, cmp_2 = source_names | ||
|
|
||
| for mode_idx in range(mode_index.size): | ||
| for f_idx, f_val in enumerate(f): |
There was a problem hiding this comment.
These two loops seem easy to vectorize out.
| if sym == 0: | ||
| colocate_coords[dim] = coords[1:-1] | ||
| else: | ||
| colocate_coords[dim] = coords[:-1] |
There was a problem hiding this comment.
Let's separate this out in a separate private method that we can call both here and in _colocate_fields to avoid code duplication.
@momchil-flex This is a very initial draft and not ready for review yet, but could you take a quick look at the code structure to see if there are any major mistakes so I can address them early on? I added a field
bend_correctionin mode_spec, and the process for mode rotation is mostly indata_rawinModeSolver.