Improve DIM performance using padding in FFT

[MatthewFilipovich]

Optimize conv2d() to improve performance for power-of-two grid shapes. Adds padding in FFT.

Closes #68

[codecov]

Codecov Report

✅ All modified and coverable lines are covered by tests.
✅ Project coverage is 99.92%. Comparing base (52d9eef) to head (748dbe0).
⚠️ Report is 1 commits behind head on main.

Additional details and impacted files

@@           Coverage Diff           @@
##             main      #92   +/-   ##
=======================================
  Coverage   99.92%   99.92%           
=======================================
  Files          38       38           
  Lines        1304     1305    +1     
=======================================
+ Hits         1303     1304    +1     
  Misses          1        1           

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

🚀 New features to boost your workflow:

• ❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.

[Copilot]

Pull request overview

This PR aims to improve Direct Integration Method (DIM) propagation performance by allowing FFT-size padding in conv2d_fft, and enabling it in DIM.

Changes:

• Extend conv2d_fft with an extra FFT-size padding parameter and adjust FFT sizing logic.
• Use padding=1 in DIM propagation to potentially get faster FFT sizes.
• Add a new functional test covering a larger-kernel convolution scenario.

Reviewed changes

Copilot reviewed 3 out of 3 changed files in this pull request and generated 5 comments.

File	Description
tests/functional/test_conv2d.py	Adds a new “large kernel” correctness test for conv2d_fft.
src/torchoptics/propagation/direct_integration_method.py	Enables FFT padding in DIM by calling conv2d_fft(..., padding=1).
src/torchoptics/functional/functional.py	Adds padding parameter to conv2d_fft and uses it to set FFT sizes.

---

💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.