Implement angular PDF evaluation for angle-energy distributions by itay-space · Pull Request #3550 · openmc-dev/openmc

itay-space · 2025-08-27T15:46:18Z

Description

This PR is the first step in breaking down the point detector tally project into smaller, reviewable pieces, following the large PR #3109
and the paper Development and benchmarking of a point detector in OpenMC.

As suggested during the earlier discussion (thanks @GuySten, @gridley, @shimwell !), the natural starting point is to add the ability to correctly calculate PDFs on the C++ side.

Accordingly, this PR introduces the get_pdf implementations for the distribution classes. This is essentially a clean cut from the original large PR, with the goal of keeping the scope focused and review manageable.

We plan to follow up with tests and additional pieces of the point detector work in subsequent PRs.

Checklist

I have performed a self-review of my own code
I have run clang-format (version 15) on any C++ source files (if applicable)
I have followed the style guidelines for Python source files (if applicable)
I have made corresponding changes to the documentation (if applicable)
I have added tests that prove my fix is effective or that my feature works (if applicable)

GuySten

Nice work @itay-space!

My review focus more on the code structure side of things.
I did not check the maths.

Co-authored-by: GuySten <62616591+GuySten@users.noreply.github.com>

Removed unnecessary break statement in probability sampling logic.

GuySten · 2026-03-02T20:19:18Z

I've removed all dynamic allocations.
@itay-space, can you check that you agree with my changes?

@paulromano, this PR is ready for review again.

Reconstructing the azimuth with acos(u.dot(v_ref_) / sqrt(1 - mu*mu)) loses the sign of the w_ref_ component, so directions at azimuth phi and 2π - phi collapse to the same value.

CoherentElasticAE::sample_energy_and_pdf returns zero whenever E_in > bragg_edges_.back(), but coherent elastic scattering is still valid above the highest Bragg edge; CoherentElasticXS::operator() continues to use the final cumulative structure factor there. Additionally, the call to bragg_edges_.slice(0, n) / factors_diff_.slice(0, n) is not valid for a 1-D Tensor. In the tensor API, slice(0, n) means two axis indices, not a range.

paulromano · 2026-03-13T16:43:14Z

@GuySten Thanks for the updates on this branch -- it was definitely looking much better. There were still a few issues that I found and fixed:

PolarAzimuthal::evaluate reconstructed phi in a way that loses the sign of the w component, so directions at phi and 2π - phi would evaluate to the same value (even if phi is asymmetric). I updated the logic to fix this.
CoherentElasticAE::sample_energy_and_pdf returned zero when E_in > bragg_edges.back() but coherent elastic scattering is still valid above the highest Bragg edge. Also, the tensor slicing was incorrect (slice(0, n) would slice single values on two dimensions, whereas we needed a range in 1D).
The thermal discrete PDF routines did not match the distributions that the actual sample routines generate. Namely, in the actual sampling, the discrete cosines are interpolated and then smeared using neighboring interpolated points. I've refactored the get_pdf_discrete functions to reflect this.

Let me know if you agree with the changes and if so, I think this is ready to merge!

GuySten · 2026-03-14T21:58:03Z

I agree with these changes.
Thanks @paulromano for the various improvements.

Implement get_pdf functionality

06ae298

itay-space changed the title ~~Implement get_pdf functionality~~ [Point Detector] Add distribution get_pdf functionality Aug 27, 2025

GuySten requested changes Aug 27, 2025

View reviewed changes

GuySten added Other Enhancement Point Detector and removed Other Enhancement labels Aug 27, 2025

itay-space and others added 6 commits August 28, 2025 08:32

Update include/openmc/distribution.h

78d841e

Co-authored-by: GuySten <62616591+GuySten@users.noreply.github.com>

Update include/openmc/particle_data.h

33a27a9

Co-authored-by: GuySten <62616591+GuySten@users.noreply.github.com>

Update include/openmc/particle_data.h

d968875

Co-authored-by: GuySten <62616591+GuySten@users.noreply.github.com>

Update src/physics.cpp

3b896c0

Co-authored-by: GuySten <62616591+GuySten@users.noreply.github.com>

Update src/physics.cpp

c034bff

Co-authored-by: GuySten <62616591+GuySten@users.noreply.github.com>

Update src/physics.cpp

a5d2f4c

Co-authored-by: GuySten <62616591+GuySten@users.noreply.github.com>

GuySten reviewed Aug 28, 2025

View reviewed changes

Comment thread src/distribution.cpp Outdated

Your Name and others added 5 commits August 31, 2025 14:50

Cleaning some code

ca5b44e

Resolving conversations

eed8b1c

cosmetic fixes

f2ecdd4

Merge branch 'develop' into get_pdf

71d5fa2

simplified code significantly

c3b4d82

GuySten requested changes Aug 31, 2025

View reviewed changes

itay-space and others added 8 commits August 31, 2025 18:06

Update src/distribution_angle.cpp

78a33cc

Co-authored-by: GuySten <62616591+GuySten@users.noreply.github.com>

ran clang format

6d70d39

fixes

acfd402

cosmetics

98ca32e

using openmc's fatal_error

26d2a56

Use accessors in powerlaw

7abe69c

clausing error for different distributions

d183cce

fixing function signature

37b41c4

GuySten requested changes Sep 1, 2025

View reviewed changes

GuySten reviewed Sep 1, 2025

View reviewed changes

Comment thread src/reaction_product.cpp Outdated

Update include/openmc/angle_energy.h

007dcf5

Co-authored-by: GuySten <62616591+GuySten@users.noreply.github.com>

Merge branch 'develop' into get_pdf

20ec56b

GuySten mentioned this pull request Jan 29, 2026

Testing point detectors #3757

Draft

5 tasks

GuySten and others added 7 commits January 29, 2026 20:02

fix for invalid energy distribution in elastic scattering case

64e23e3

Refactor probability sampling logic in reaction_product

1f36c89

Removed unnecessary break statement in probability sampling logic.

Merge branch 'develop' into get_pdf

8ee06a6

decrease dynamic allocations

d11f2e5

decrease dynamic allocations further

321d188

decrease dynamic allocations further

7b9777e

decrease

1f5e24e

GuySten requested a review from paulromano March 2, 2026 20:19

paulromano added 13 commits March 12, 2026 14:34

Merge branch 'develop' into pr/itay-space/3550

751202a

Store copy of Bragg edges to avoid allocation

80413b0

Add comments and docstrings

e3f2114

Avoid falling off non-void function without returning

02ae3ac

Fix PolarAzimuthal::evaluate

85f0fcd

Reconstructing the azimuth with acos(u.dot(v_ref_) / sqrt(1 - mu*mu)) loses the sign of the w_ref_ component, so directions at azimuth phi and 2π - phi collapse to the same value.

list clamp

722bdd4

Fix incoherent elastic/inelastic PDF calculation

57da912

Single implementation for get_pdf_disrete

88ff94f

Remove discrete_half_width

61563a6

Move evaluate_bin into implementation

ff29e60

Eliminate DoubleVector

986acd1

Fix up Doxygen comments

696fb65

paulromano changed the title ~~[Point Detector] Add distribution get_pdf functionality~~ Implement angular PDF evaluation for angle-energy distributions Mar 13, 2026

paulromano approved these changes Mar 13, 2026

View reviewed changes

GuySten merged commit 1578698 into openmc-dev:develop Mar 14, 2026
17 checks passed

GuySten mentioned this pull request May 6, 2026

Implement jacobian calculation and energy and angle transformation #3935

Draft

5 tasks

Conversation

itay-space commented Aug 27, 2025 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Description

Checklist

Uh oh!

GuySten left a comment • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

GuySten commented Mar 2, 2026

Uh oh!

paulromano commented Mar 13, 2026

Uh oh!

GuySten commented Mar 14, 2026

Uh oh!

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

7 participants

itay-space commented Aug 27, 2025 •

edited

Loading

GuySten left a comment •

edited

Loading