Adds Math::sin_cos() - This computes sine and cosine of the same angle together, which reduces cycles and should improve performance slightly.

[mcdubhghlas]

I also replaced certain individual calls of sin and cos with sin_cos. I did not do every single instance of sine and cosine because it is mostly not necessary in every case and I did not want to add more code to review. Plus I wanted to not touch tons of files. We can, if someone wanted to, but the pay off is a one-off and I don't think it will be impactful outside of hot paths.

Summary by CodeRabbit

• Performance Improvements
  • Optimized trigonometric usage by introducing Math::sin_cos (leveraging compiler builtins when available).
  • Updated rotation, Euler/axis-angle conversions, circle sampling, particle initialization/orientation, and various rendering/audio/physics computations to use paired sine/cosine calculations.
  • Improves efficiency in drawing arcs/circles, generating particle motion, computing collision contact points, and running audio effects, with no expected visible behavior changes.

[coderabbitai]

No actionable comments were generated in the recent review. 🎉

ℹ️ Recent review info

⚙️ Run configuration

Configuration used: Path: .coderabbit.yaml

Review profile: CHILL

Plan: Pro

Run ID: 0d0396fa-c250-4027-bfe8-ab30d3694ad8

📥 Commits

Reviewing files that changed from the base of the PR and between 68a0610 and 4f5a0f3.

📒 Files selected for processing (3)

• scene/2d/cpu_particles_2d.cpp
• servers/audio/effects/audio_effect_pitch_shift.cpp
• servers/audio/effects/audio_effect_spectrum_analyzer.cpp

🚧 Files skipped from review as they are similar to previous changes (3)

• servers/audio/effects/audio_effect_spectrum_analyzer.cpp
• servers/audio/effects/audio_effect_pitch_shift.cpp
• scene/2d/cpu_particles_2d.cpp

---

Walkthrough

This PR adds Math::sin_cos and switches multiple math, physics, rendering, particle, and audio code paths to compute sine and cosine together instead of separately.

Changes

Math::sin_cos rollout

Layer / File(s)	Summary
sin_cos helper definition core/math/math_funcs.h	Adds Math::sin_cos template overloads for float and double, with builtin support when available.
Core math types core/math/basis.cpp, core/math/quaternion.cpp, core/math/vector2.cpp, core/math/transform_2d.cpp, core/math/transform_2d.h	Basis, Quaternion, Vector2, and Transform2D rotation and skew code now uses paired sine/cosine values.
Physics collision solvers modules/godot_physics_3d/godot_collision_solver_3d.cpp, modules/godot_physics_3d/godot_collision_solver_3d_sat.cpp	Circle support and contact point generation now uses sin_cos.
CPU particle math scene/2d/cpu_particles_2d.cpp, scene/3d/cpu_particles_3d.cpp	Particle rotation, emission placement, hue rotation, and basis construction now use shared sine/cosine results.
Canvas and rendering paths scene/main/canvas_item.cpp, scene/resources/style_box_flat.cpp, servers/rendering/renderer_canvas_cull.cpp, servers/rendering/renderer_rd/renderer_compositor_rd.cpp, servers/rendering/renderer_scene_cull.cpp	Arc/circle drawing, rounded rectangles, screen rotation constants, circle polygons, and spotlight geometry now use sin_cos.
Audio FFT computations servers/audio/effects/audio_effect_pitch_shift.cpp, servers/audio/effects/audio_effect_spectrum_analyzer.cpp	Pitch-shift synthesis and FFT twiddle factors now use Math::sin_cos.

Estimated code review effort: 2 (Simple) | ~15 minutes

🚥 Pre-merge checks | ✅ 4 | ❌ 1

❌ Failed checks (1 warning)

Check name	Status	Explanation	Resolution
Docstring Coverage	⚠️ Warning	Docstring coverage is 14.29% which is insufficient. The required threshold is 80.00%.	Write docstrings for the functions missing them to satisfy the coverage threshold.

✅ Passed checks (4 passed)

Check name	Status	Explanation
Description Check	✅ Passed	Check skipped - CodeRabbit’s high-level summary is enabled.
Title check	✅ Passed	The title accurately describes the main change: adding Math::sin_cos() and using it for performance-sensitive trig computations.
Linked Issues check	✅ Passed	Check skipped because no linked issues were found for this pull request.
Out of Scope Changes check	✅ Passed	Check skipped because no linked issues were found for this pull request.

✨ Finishing Touches

🧪 Generate unit tests (beta)

• Create PR with unit tests

---

Thanks for using CodeRabbit! It's free for OSS, and your support helps us grow. If you like it, consider giving us a shout-out.

❤️ Share

• X
• Mastodon
• Reddit
• LinkedIn

_{Comment @coderabbitai help to get the list of available commands.}

[coderabbitai]

🧹 Nitpick comments (2)

servers/audio/effects/audio_effect_spectrum_analyzer.cpp (1)

84-87: 📐 Maintainability & Code Quality | 🔵 Trivial | 💤 Low value

Consider using float instead of real_t for sc_sin/sc_cos.

Same observation as the smbFft in audio_effect_pitch_shift.cpp: all surrounding variables are float, so declaring sc_sin/sc_cos as real_t is inconsistent and may cause an unnecessary promote-narrow if real_t is double.

♻️ Proposed refactor

-	real_t sc_sin, sc_cos;
+	float sc_sin, sc_cos;

🤖 Prompt for AI Agents

Verify each finding against current code. Fix only still-valid issues, skip the
rest with a brief reason, keep changes minimal, and validate.

In `@servers/audio/effects/audio_effect_spectrum_analyzer.cpp` around lines 84 -
87, The `AudioEffectSpectrumAnalyzer` FFT twiddle setup uses `real_t` for
`sc_sin` and `sc_cos`, which is inconsistent with the surrounding `float`
variables and may introduce unnecessary conversions. Update the `Math::sin_cos`
locals in the spectrum analyzer code path to use `float`, matching the `smbFft`
style in `AudioEffectPitchShift`, and keep the subsequent assignments to `wr`
and `wi` in the same `float` type flow.

servers/audio/effects/audio_effect_pitch_shift.cpp (1)

262-265: 📐 Maintainability & Code Quality | 🔵 Trivial | 💤 Low value

Consider using float instead of real_t for sc_sin/sc_cos.

All surrounding variables in smbFft (wr, wi, arg, ur, ui) are float, but sc_sin/sc_cos are declared as real_t. If real_t is configured as double, this triggers the sin_cos(float, double&, double&) overload which computes in float via sincosf then promotes to double, only to narrow back to float on assignment to wr/wi. Using float directly is consistent with the rest of the function and avoids the unnecessary promotion.

♻️ Proposed refactor

-		real_t sc_sin, sc_cos;
+		float sc_sin, sc_cos;

🤖 Prompt for AI Agents

Verify each finding against current code. Fix only still-valid issues, skip the
rest with a brief reason, keep changes minimal, and validate.

In `@servers/audio/effects/audio_effect_pitch_shift.cpp` around lines 262 - 265,
The smbFft implementation in audio_effect_pitch_shift.cpp uses real_t for
sc_sin/sc_cos even though the surrounding values are float. Update the sc_sin
and sc_cos declarations in smbFft to float so Math::sin_cos uses the float
overload consistently, matching wr, wi, arg, ur, and ui and avoiding the extra
promote/narrow path.

🤖 Prompt for all review comments with AI agents

Verify each finding against current code. Fix only still-valid issues, skip the
rest with a brief reason, keep changes minimal, and validate.

Nitpick comments:
In `@servers/audio/effects/audio_effect_pitch_shift.cpp`:
- Around line 262-265: The smbFft implementation in audio_effect_pitch_shift.cpp
uses real_t for sc_sin/sc_cos even though the surrounding values are float.
Update the sc_sin and sc_cos declarations in smbFft to float so Math::sin_cos
uses the float overload consistently, matching wr, wi, arg, ur, and ui and
avoiding the extra promote/narrow path.

In `@servers/audio/effects/audio_effect_spectrum_analyzer.cpp`:
- Around line 84-87: The `AudioEffectSpectrumAnalyzer` FFT twiddle setup uses
`real_t` for `sc_sin` and `sc_cos`, which is inconsistent with the surrounding
`float` variables and may introduce unnecessary conversions. Update the
`Math::sin_cos` locals in the spectrum analyzer code path to use `float`,
matching the `smbFft` style in `AudioEffectPitchShift`, and keep the subsequent
assignments to `wr` and `wi` in the same `float` type flow.

---

ℹ️ Review info

⚙️ Run configuration

Configuration used: Path: .coderabbit.yaml

Review profile: CHILL

Plan: Pro

Run ID: 1553b08b-27c8-4218-ad72-fa90c4572700

📥 Commits

Reviewing files that changed from the base of the PR and between d93bf62 and 68a0610.

📒 Files selected for processing (17)

• core/math/basis.cpp
• core/math/math_funcs.h
• core/math/quaternion.cpp
• core/math/transform_2d.cpp
• core/math/transform_2d.h
• core/math/vector2.cpp
• modules/godot_physics_3d/godot_collision_solver_3d.cpp
• modules/godot_physics_3d/godot_collision_solver_3d_sat.cpp
• scene/2d/cpu_particles_2d.cpp
• scene/3d/cpu_particles_3d.cpp
• scene/main/canvas_item.cpp
• scene/resources/style_box_flat.cpp
• servers/audio/effects/audio_effect_pitch_shift.cpp
• servers/audio/effects/audio_effect_spectrum_analyzer.cpp
• servers/rendering/renderer_canvas_cull.cpp
• servers/rendering/renderer_rd/renderer_compositor_rd.cpp
• servers/rendering/renderer_scene_cull.cpp

[Arctis-Fireblight]

Fantastic work McDubh!

I am seeing a pretty good speed up here compared to #1273 .
We were sitting at 253FPS on the last benchmark, and this PR raises it to 264!