Refactor renderer implementations

[tarcila]

Consolidates duplicated ray traversal logic across renderers (Raycast, AmbientOcclusion, DirectLight) into a templated rendering loop in gpu/renderer/raygen_helpers.h.

Each renderer now provides a simple ShadingPolicy that implements surface shading, while the common loop handles ray marching, transparency traversal, and volume integration.

Key changes:

• Extract shared rendering loop into template accepting ShadingPolicy
• Standardize ray types across renderers (PRIMARY/SHADOW)
• Separate buffer ID assignment (setPixelIds) from pixel accumulation (accumPixelSample)
• Fix frameID sample count discrepancy by incrementing after render completes
• Simplify DirectLight indirect lighting to single-bounce cosine-weighted diffuse
• Add shared constants (OPACITY_THRESHOLD, MIN_CONTRIBUTION_EPSILON)
• Add opacity pass-through check to material shaders (MDL, PhysicallyBased)

PT, Debug and Test renderers are purposefully excluded from this rework as they don't fit into the exposed paradigm.
A quick assessment shows that performances are on par with our previous implementation.

[Copilot]

Pull request overview

This PR refactors renderer implementations by consolidating duplicated ray traversal logic into a templated rendering loop in raygen_helpers.h. Each renderer now implements a lightweight ShadingPolicy for surface shading while the common loop handles ray marching, transparency traversal, and volume integration.

Changes:

• Extracted shared rendering loop into template accepting ShadingPolicy parameter
• Standardized ray types across renderers (PRIMARY/SHADOW) and added shared constants
• Separated buffer ID assignment (setPixelIds) from pixel accumulation (accumPixelSample)
• Simplified DirectLight indirect lighting to single-bounce cosine-weighted diffuse
• Added opacity pass-through check to material shaders (MDL, PhysicallyBased)

Reviewed changes

Copilot reviewed 18 out of 18 changed files in this pull request and generated 2 comments.

Show a summary per file

File	Description
devices/rtx/device/renderer/raygen_helpers.h	New templated rendering loop with shared ray traversal logic
devices/rtx/device/renderer/common.h	New shared constants and standardized RayType enum
devices/rtx/device/renderer/Raycast_ptx.cu	Refactored to use templated rendering loop with RaycastShadingPolicy
devices/rtx/device/renderer/AmbientOcclusion_ptx.cu	Refactored to use templated rendering loop with AmbientOcclusionShadingPolicy
devices/rtx/device/renderer/DirectLight_ptx.cu	Refactored to use templated rendering loop with DirectLightShadingPolicy, simplified indirect lighting
devices/rtx/device/renderer/Test_ptx.cu	Updated to use separated setPixelIds and accumPixelSample functions
devices/rtx/device/renderer/PathTracer_ptx.cu	Updated to use separated setPixelIds and accumPixelSample functions
devices/rtx/device/renderer/Debug_ptx.cu	Updated to use separated setPixelIds and accumPixelSample functions
devices/rtx/device/renderer/DirectLight.cpp	Updated hitgroup and miss function names
devices/rtx/device/renderer/AmbientOcclusion.cpp	Updated hitgroup function names
devices/rtx/device/gpu/gpu_util.h	Split accumResults into setPixelIds and accumPixelSample, added accumulateNormal
devices/rtx/device/gpu/computeAO.h	Updated to use function pointer for surface attenuation
devices/rtx/device/gpu/intersectRay.h	Removed surfaceAttenuation function (moved to raygen_helpers.h)
devices/rtx/device/gpu/volumeIntegration.h	Adjusted depth recording logic for minimum opacity threshold
devices/rtx/device/material/shaders/PhysicallyBasedShader_ptx.cu	Added opacity pass-through check to nextRay function
devices/rtx/device/material/shaders/MDLShader_ptx.cu	Added opacity pass-through check to nextRay function
devices/rtx/device/mdl/MaterialRegistry.cpp	Added validation for PTX target code generation
devices/rtx/device/frame/Frame.cu	Moved frameID increment to after rendering completes, changed normal buffer to use normalize

Comments suppressed due to low confidence (2)

devices/rtx/device/renderer/DirectLight_ptx.cu:1

• Inconsistent indentation in ternary operator. The false branch should align with the true branch for better readability.

/*

devices/rtx/device/renderer/DirectLight_ptx.cu:1

• Inconsistent indentation in ternary operator. The false branch should align with the true branch for better readability.

/*

---

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

[jeffamstutz]

LGTM, thanks!