Cycles: simplify handling of ray differentials

* Store compact ray differentials in ShaderData and compute full differentials
  on demand. This reduces register pressure on the GPU.
* Remove BSDF differential code that was effectively doing nothing as the
  differential orientation was discarded when making it compact.

This gives a 1-5% speedup with RTX A6000 + OptiX in our benchmarks, with the
bigger speedups in simpler scenes.

Renders appear to be identical except for the Both displacement option that
does both displacement and bump.

Differential Revision: https://developer.blender.org/D15677

1 files changed, 0 insertions, 8 deletions

diff --git a/intern/cycles/kernel/closure/bsdf_transparent.h b/intern/cycles/kernel/closure/bsdf_transparent.h
index 89b36c709e4..c2aee1e1633 100644
--- a/intern/cycles/kernel/closure/bsdf_transparent.h
+++ b/intern/cycles/kernel/closure/bsdf_transparent.h
@@ -80,22 +80,14 @@ ccl_device Spectrum bsdf_transparent_eval_transmit(ccl_private const ShaderClosu
 ccl_device int bsdf_transparent_sample(ccl_private const ShaderClosure *sc,
                                        float3 Ng,
                                        float3 I,
-                                       float3 dIdx,
-                                       float3 dIdy,
                                        float randu,
                                        float randv,
                                        ccl_private Spectrum *eval,
                                        ccl_private float3 *omega_in,
-                                       ccl_private float3 *domega_in_dx,
-                                       ccl_private float3 *domega_in_dy,
                                        ccl_private float *pdf)
 {
   // only one direction is possible
   *omega_in = -I;
-#ifdef __RAY_DIFFERENTIALS__
-  *domega_in_dx = -dIdx;
-  *domega_in_dy = -dIdy;
-#endif
   *pdf = 1;
   *eval = one_spectrum();
   return LABEL_TRANSMIT | LABEL_TRANSPARENT;