Cycles: bake transparent shadows for hair

These transparent shadows can be expansive to evaluate. Especially on the
GPU they can lead to poor occupancy when only some pixels require many kernel
launches to trace and evaluate many layers of transparency.

Baked transparency allows tracing a single ray in many cases by accumulating
the throughput directly in the intersection program without recording hits
or evaluating shaders. Transparency is baked at curve vertices and
interpolated, for most shaders this will look practically the same as actual
shader evaluation.

Fixes T91428, performance regression with spring demo file due to transparent
hair, and makes it render significantly faster than Blender 2.93.

Differential Revision: https://developer.blender.org/D12880

2 files changed, 18 insertions, 0 deletions

diff --git a/intern/cycles/kernel/device/cpu/kernel_arch.h b/intern/cycles/kernel/device/cpu/kernel_arch.h
index ae7fab65100..7a438b58e73 100644
--- a/intern/cycles/kernel/device/cpu/kernel_arch.h
+++ b/intern/cycles/kernel/device/cpu/kernel_arch.h
@@ -64,6 +64,11 @@ void KERNEL_FUNCTION_FULL_NAME(shader_eval_displace)(const KernelGlobalsCPU *kg,
                                                      const KernelShaderEvalInput *input,
                                                      float *output,
                                                      const int offset);
+void KERNEL_FUNCTION_FULL_NAME(shader_eval_curve_shadow_transparency)(
+    const KernelGlobalsCPU *kg,
+    const KernelShaderEvalInput *input,
+    float *output,
+    const int offset);
 
 /* --------------------------------------------------------------------
  * Adaptive sampling.
diff --git a/intern/cycles/kernel/device/cpu/kernel_arch_impl.h b/intern/cycles/kernel/device/cpu/kernel_arch_impl.h
index 2b0eea4fb61..ac606c768db 100644
--- a/intern/cycles/kernel/device/cpu/kernel_arch_impl.h
+++ b/intern/cycles/kernel/device/cpu/kernel_arch_impl.h
@@ -150,6 +150,19 @@ void KERNEL_FUNCTION_FULL_NAME(shader_eval_background)(const KernelGlobalsCPU *k
 #endif
 }
 
+void KERNEL_FUNCTION_FULL_NAME(shader_eval_curve_shadow_transparency)(
+    const KernelGlobalsCPU *kg,
+    const KernelShaderEvalInput *input,
+    float *output,
+    const int offset)
+{
+#ifdef KERNEL_STUB
+  STUB_ASSERT(KERNEL_ARCH, shader_eval_curve_shadow_transparency);
+#else
+  kernel_curve_shadow_transparency_evaluate(kg, input, output, offset);
+#endif
+}
+
 /* --------------------------------------------------------------------
  * Adaptive sampling.
  */