Cycles: Branched path tracing for the split kernel

This implements branched path tracing for the split kernel.

General approach is to store the ray state at a branch point, trace the
branched ray as normal, then restore the state as necessary before iterating
to the next part of the path. A state machine is used to advance the indirect
loop state, which avoids the need to add any new kernels. Each iteration the
state machine recreates as much state as possible from the stored ray to keep
overall storage down.

Its kind of hard to keep all the different integration loops in sync, so this
needs lots of testing to make sure everything is working correctly. We should
probably start trying to deduplicate the integration loops more now.

Nonbranched BMW is ~2% slower, while classroom is ~2% faster, other scenes
could use more testing still.

Reviewers: sergey, nirved

Reviewed By: nirved

Subscribers: Blendify, bliblubli

Differential Revision: https://developer.blender.org/D2611

intern/cycles/kernel/kernel_path.h

@@ -58,7 +58,7 @@ ccl_device_noinline void kernel_path_ao(KernelGlobals *kg,
                                         ShaderData *sd,
                                         ShaderData *emission_sd,
                                         PathRadiance *L,
-                                        PathState *state,
+                                        ccl_addr_space PathState *state,
                                         RNG *rng,
                                         float3 throughput,
                                         float3 ao_alpha)
@@ -98,6 +98,8 @@ ccl_device_noinline void kernel_path_ao(KernelGlobals *kg,
 	}
 }
 
+#ifndef __SPLIT_KERNEL__
+
 ccl_device void kernel_path_indirect(KernelGlobals *kg,
                                      ShaderData *sd,
                                      ShaderData *emission_sd,
@@ -818,5 +820,7 @@ ccl_device void kernel_path_trace(KernelGlobals *kg,
 	path_rng_end(kg, rng_state, rng);
 }
 
+#endif  /* __SPLIT_KERNEL__ */
+
 CCL_NAMESPACE_END