Cr8-xyz/blender
1
0
Fork 0
Code Issues Pull Requests Actions 1 Packages Projects Releases Wiki Activity

Cycles: code refactoring to deduplicate the various BVH traversal variations.

Browse Source 
Now there is a single BVH traversal code with #ifdefs for various features.
At runtime it will then select the appropriate variation to use depending if
instancing, hair or motion blur is in use.

This makes scenes without hair render a bit faster, especially after the
minimum width feature was added. It's not the most beautiful code, but we can't
use c++ templates and there were already 4 copies, adding 4 more to handle the
hair case separately would be too much.
This commit is contained in:
Brecht Van Lommel
2013-04-17 20:07:22 +00:00
parent f3f5e9553e
commit ed1a08382f
8 changed files with 421 additions and 569 deletions
Download Patch File Download Diff File Expand all files Collapse all files

44
intern/cycles/kernel/kernel_path.h
View File

@@ -251,14 +251,19 @@ __device float4 kernel_path_progressive(KernelGlobals *kg, RNG *rng, int sample,
uint visibility = path_state_ray_visibility(kg, &state);
#ifdef __HAIR__
float difl = 0.0f;
if((kernel_data.cam.resolution == 1) && (state.flag & PATH_RAY_CAMERA)) {
float3 pixdiff = ray.dD.dx + ray.dD.dy;
/*pixdiff = pixdiff - dot(pixdiff, ray.D)*ray.D;*/
difl = kernel_data.curve_kernel_data.minimum_width * len(pixdiff) * 0.5f;
float difl = 0.0f, extmax = 0.0f;
uint lcg_state = 0;
if(kernel_data.bvh.have_curves) {
if((kernel_data.cam.resolution == 1) && (state.flag & PATH_RAY_CAMERA)) {
float3 pixdiff = ray.dD.dx + ray.dD.dy;
/*pixdiff = pixdiff - dot(pixdiff, ray.D)*ray.D;*/
difl = kernel_data.curve_kernel_data.minimum_width * len(pixdiff) * 0.5f;
}
extmax = kernel_data.curve_kernel_data.maximum_width;
lcg_state = lcg_init(*rng + rng_offset + sample*0x51633e2d);
}
float extmax = kernel_data.curve_kernel_data.maximum_width;
uint lcg_state = lcg_init(*rng + rng_offset + sample);
bool hit = scene_intersect(kg, &ray, visibility, &isect, &lcg_state, difl, extmax);
#else
@@ -376,7 +381,7 @@ __device float4 kernel_path_progressive(KernelGlobals *kg, RNG *rng, int sample,
/* do bssrdf scatter step if we picked a bssrdf closure */
if(sc) {
uint lcg_state = lcg_init(*rng + rng_offset + sample);
uint lcg_state = lcg_init(*rng + rng_offset + sample*0x68bc21eb);
subsurface_scatter_step(kg, &sd, state.flag, sc, &lcg_state, false);
}
}
@@ -600,7 +605,7 @@ __device void kernel_path_indirect(KernelGlobals *kg, RNG *rng, int sample, Ray
/* do bssrdf scatter step if we picked a bssrdf closure */
if(sc) {
uint lcg_state = lcg_init(*rng + rng_offset + sample);
uint lcg_state = lcg_init(*rng + rng_offset + sample*0x68bc21eb);
subsurface_scatter_step(kg, &sd, state.flag, sc, &lcg_state, false);
}
}
@@ -922,14 +927,19 @@ __device float4 kernel_path_non_progressive(KernelGlobals *kg, RNG *rng, int sam
uint visibility = path_state_ray_visibility(kg, &state);
#ifdef __HAIR__
float difl = 0.0f;
if((kernel_data.cam.resolution == 1) && (state.flag & PATH_RAY_CAMERA)) {
float3 pixdiff = ray.dD.dx + ray.dD.dy;
/*pixdiff = pixdiff - dot(pixdiff, ray.D)*ray.D;*/
difl = kernel_data.curve_kernel_data.minimum_width * len(pixdiff) * 0.5f;
float difl = 0.0f, extmax = 0.0f;
uint lcg_state = 0;
if(kernel_data.bvh.have_curves) {
if((kernel_data.cam.resolution == 1) && (state.flag & PATH_RAY_CAMERA)) {
float3 pixdiff = ray.dD.dx + ray.dD.dy;
/*pixdiff = pixdiff - dot(pixdiff, ray.D)*ray.D;*/
difl = kernel_data.curve_kernel_data.minimum_width * len(pixdiff) * 0.5f;
}
extmax = kernel_data.curve_kernel_data.maximum_width;
lcg_state = lcg_init(*rng + rng_offset + sample*0x51633e2d);
}
float extmax = kernel_data.curve_kernel_data.maximum_width;
uint lcg_state = lcg_init(*rng + rng_offset + sample);
if(!scene_intersect(kg, &ray, visibility, &isect, &lcg_state, difl, extmax)) {
#else
@@ -1015,7 +1025,7 @@ __device float4 kernel_path_non_progressive(KernelGlobals *kg, RNG *rng, int sam
continue;
/* set up random number generator */
uint lcg_state = lcg_init(*rng + rng_offset + sample);
uint lcg_state = lcg_init(*rng + rng_offset + sample*0x68bc21eb);
int num_samples = kernel_data.integrator.subsurface_samples;
float num_samples_inv = 1.0f/num_samples;