bartvdbraak/blender
1
0
Fork 4
Code Issues 4 Pull Requests 1 Actions 1 Packages Projects Releases Wiki Activity
fa46f5a289
blender/intern/cycles/bvh/bvh.cpp

966 lines
25 KiB
C++
Raw Normal View History

Cycles render engine, initial commit. This is the engine itself, blender modifications and build instructions will follow later. Cycles uses code from some great open source projects, many thanks them: * BVH building and traversal code from NVidia's "Understanding the Efficiency of Ray Traversal on GPUs": http://code.google.com/p/understanding-the-efficiency-of-ray-traversal-on-gpus/ * Open Shading Language for a large part of the shading system: http://code.google.com/p/openshadinglanguage/ * Blender for procedural textures and a few other nodes. * Approximate Catmull Clark subdivision from NVidia Mesh tools: http://code.google.com/p/nvidia-mesh-tools/ * Sobol direction vectors from: http://web.maths.unsw.edu.au/~fkuo/sobol/ * Film response functions from: http://www.cs.columbia.edu/CAVE/software/softlib/dorf.php
2011-04-27 11:58:34 +00:00
/*
* Adapted from code copyright 2009-2010 NVIDIA Corporation
* Modifications Copyright 2011, Blender Foundation.
*
* Licensed under the Apache License, Version 2.0 (the "License");
* you may not use this file except in compliance with the License.
* You may obtain a copy of the License at
*
* http://www.apache.org/licenses/LICENSE-2.0
*
* Unless required by applicable law or agreed to in writing, software
* distributed under the License is distributed on an "AS IS" BASIS,
* WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
* See the License for the specific language governing permissions and
* limitations under the License.
*/
#include "mesh.h"
#include "object.h"
#include "scene.h"
Cycles Hair: Introduction of Cardinal Spline Curve Segments and minor fixes. The curve segment primitive has been added. This includes an intersection function and changes to the BVH. A few small errors in the line segment intersection routine are also fixed.
2013-01-15 19:44:41 +00:00
#include "curves.h"
Cycles render engine, initial commit. This is the engine itself, blender modifications and build instructions will follow later. Cycles uses code from some great open source projects, many thanks them: * BVH building and traversal code from NVidia's "Understanding the Efficiency of Ray Traversal on GPUs": http://code.google.com/p/understanding-the-efficiency-of-ray-traversal-on-gpus/ * Open Shading Language for a large part of the shading system: http://code.google.com/p/openshadinglanguage/ * Blender for procedural textures and a few other nodes. * Approximate Catmull Clark subdivision from NVidia Mesh tools: http://code.google.com/p/nvidia-mesh-tools/ * Sobol direction vectors from: http://web.maths.unsw.edu.au/~fkuo/sobol/ * Film response functions from: http://www.cs.columbia.edu/CAVE/software/softlib/dorf.php
2011-04-27 11:58:34 +00:00
#include "bvh.h"
#include "bvh_build.h"
#include "bvh_node.h"
#include "bvh_params.h"
#include "util_cache.h"
#include "util_debug.h"
#include "util_foreach.h"
Cycles: * Fix crash in light path node * Fix struct alignment issue for cuda * Fix issue with instances taking up too much memory * Fix issue with ray visibility working incorrect on some objects * Enable OpenCL always and remove option, it has no dependencies so may as well * Refuse to load kernel if OpenCL version < 1.1, recent drivers are needed * Better error handling for OpenCL device * 3D views with rendered draw mode will now revert to wireframe on file load
2011-09-02 14:55:06 +00:00
#include "util_map.h"
Cycles render engine, initial commit. This is the engine itself, blender modifications and build instructions will follow later. Cycles uses code from some great open source projects, many thanks them: * BVH building and traversal code from NVidia's "Understanding the Efficiency of Ray Traversal on GPUs": http://code.google.com/p/understanding-the-efficiency-of-ray-traversal-on-gpus/ * Open Shading Language for a large part of the shading system: http://code.google.com/p/openshadinglanguage/ * Blender for procedural textures and a few other nodes. * Approximate Catmull Clark subdivision from NVidia Mesh tools: http://code.google.com/p/nvidia-mesh-tools/ * Sobol direction vectors from: http://web.maths.unsw.edu.au/~fkuo/sobol/ * Film response functions from: http://www.cs.columbia.edu/CAVE/software/softlib/dorf.php
2011-04-27 11:58:34 +00:00
#include "util_progress.h"
Fix cycles issue with BVH cache created with 64 bits and used for 32 bits binary, and vice versa.
2013-02-13 11:02:51 +00:00
#include "util_system.h"
Cycles render engine, initial commit. This is the engine itself, blender modifications and build instructions will follow later. Cycles uses code from some great open source projects, many thanks them: * BVH building and traversal code from NVidia's "Understanding the Efficiency of Ray Traversal on GPUs": http://code.google.com/p/understanding-the-efficiency-of-ray-traversal-on-gpus/ * Open Shading Language for a large part of the shading system: http://code.google.com/p/openshadinglanguage/ * Blender for procedural textures and a few other nodes. * Approximate Catmull Clark subdivision from NVidia Mesh tools: http://code.google.com/p/nvidia-mesh-tools/ * Sobol direction vectors from: http://web.maths.unsw.edu.au/~fkuo/sobol/ * Film response functions from: http://www.cs.columbia.edu/CAVE/software/softlib/dorf.php
2011-04-27 11:58:34 +00:00
#include "util_types.h"
Cycles : * Use is_zero(a) rather than dot(a, a) == 0, saves some calculations.
2013-05-14 18:31:55 +00:00
#include "util_math.h"
Cycles render engine, initial commit. This is the engine itself, blender modifications and build instructions will follow later. Cycles uses code from some great open source projects, many thanks them: * BVH building and traversal code from NVidia's "Understanding the Efficiency of Ray Traversal on GPUs": http://code.google.com/p/understanding-the-efficiency-of-ray-traversal-on-gpus/ * Open Shading Language for a large part of the shading system: http://code.google.com/p/openshadinglanguage/ * Blender for procedural textures and a few other nodes. * Approximate Catmull Clark subdivision from NVidia Mesh tools: http://code.google.com/p/nvidia-mesh-tools/ * Sobol direction vectors from: http://web.maths.unsw.edu.au/~fkuo/sobol/ * Film response functions from: http://www.cs.columbia.edu/CAVE/software/softlib/dorf.php
2011-04-27 11:58:34 +00:00
CCL_NAMESPACE_BEGIN
/* Pack Utility */
struct BVHStackEntry
{
const BVHNode *node;
int idx;
BVHStackEntry(const BVHNode* n = 0, int i = 0)
: node(n), idx(i)
{
}
int encodeIdx() const
{
return (node->is_leaf())? ~idx: idx;
}
};
/* BVH */
BVH::BVH(const BVHParams& params_, const vector<Object*>& objects_)
: params(params_), objects(objects_)
{
}
BVH *BVH::create(const BVHParams& params, const vector<Object*>& objects)
{
if(params.use_qbvh)
return new QBVH(params, objects);
else
return new RegularBVH(params, objects);
}
/* Cache */
bool BVH::cache_read(CacheData& key)
{
Fix cycles issue with BVH cache created with 64 bits and used for 32 bits binary, and vice versa.
2013-02-13 11:02:51 +00:00
key.add(system_cpu_bits());
Cycles render engine, initial commit. This is the engine itself, blender modifications and build instructions will follow later. Cycles uses code from some great open source projects, many thanks them: * BVH building and traversal code from NVidia's "Understanding the Efficiency of Ray Traversal on GPUs": http://code.google.com/p/understanding-the-efficiency-of-ray-traversal-on-gpus/ * Open Shading Language for a large part of the shading system: http://code.google.com/p/openshadinglanguage/ * Blender for procedural textures and a few other nodes. * Approximate Catmull Clark subdivision from NVidia Mesh tools: http://code.google.com/p/nvidia-mesh-tools/ * Sobol direction vectors from: http://web.maths.unsw.edu.au/~fkuo/sobol/ * Film response functions from: http://www.cs.columbia.edu/CAVE/software/softlib/dorf.php
2011-04-27 11:58:34 +00:00
key.add(&params, sizeof(params));
foreach(Object *ob, objects) {
Fix T39523: cycles cache BVH not working correct with deformation motion blur.
2014-04-02 15:06:23 +00:00
Mesh *mesh = ob->mesh;
key.add(mesh->verts);
key.add(mesh->triangles);
key.add(mesh->curve_keys);
key.add(mesh->curves);
Cycles: add option to cache BVH's between subsequent renders, storing the BVH on disk to be reused by the next render. This is useful for rendering animations where only the camera or materials change. Note that saving the BVH to disk only to be removed for the next frame is slower if this is not the case and the meshes do actually change. For a render, it will save bvh files to the cache user directory, and remove all cache files from other renders. The files are named using a MD5 hash based on the mesh, to verify if the meshes are still the same.
2012-01-16 13:13:37 +00:00
key.add(&ob->bounds, sizeof(ob->bounds));
key.add(&ob->visibility, sizeof(ob->visibility));
Fix T39523: cycles cache BVH not working correct with deformation motion blur.
2014-04-02 15:06:23 +00:00
key.add(&mesh->transform_applied, sizeof(bool));
if(mesh->use_motion_blur) {
Attribute *attr = mesh->attributes.find(ATTR_STD_MOTION_VERTEX_POSITION);
if(attr)
key.add(attr->buffer);
attr = mesh->curve_attributes.find(ATTR_STD_MOTION_VERTEX_POSITION);
if(attr)
key.add(attr->buffer);
}
Cycles render engine, initial commit. This is the engine itself, blender modifications and build instructions will follow later. Cycles uses code from some great open source projects, many thanks them: * BVH building and traversal code from NVidia's "Understanding the Efficiency of Ray Traversal on GPUs": http://code.google.com/p/understanding-the-efficiency-of-ray-traversal-on-gpus/ * Open Shading Language for a large part of the shading system: http://code.google.com/p/openshadinglanguage/ * Blender for procedural textures and a few other nodes. * Approximate Catmull Clark subdivision from NVidia Mesh tools: http://code.google.com/p/nvidia-mesh-tools/ * Sobol direction vectors from: http://web.maths.unsw.edu.au/~fkuo/sobol/ * Film response functions from: http://www.cs.columbia.edu/CAVE/software/softlib/dorf.php
2011-04-27 11:58:34 +00:00
}
CacheData value;
if(Cache::global.lookup(key, value)) {
Cycles: add option to cache BVH's between subsequent renders, storing the BVH on disk to be reused by the next render. This is useful for rendering animations where only the camera or materials change. Note that saving the BVH to disk only to be removed for the next frame is slower if this is not the case and the meshes do actually change. For a render, it will save bvh files to the cache user directory, and remove all cache files from other renders. The files are named using a MD5 hash based on the mesh, to verify if the meshes are still the same.
2012-01-16 13:13:37 +00:00
cache_filename = key.get_filename();
Cycles: Rebuild BVH from scratch if loading cache failed Before this Cycles used to try using the cache even so it knew for the fact that reading it from the disk failed. This change doesn't make it more stable if someone will try to trick Cycles and give malformed data but it solves general cases when Blender crashed during the cache write and will preserve rendering from crashing when trying to use that partial cache.
2014-09-01 12:03:03 +00:00
if(!(value.read(pack.root_index) &&
value.read(pack.SAH) &&
value.read(pack.nodes) &&
value.read(pack.object_node) &&
value.read(pack.tri_woop) &&
value.read(pack.prim_type) &&
value.read(pack.prim_visibility) &&
value.read(pack.prim_index) &&
value.read(pack.prim_object) &&
value.read(pack.is_leaf)))
{
/* Clear the pack if load failed. */
pack.root_index = 0;
pack.SAH = 0.0f;
pack.nodes.clear();
pack.object_node.clear();
pack.tri_woop.clear();
pack.prim_type.clear();
pack.prim_visibility.clear();
pack.prim_index.clear();
pack.prim_object.clear();
pack.is_leaf.clear();
return false;
}
Cycles render engine, initial commit. This is the engine itself, blender modifications and build instructions will follow later. Cycles uses code from some great open source projects, many thanks them: * BVH building and traversal code from NVidia's "Understanding the Efficiency of Ray Traversal on GPUs": http://code.google.com/p/understanding-the-efficiency-of-ray-traversal-on-gpus/ * Open Shading Language for a large part of the shading system: http://code.google.com/p/openshadinglanguage/ * Blender for procedural textures and a few other nodes. * Approximate Catmull Clark subdivision from NVidia Mesh tools: http://code.google.com/p/nvidia-mesh-tools/ * Sobol direction vectors from: http://web.maths.unsw.edu.au/~fkuo/sobol/ * Film response functions from: http://www.cs.columbia.edu/CAVE/software/softlib/dorf.php
2011-04-27 11:58:34 +00:00
return true;
}
return false;
}
void BVH::cache_write(CacheData& key)
{
CacheData value;
value.add(pack.root_index);
Cycles: add option to cache BVH's between subsequent renders, storing the BVH on disk to be reused by the next render. This is useful for rendering animations where only the camera or materials change. Note that saving the BVH to disk only to be removed for the next frame is slower if this is not the case and the meshes do actually change. For a render, it will save bvh files to the cache user directory, and remove all cache files from other renders. The files are named using a MD5 hash based on the mesh, to verify if the meshes are still the same.
2012-01-16 13:13:37 +00:00
value.add(pack.SAH);
Cycles render engine, initial commit. This is the engine itself, blender modifications and build instructions will follow later. Cycles uses code from some great open source projects, many thanks them: * BVH building and traversal code from NVidia's "Understanding the Efficiency of Ray Traversal on GPUs": http://code.google.com/p/understanding-the-efficiency-of-ray-traversal-on-gpus/ * Open Shading Language for a large part of the shading system: http://code.google.com/p/openshadinglanguage/ * Blender for procedural textures and a few other nodes. * Approximate Catmull Clark subdivision from NVidia Mesh tools: http://code.google.com/p/nvidia-mesh-tools/ * Sobol direction vectors from: http://web.maths.unsw.edu.au/~fkuo/sobol/ * Film response functions from: http://www.cs.columbia.edu/CAVE/software/softlib/dorf.php
2011-04-27 11:58:34 +00:00
value.add(pack.nodes);
value.add(pack.object_node);
value.add(pack.tri_woop);
Cycles code refactor: changes to make adding new primitive types easier.
2014-03-29 12:03:46 +00:00
value.add(pack.prim_type);
Cycles: * Add max diffuse/glossy/transmission bounces * Add separate min/max for transparent depth * Updated/added some presets that use these options * Add ray visibility options for objects, to hide them from camera/diffuse/glossy/transmission/shadow rays * Is singular ray output for light path node Details here: http://wiki.blender.org/index.php/Dev:2.5/Source/Render/Cycles/LightPaths
2011-09-01 15:53:36 +00:00
value.add(pack.prim_visibility);
Cycles render engine, initial commit. This is the engine itself, blender modifications and build instructions will follow later. Cycles uses code from some great open source projects, many thanks them: * BVH building and traversal code from NVidia's "Understanding the Efficiency of Ray Traversal on GPUs": http://code.google.com/p/understanding-the-efficiency-of-ray-traversal-on-gpus/ * Open Shading Language for a large part of the shading system: http://code.google.com/p/openshadinglanguage/ * Blender for procedural textures and a few other nodes. * Approximate Catmull Clark subdivision from NVidia Mesh tools: http://code.google.com/p/nvidia-mesh-tools/ * Sobol direction vectors from: http://web.maths.unsw.edu.au/~fkuo/sobol/ * Film response functions from: http://www.cs.columbia.edu/CAVE/software/softlib/dorf.php
2011-04-27 11:58:34 +00:00
value.add(pack.prim_index);
value.add(pack.prim_object);
value.add(pack.is_leaf);
Cache::global.insert(key, value);
Cycles: add option to cache BVH's between subsequent renders, storing the BVH on disk to be reused by the next render. This is useful for rendering animations where only the camera or materials change. Note that saving the BVH to disk only to be removed for the next frame is slower if this is not the case and the meshes do actually change. For a render, it will save bvh files to the cache user directory, and remove all cache files from other renders. The files are named using a MD5 hash based on the mesh, to verify if the meshes are still the same.
2012-01-16 13:13:37 +00:00
cache_filename = key.get_filename();
}
void BVH::clear_cache_except()
{
set<string> except;
if(!cache_filename.empty())
except.insert(cache_filename);
foreach(Object *ob, objects) {
Mesh *mesh = ob->mesh;
BVH *bvh = mesh->bvh;
if(bvh && !bvh->cache_filename.empty())
except.insert(bvh->cache_filename);
}
Cache::global.clear_except("bvh", except);
Cycles render engine, initial commit. This is the engine itself, blender modifications and build instructions will follow later. Cycles uses code from some great open source projects, many thanks them: * BVH building and traversal code from NVidia's "Understanding the Efficiency of Ray Traversal on GPUs": http://code.google.com/p/understanding-the-efficiency-of-ray-traversal-on-gpus/ * Open Shading Language for a large part of the shading system: http://code.google.com/p/openshadinglanguage/ * Blender for procedural textures and a few other nodes. * Approximate Catmull Clark subdivision from NVidia Mesh tools: http://code.google.com/p/nvidia-mesh-tools/ * Sobol direction vectors from: http://web.maths.unsw.edu.au/~fkuo/sobol/ * Film response functions from: http://www.cs.columbia.edu/CAVE/software/softlib/dorf.php
2011-04-27 11:58:34 +00:00
}
/* Building */
void BVH::build(Progress& progress)
{
progress.set_substatus("Building BVH");
/* cache read */
CacheData key("bvh");
if(params.use_cache) {
progress.set_substatus("Looking in BVH cache");
if(cache_read(key))
return;
}
/* build nodes */
Cycles code refactor: changes to make adding new primitive types easier.
2014-03-29 12:03:46 +00:00
vector<int> prim_type;
Cycles render engine, initial commit. This is the engine itself, blender modifications and build instructions will follow later. Cycles uses code from some great open source projects, many thanks them: * BVH building and traversal code from NVidia's "Understanding the Efficiency of Ray Traversal on GPUs": http://code.google.com/p/understanding-the-efficiency-of-ray-traversal-on-gpus/ * Open Shading Language for a large part of the shading system: http://code.google.com/p/openshadinglanguage/ * Blender for procedural textures and a few other nodes. * Approximate Catmull Clark subdivision from NVidia Mesh tools: http://code.google.com/p/nvidia-mesh-tools/ * Sobol direction vectors from: http://web.maths.unsw.edu.au/~fkuo/sobol/ * Film response functions from: http://www.cs.columbia.edu/CAVE/software/softlib/dorf.php
2011-04-27 11:58:34 +00:00
vector<int> prim_index;
vector<int> prim_object;
Cycles code refactor: changes to make adding new primitive types easier.
2014-03-29 12:03:46 +00:00
BVHBuild bvh_build(objects, prim_type, prim_index, prim_object, params, progress);
Cycles render engine, initial commit. This is the engine itself, blender modifications and build instructions will follow later. Cycles uses code from some great open source projects, many thanks them: * BVH building and traversal code from NVidia's "Understanding the Efficiency of Ray Traversal on GPUs": http://code.google.com/p/understanding-the-efficiency-of-ray-traversal-on-gpus/ * Open Shading Language for a large part of the shading system: http://code.google.com/p/openshadinglanguage/ * Blender for procedural textures and a few other nodes. * Approximate Catmull Clark subdivision from NVidia Mesh tools: http://code.google.com/p/nvidia-mesh-tools/ * Sobol direction vectors from: http://web.maths.unsw.edu.au/~fkuo/sobol/ * Film response functions from: http://www.cs.columbia.edu/CAVE/software/softlib/dorf.php
2011-04-27 11:58:34 +00:00
BVHNode *root = bvh_build.run();
if(progress.get_cancel()) {
if(root) root->deleteSubtree();
return;
}
/* todo: get rid of this copy */
Cycles code refactor: changes to make adding new primitive types easier.
2014-03-29 12:03:46 +00:00
pack.prim_type = prim_type;
Cycles render engine, initial commit. This is the engine itself, blender modifications and build instructions will follow later. Cycles uses code from some great open source projects, many thanks them: * BVH building and traversal code from NVidia's "Understanding the Efficiency of Ray Traversal on GPUs": http://code.google.com/p/understanding-the-efficiency-of-ray-traversal-on-gpus/ * Open Shading Language for a large part of the shading system: http://code.google.com/p/openshadinglanguage/ * Blender for procedural textures and a few other nodes. * Approximate Catmull Clark subdivision from NVidia Mesh tools: http://code.google.com/p/nvidia-mesh-tools/ * Sobol direction vectors from: http://web.maths.unsw.edu.au/~fkuo/sobol/ * Film response functions from: http://www.cs.columbia.edu/CAVE/software/softlib/dorf.php
2011-04-27 11:58:34 +00:00
pack.prim_index = prim_index;
pack.prim_object = prim_object;
/* compute SAH */
if(!params.top_level)
pack.SAH = root->computeSubtreeSAHCost(params);
if(progress.get_cancel()) {
root->deleteSubtree();
return;
}
/* pack triangles */
New feature Patch [#33445] - Experimental Cycles Hair Rendering (CPU only) This patch allows hair data to be exported to cycles and introduces a new line segment primitive to render with. The UI appears under the particle tab and there is a new hair info node available. It is only available under the experimental feature set and for cpu rendering.
2012-12-28 14:21:30 +00:00
progress.set_substatus("Packing BVH triangles and strands");
pack_primitives();
Cycles render engine, initial commit. This is the engine itself, blender modifications and build instructions will follow later. Cycles uses code from some great open source projects, many thanks them: * BVH building and traversal code from NVidia's "Understanding the Efficiency of Ray Traversal on GPUs": http://code.google.com/p/understanding-the-efficiency-of-ray-traversal-on-gpus/ * Open Shading Language for a large part of the shading system: http://code.google.com/p/openshadinglanguage/ * Blender for procedural textures and a few other nodes. * Approximate Catmull Clark subdivision from NVidia Mesh tools: http://code.google.com/p/nvidia-mesh-tools/ * Sobol direction vectors from: http://web.maths.unsw.edu.au/~fkuo/sobol/ * Film response functions from: http://www.cs.columbia.edu/CAVE/software/softlib/dorf.php
2011-04-27 11:58:34 +00:00
if(progress.get_cancel()) {
root->deleteSubtree();
return;
}
/* pack nodes */
progress.set_substatus("Packing BVH nodes");
array<int> tmp_prim_object = pack.prim_object;
pack_nodes(tmp_prim_object, root);
/* free build nodes */
root->deleteSubtree();
if(progress.get_cancel()) return;
/* cache write */
if(params.use_cache) {
progress.set_substatus("Writing BVH cache");
cache_write(key);
Cycles: add option to cache BVH's between subsequent renders, storing the BVH on disk to be reused by the next render. This is useful for rendering animations where only the camera or materials change. Note that saving the BVH to disk only to be removed for the next frame is slower if this is not the case and the meshes do actually change. For a render, it will save bvh files to the cache user directory, and remove all cache files from other renders. The files are named using a MD5 hash based on the mesh, to verify if the meshes are still the same.
2012-01-16 13:13:37 +00:00
/* clear other bvh files from cache */
if(params.top_level)
clear_cache_except();
Cycles render engine, initial commit. This is the engine itself, blender modifications and build instructions will follow later. Cycles uses code from some great open source projects, many thanks them: * BVH building and traversal code from NVidia's "Understanding the Efficiency of Ray Traversal on GPUs": http://code.google.com/p/understanding-the-efficiency-of-ray-traversal-on-gpus/ * Open Shading Language for a large part of the shading system: http://code.google.com/p/openshadinglanguage/ * Blender for procedural textures and a few other nodes. * Approximate Catmull Clark subdivision from NVidia Mesh tools: http://code.google.com/p/nvidia-mesh-tools/ * Sobol direction vectors from: http://web.maths.unsw.edu.au/~fkuo/sobol/ * Film response functions from: http://www.cs.columbia.edu/CAVE/software/softlib/dorf.php
2011-04-27 11:58:34 +00:00
}
}
/* Refitting */
void BVH::refit(Progress& progress)
{
New feature Patch [#33445] - Experimental Cycles Hair Rendering (CPU only) This patch allows hair data to be exported to cycles and introduces a new line segment primitive to render with. The UI appears under the particle tab and there is a new hair info node available. It is only available under the experimental feature set and for cpu rendering.
2012-12-28 14:21:30 +00:00
progress.set_substatus("Packing BVH primitives");
pack_primitives();
Cycles render engine, initial commit. This is the engine itself, blender modifications and build instructions will follow later. Cycles uses code from some great open source projects, many thanks them: * BVH building and traversal code from NVidia's "Understanding the Efficiency of Ray Traversal on GPUs": http://code.google.com/p/understanding-the-efficiency-of-ray-traversal-on-gpus/ * Open Shading Language for a large part of the shading system: http://code.google.com/p/openshadinglanguage/ * Blender for procedural textures and a few other nodes. * Approximate Catmull Clark subdivision from NVidia Mesh tools: http://code.google.com/p/nvidia-mesh-tools/ * Sobol direction vectors from: http://web.maths.unsw.edu.au/~fkuo/sobol/ * Film response functions from: http://www.cs.columbia.edu/CAVE/software/softlib/dorf.php
2011-04-27 11:58:34 +00:00
if(progress.get_cancel()) return;
progress.set_substatus("Refitting BVH nodes");
refit_nodes();
}
/* Triangles */
void BVH::pack_triangle(int idx, float4 woop[3])
{
int tob = pack.prim_object[idx];
Cycles: Add assert check to triangle packing Handy for troubleshooting.
2015-01-22 09:27:13 +00:00
assert(tob >= 0 && tob < objects.size());
Cycles render engine, initial commit. This is the engine itself, blender modifications and build instructions will follow later. Cycles uses code from some great open source projects, many thanks them: * BVH building and traversal code from NVidia's "Understanding the Efficiency of Ray Traversal on GPUs": http://code.google.com/p/understanding-the-efficiency-of-ray-traversal-on-gpus/ * Open Shading Language for a large part of the shading system: http://code.google.com/p/openshadinglanguage/ * Blender for procedural textures and a few other nodes. * Approximate Catmull Clark subdivision from NVidia Mesh tools: http://code.google.com/p/nvidia-mesh-tools/ * Sobol direction vectors from: http://web.maths.unsw.edu.au/~fkuo/sobol/ * Film response functions from: http://www.cs.columbia.edu/CAVE/software/softlib/dorf.php
2011-04-27 11:58:34 +00:00
const Mesh *mesh = objects[tob]->mesh;
Cycles: add support for mesh deformation motion blur.
2014-03-29 12:03:47 +00:00
if(mesh->has_motion_blur())
return;
Cycles render engine, initial commit. This is the engine itself, blender modifications and build instructions will follow later. Cycles uses code from some great open source projects, many thanks them: * BVH building and traversal code from NVidia's "Understanding the Efficiency of Ray Traversal on GPUs": http://code.google.com/p/understanding-the-efficiency-of-ray-traversal-on-gpus/ * Open Shading Language for a large part of the shading system: http://code.google.com/p/openshadinglanguage/ * Blender for procedural textures and a few other nodes. * Approximate Catmull Clark subdivision from NVidia Mesh tools: http://code.google.com/p/nvidia-mesh-tools/ * Sobol direction vectors from: http://web.maths.unsw.edu.au/~fkuo/sobol/ * Film response functions from: http://www.cs.columbia.edu/CAVE/software/softlib/dorf.php
2011-04-27 11:58:34 +00:00
int tidx = pack.prim_index[idx];
const int *vidx = mesh->triangles[tidx].v;
const float3* vpos = &mesh->verts[0];
float3 v0 = vpos[vidx[0]];
float3 v1 = vpos[vidx[1]];
float3 v2 = vpos[vidx[2]];
Cycles: Implement watertight ray/triangle intersection Using this paper: Sven Woop, Watertight Ray/Triangle Intersection http://jcgt.org/published/0002/01/05/paper.pdf This change is expected to address quite reasonable amount of reports from the bug tracker, plus it might help reducing the noise in some scenes. Unfortunately, it's currently about 7% slower than the previous solution with pre-computed triangle plane equations, but maybe with some smart tweaks to the code (tests reshuffle, using SIMD in a nice way or so) we can avoid the speed regression. But perhaps smartest thing to do here would be to change single triangle / ray intersection with multiple triangles / ray intersections. That's how Embree does this and it's watertight single ray intersection is not any faster that this. Currently only triangle intersection is modified accordingly to the paper, in the future we would also want to modify the node / ray intersection. Reviewers: brecht, juicyfruit Subscribers: dingto, ton Differential Revision: https://developer.blender.org/D819
2014-12-15 16:18:01 +00:00
woop[0] = float3_to_float4(v0);
woop[1] = float3_to_float4(v1);
woop[2] = float3_to_float4(v2);
Cycles render engine, initial commit. This is the engine itself, blender modifications and build instructions will follow later. Cycles uses code from some great open source projects, many thanks them: * BVH building and traversal code from NVidia's "Understanding the Efficiency of Ray Traversal on GPUs": http://code.google.com/p/understanding-the-efficiency-of-ray-traversal-on-gpus/ * Open Shading Language for a large part of the shading system: http://code.google.com/p/openshadinglanguage/ * Blender for procedural textures and a few other nodes. * Approximate Catmull Clark subdivision from NVidia Mesh tools: http://code.google.com/p/nvidia-mesh-tools/ * Sobol direction vectors from: http://web.maths.unsw.edu.au/~fkuo/sobol/ * Film response functions from: http://www.cs.columbia.edu/CAVE/software/softlib/dorf.php
2011-04-27 11:58:34 +00:00
}
New feature Patch [#33445] - Experimental Cycles Hair Rendering (CPU only) This patch allows hair data to be exported to cycles and introduces a new line segment primitive to render with. The UI appears under the particle tab and there is a new hair info node available. It is only available under the experimental feature set and for cpu rendering.
2012-12-28 14:21:30 +00:00
/* Curves*/
void BVH::pack_primitives()
Cycles render engine, initial commit. This is the engine itself, blender modifications and build instructions will follow later. Cycles uses code from some great open source projects, many thanks them: * BVH building and traversal code from NVidia's "Understanding the Efficiency of Ray Traversal on GPUs": http://code.google.com/p/understanding-the-efficiency-of-ray-traversal-on-gpus/ * Open Shading Language for a large part of the shading system: http://code.google.com/p/openshadinglanguage/ * Blender for procedural textures and a few other nodes. * Approximate Catmull Clark subdivision from NVidia Mesh tools: http://code.google.com/p/nvidia-mesh-tools/ * Sobol direction vectors from: http://web.maths.unsw.edu.au/~fkuo/sobol/ * Film response functions from: http://www.cs.columbia.edu/CAVE/software/softlib/dorf.php
2011-04-27 11:58:34 +00:00
{
int nsize = TRI_NODE_SIZE;
size_t tidx_size = pack.prim_index.size();
pack.tri_woop.clear();
pack.tri_woop.resize(tidx_size * nsize);
Cycles: * Add max diffuse/glossy/transmission bounces * Add separate min/max for transparent depth * Updated/added some presets that use these options * Add ray visibility options for objects, to hide them from camera/diffuse/glossy/transmission/shadow rays * Is singular ray output for light path node Details here: http://wiki.blender.org/index.php/Dev:2.5/Source/Render/Cycles/LightPaths
2011-09-01 15:53:36 +00:00
pack.prim_visibility.clear();
pack.prim_visibility.resize(tidx_size);
Cycles render engine, initial commit. This is the engine itself, blender modifications and build instructions will follow later. Cycles uses code from some great open source projects, many thanks them: * BVH building and traversal code from NVidia's "Understanding the Efficiency of Ray Traversal on GPUs": http://code.google.com/p/understanding-the-efficiency-of-ray-traversal-on-gpus/ * Open Shading Language for a large part of the shading system: http://code.google.com/p/openshadinglanguage/ * Blender for procedural textures and a few other nodes. * Approximate Catmull Clark subdivision from NVidia Mesh tools: http://code.google.com/p/nvidia-mesh-tools/ * Sobol direction vectors from: http://web.maths.unsw.edu.au/~fkuo/sobol/ * Film response functions from: http://www.cs.columbia.edu/CAVE/software/softlib/dorf.php
2011-04-27 11:58:34 +00:00
for(unsigned int i = 0; i < tidx_size; i++) {
if(pack.prim_index[i] != -1) {
float4 woop[3];
Cycles: Implement watertight ray/triangle intersection Using this paper: Sven Woop, Watertight Ray/Triangle Intersection http://jcgt.org/published/0002/01/05/paper.pdf This change is expected to address quite reasonable amount of reports from the bug tracker, plus it might help reducing the noise in some scenes. Unfortunately, it's currently about 7% slower than the previous solution with pre-computed triangle plane equations, but maybe with some smart tweaks to the code (tests reshuffle, using SIMD in a nice way or so) we can avoid the speed regression. But perhaps smartest thing to do here would be to change single triangle / ray intersection with multiple triangles / ray intersections. That's how Embree does this and it's watertight single ray intersection is not any faster that this. Currently only triangle intersection is modified accordingly to the paper, in the future we would also want to modify the node / ray intersection. Reviewers: brecht, juicyfruit Subscribers: dingto, ton Differential Revision: https://developer.blender.org/D819
2014-12-15 16:18:01 +00:00
if(pack.prim_type[i] & PRIMITIVE_ALL_TRIANGLE)
New feature Patch [#33445] - Experimental Cycles Hair Rendering (CPU only) This patch allows hair data to be exported to cycles and introduces a new line segment primitive to render with. The UI appears under the particle tab and there is a new hair info node available. It is only available under the experimental feature set and for cpu rendering.
2012-12-28 14:21:30 +00:00
pack_triangle(i, woop);
Cycles render engine, initial commit. This is the engine itself, blender modifications and build instructions will follow later. Cycles uses code from some great open source projects, many thanks them: * BVH building and traversal code from NVidia's "Understanding the Efficiency of Ray Traversal on GPUs": http://code.google.com/p/understanding-the-efficiency-of-ray-traversal-on-gpus/ * Open Shading Language for a large part of the shading system: http://code.google.com/p/openshadinglanguage/ * Blender for procedural textures and a few other nodes. * Approximate Catmull Clark subdivision from NVidia Mesh tools: http://code.google.com/p/nvidia-mesh-tools/ * Sobol direction vectors from: http://web.maths.unsw.edu.au/~fkuo/sobol/ * Film response functions from: http://www.cs.columbia.edu/CAVE/software/softlib/dorf.php
2011-04-27 11:58:34 +00:00
memcpy(&pack.tri_woop[i * nsize], woop, sizeof(float4)*3);
Cycles: * Add max diffuse/glossy/transmission bounces * Add separate min/max for transparent depth * Updated/added some presets that use these options * Add ray visibility options for objects, to hide them from camera/diffuse/glossy/transmission/shadow rays * Is singular ray output for light path node Details here: http://wiki.blender.org/index.php/Dev:2.5/Source/Render/Cycles/LightPaths
2011-09-01 15:53:36 +00:00
int tob = pack.prim_object[i];
Object *ob = objects[tob];
pack.prim_visibility[i] = ob->visibility;
Cycles Hair: Strand Minimum Pixel Size Code is added to restrict the pixel size of strands in cycles. It works best with ribbon primitives and a preset for these is included. It uses distance dependent expansion of the strands and then stochastic strand removal to give a fading. To prevent a slowdown for triangle mesh objects in the BVH an extra visibility flag has been added. It is also only applied for camera rays. The strand width settings are also changed, so that the particle size is not included in the width calculation. Instead there is a separate particle system parameter for width scaling.
2013-04-15 21:38:31 +00:00
Cycles code refactor: changes to make adding new primitive types easier.
2014-03-29 12:03:46 +00:00
if(pack.prim_type[i] & PRIMITIVE_ALL_CURVE)
Cycles Hair: Strand Minimum Pixel Size Code is added to restrict the pixel size of strands in cycles. It works best with ribbon primitives and a preset for these is included. It uses distance dependent expansion of the strands and then stochastic strand removal to give a fading. To prevent a slowdown for triangle mesh objects in the BVH an extra visibility flag has been added. It is also only applied for camera rays. The strand width settings are also changed, so that the particle size is not included in the width calculation. Instead there is a separate particle system parameter for width scaling.
2013-04-15 21:38:31 +00:00
pack.prim_visibility[i] |= PATH_RAY_CURVE;
Cycles render engine, initial commit. This is the engine itself, blender modifications and build instructions will follow later. Cycles uses code from some great open source projects, many thanks them: * BVH building and traversal code from NVidia's "Understanding the Efficiency of Ray Traversal on GPUs": http://code.google.com/p/understanding-the-efficiency-of-ray-traversal-on-gpus/ * Open Shading Language for a large part of the shading system: http://code.google.com/p/openshadinglanguage/ * Blender for procedural textures and a few other nodes. * Approximate Catmull Clark subdivision from NVidia Mesh tools: http://code.google.com/p/nvidia-mesh-tools/ * Sobol direction vectors from: http://web.maths.unsw.edu.au/~fkuo/sobol/ * Film response functions from: http://www.cs.columbia.edu/CAVE/software/softlib/dorf.php
2011-04-27 11:58:34 +00:00
}
Fix some (quite harmless) use of uninitialized memory reported by valgrind.
2013-02-12 13:48:04 +00:00
else {
memset(&pack.tri_woop[i * nsize], 0, sizeof(float4)*3);
pack.prim_visibility[i] = 0;
}
Cycles render engine, initial commit. This is the engine itself, blender modifications and build instructions will follow later. Cycles uses code from some great open source projects, many thanks them: * BVH building and traversal code from NVidia's "Understanding the Efficiency of Ray Traversal on GPUs": http://code.google.com/p/understanding-the-efficiency-of-ray-traversal-on-gpus/ * Open Shading Language for a large part of the shading system: http://code.google.com/p/openshadinglanguage/ * Blender for procedural textures and a few other nodes. * Approximate Catmull Clark subdivision from NVidia Mesh tools: http://code.google.com/p/nvidia-mesh-tools/ * Sobol direction vectors from: http://web.maths.unsw.edu.au/~fkuo/sobol/ * Film response functions from: http://www.cs.columbia.edu/CAVE/software/softlib/dorf.php
2011-04-27 11:58:34 +00:00
}
}
/* Pack Instances */
void BVH::pack_instances(size_t nodes_size)
{
/* The BVH's for instances are built separately, but for traversal all
style cleanup: block comments
2012-06-09 17:22:52 +00:00
* BVH's are stored in global arrays. This function merges them into the
* top level BVH, adjusting indexes and offsets where appropriate. */
Cycles render engine, initial commit. This is the engine itself, blender modifications and build instructions will follow later. Cycles uses code from some great open source projects, many thanks them: * BVH building and traversal code from NVidia's "Understanding the Efficiency of Ray Traversal on GPUs": http://code.google.com/p/understanding-the-efficiency-of-ray-traversal-on-gpus/ * Open Shading Language for a large part of the shading system: http://code.google.com/p/openshadinglanguage/ * Blender for procedural textures and a few other nodes. * Approximate Catmull Clark subdivision from NVidia Mesh tools: http://code.google.com/p/nvidia-mesh-tools/ * Sobol direction vectors from: http://web.maths.unsw.edu.au/~fkuo/sobol/ * Film response functions from: http://www.cs.columbia.edu/CAVE/software/softlib/dorf.php
2011-04-27 11:58:34 +00:00
bool use_qbvh = params.use_qbvh;
size_t nsize = (use_qbvh)? BVH_QNODE_SIZE: BVH_NODE_SIZE;
/* adjust primitive index to point to the triangle in the global array, for
style cleanup: block comments
2012-06-09 17:22:52 +00:00
* meshes with transform applied and already in the top level BVH */
Cycles render engine, initial commit. This is the engine itself, blender modifications and build instructions will follow later. Cycles uses code from some great open source projects, many thanks them: * BVH building and traversal code from NVidia's "Understanding the Efficiency of Ray Traversal on GPUs": http://code.google.com/p/understanding-the-efficiency-of-ray-traversal-on-gpus/ * Open Shading Language for a large part of the shading system: http://code.google.com/p/openshadinglanguage/ * Blender for procedural textures and a few other nodes. * Approximate Catmull Clark subdivision from NVidia Mesh tools: http://code.google.com/p/nvidia-mesh-tools/ * Sobol direction vectors from: http://web.maths.unsw.edu.au/~fkuo/sobol/ * Film response functions from: http://www.cs.columbia.edu/CAVE/software/softlib/dorf.php
2011-04-27 11:58:34 +00:00
for(size_t i = 0; i < pack.prim_index.size(); i++)
New feature Patch [#33445] - Experimental Cycles Hair Rendering (CPU only) This patch allows hair data to be exported to cycles and introduces a new line segment primitive to render with. The UI appears under the particle tab and there is a new hair info node available. It is only available under the experimental feature set and for cpu rendering.
2012-12-28 14:21:30 +00:00
if(pack.prim_index[i] != -1) {
Cycles code refactor: changes to make adding new primitive types easier.
2014-03-29 12:03:46 +00:00
if(pack.prim_type[i] & PRIMITIVE_ALL_CURVE)
Cycles Hair: refactoring to store curves with the index of the first key and the number of keys in the curve, rather than curve segments with the indices of two keys. ShaderData.segment now stores the segment number in the curve.
2013-01-03 12:09:09 +00:00
pack.prim_index[i] += objects[pack.prim_object[i]]->mesh->curve_offset;
New feature Patch [#33445] - Experimental Cycles Hair Rendering (CPU only) This patch allows hair data to be exported to cycles and introduces a new line segment primitive to render with. The UI appears under the particle tab and there is a new hair info node available. It is only available under the experimental feature set and for cpu rendering.
2012-12-28 14:21:30 +00:00
else
pack.prim_index[i] += objects[pack.prim_object[i]]->mesh->tri_offset;
}
Cycles render engine, initial commit. This is the engine itself, blender modifications and build instructions will follow later. Cycles uses code from some great open source projects, many thanks them: * BVH building and traversal code from NVidia's "Understanding the Efficiency of Ray Traversal on GPUs": http://code.google.com/p/understanding-the-efficiency-of-ray-traversal-on-gpus/ * Open Shading Language for a large part of the shading system: http://code.google.com/p/openshadinglanguage/ * Blender for procedural textures and a few other nodes. * Approximate Catmull Clark subdivision from NVidia Mesh tools: http://code.google.com/p/nvidia-mesh-tools/ * Sobol direction vectors from: http://web.maths.unsw.edu.au/~fkuo/sobol/ * Film response functions from: http://www.cs.columbia.edu/CAVE/software/softlib/dorf.php
2011-04-27 11:58:34 +00:00
/* track offsets of instanced BVH data in global array */
New feature Patch [#33445] - Experimental Cycles Hair Rendering (CPU only) This patch allows hair data to be exported to cycles and introduces a new line segment primitive to render with. The UI appears under the particle tab and there is a new hair info node available. It is only available under the experimental feature set and for cpu rendering.
2012-12-28 14:21:30 +00:00
size_t prim_offset = pack.prim_index.size();
Cycles render engine, initial commit. This is the engine itself, blender modifications and build instructions will follow later. Cycles uses code from some great open source projects, many thanks them: * BVH building and traversal code from NVidia's "Understanding the Efficiency of Ray Traversal on GPUs": http://code.google.com/p/understanding-the-efficiency-of-ray-traversal-on-gpus/ * Open Shading Language for a large part of the shading system: http://code.google.com/p/openshadinglanguage/ * Blender for procedural textures and a few other nodes. * Approximate Catmull Clark subdivision from NVidia Mesh tools: http://code.google.com/p/nvidia-mesh-tools/ * Sobol direction vectors from: http://web.maths.unsw.edu.au/~fkuo/sobol/ * Film response functions from: http://www.cs.columbia.edu/CAVE/software/softlib/dorf.php
2011-04-27 11:58:34 +00:00
size_t nodes_offset = nodes_size;
/* clear array that gives the node indexes for instanced objects */
pack.object_node.clear();
/* reserve */
size_t prim_index_size = pack.prim_index.size();
size_t tri_woop_size = pack.tri_woop.size();
size_t pack_prim_index_offset = prim_index_size;
size_t pack_tri_woop_offset = tri_woop_size;
size_t pack_nodes_offset = nodes_size;
size_t object_offset = 0;
Cycles: * Fix crash in light path node * Fix struct alignment issue for cuda * Fix issue with instances taking up too much memory * Fix issue with ray visibility working incorrect on some objects * Enable OpenCL always and remove option, it has no dependencies so may as well * Refuse to load kernel if OpenCL version < 1.1, recent drivers are needed * Better error handling for OpenCL device * 3D views with rendered draw mode will now revert to wireframe on file load
2011-09-02 14:55:06 +00:00
map<Mesh*, int> mesh_map;
Cycles render engine, initial commit. This is the engine itself, blender modifications and build instructions will follow later. Cycles uses code from some great open source projects, many thanks them: * BVH building and traversal code from NVidia's "Understanding the Efficiency of Ray Traversal on GPUs": http://code.google.com/p/understanding-the-efficiency-of-ray-traversal-on-gpus/ * Open Shading Language for a large part of the shading system: http://code.google.com/p/openshadinglanguage/ * Blender for procedural textures and a few other nodes. * Approximate Catmull Clark subdivision from NVidia Mesh tools: http://code.google.com/p/nvidia-mesh-tools/ * Sobol direction vectors from: http://web.maths.unsw.edu.au/~fkuo/sobol/ * Film response functions from: http://www.cs.columbia.edu/CAVE/software/softlib/dorf.php
2011-04-27 11:58:34 +00:00
foreach(Object *ob, objects) {
Mesh *mesh = ob->mesh;
BVH *bvh = mesh->bvh;
if(!mesh->transform_applied) {
Cycles: * Fix crash in light path node * Fix struct alignment issue for cuda * Fix issue with instances taking up too much memory * Fix issue with ray visibility working incorrect on some objects * Enable OpenCL always and remove option, it has no dependencies so may as well * Refuse to load kernel if OpenCL version < 1.1, recent drivers are needed * Better error handling for OpenCL device * 3D views with rendered draw mode will now revert to wireframe on file load
2011-09-02 14:55:06 +00:00
if(mesh_map.find(mesh) == mesh_map.end()) {
prim_index_size += bvh->pack.prim_index.size();
tri_woop_size += bvh->pack.tri_woop.size();
Cycles: Avoid over-allocation in packing BVH instances This solves quite an over-allocation in BVH instances packing code, unfortunately, it's not a magic bullet to solve memory bump caused by the recent QBVH changes. For that we'll likely need to decouple storage for leaf and inner nodes. However, it's not really clear for now if it's something important since that'd still be just a fraction of memory comparing to all the hi-res textures.
2014-12-17 18:31:12 +00:00
nodes_size += bvh->pack.nodes.size();
Cycles: * Fix crash in light path node * Fix struct alignment issue for cuda * Fix issue with instances taking up too much memory * Fix issue with ray visibility working incorrect on some objects * Enable OpenCL always and remove option, it has no dependencies so may as well * Refuse to load kernel if OpenCL version < 1.1, recent drivers are needed * Better error handling for OpenCL device * 3D views with rendered draw mode will now revert to wireframe on file load
2011-09-02 14:55:06 +00:00
mesh_map[mesh] = 1;
}
Cycles render engine, initial commit. This is the engine itself, blender modifications and build instructions will follow later. Cycles uses code from some great open source projects, many thanks them: * BVH building and traversal code from NVidia's "Understanding the Efficiency of Ray Traversal on GPUs": http://code.google.com/p/understanding-the-efficiency-of-ray-traversal-on-gpus/ * Open Shading Language for a large part of the shading system: http://code.google.com/p/openshadinglanguage/ * Blender for procedural textures and a few other nodes. * Approximate Catmull Clark subdivision from NVidia Mesh tools: http://code.google.com/p/nvidia-mesh-tools/ * Sobol direction vectors from: http://web.maths.unsw.edu.au/~fkuo/sobol/ * Film response functions from: http://www.cs.columbia.edu/CAVE/software/softlib/dorf.php
2011-04-27 11:58:34 +00:00
}
}
Cycles: * Fix crash in light path node * Fix struct alignment issue for cuda * Fix issue with instances taking up too much memory * Fix issue with ray visibility working incorrect on some objects * Enable OpenCL always and remove option, it has no dependencies so may as well * Refuse to load kernel if OpenCL version < 1.1, recent drivers are needed * Better error handling for OpenCL device * 3D views with rendered draw mode will now revert to wireframe on file load
2011-09-02 14:55:06 +00:00
mesh_map.clear();
Cycles render engine, initial commit. This is the engine itself, blender modifications and build instructions will follow later. Cycles uses code from some great open source projects, many thanks them: * BVH building and traversal code from NVidia's "Understanding the Efficiency of Ray Traversal on GPUs": http://code.google.com/p/understanding-the-efficiency-of-ray-traversal-on-gpus/ * Open Shading Language for a large part of the shading system: http://code.google.com/p/openshadinglanguage/ * Blender for procedural textures and a few other nodes. * Approximate Catmull Clark subdivision from NVidia Mesh tools: http://code.google.com/p/nvidia-mesh-tools/ * Sobol direction vectors from: http://web.maths.unsw.edu.au/~fkuo/sobol/ * Film response functions from: http://www.cs.columbia.edu/CAVE/software/softlib/dorf.php
2011-04-27 11:58:34 +00:00
pack.prim_index.resize(prim_index_size);
Cycles code refactor: changes to make adding new primitive types easier.
2014-03-29 12:03:46 +00:00
pack.prim_type.resize(prim_index_size);
Cycles render engine, initial commit. This is the engine itself, blender modifications and build instructions will follow later. Cycles uses code from some great open source projects, many thanks them: * BVH building and traversal code from NVidia's "Understanding the Efficiency of Ray Traversal on GPUs": http://code.google.com/p/understanding-the-efficiency-of-ray-traversal-on-gpus/ * Open Shading Language for a large part of the shading system: http://code.google.com/p/openshadinglanguage/ * Blender for procedural textures and a few other nodes. * Approximate Catmull Clark subdivision from NVidia Mesh tools: http://code.google.com/p/nvidia-mesh-tools/ * Sobol direction vectors from: http://web.maths.unsw.edu.au/~fkuo/sobol/ * Film response functions from: http://www.cs.columbia.edu/CAVE/software/softlib/dorf.php
2011-04-27 11:58:34 +00:00
pack.prim_object.resize(prim_index_size);
Cycles: * Add max diffuse/glossy/transmission bounces * Add separate min/max for transparent depth * Updated/added some presets that use these options * Add ray visibility options for objects, to hide them from camera/diffuse/glossy/transmission/shadow rays * Is singular ray output for light path node Details here: http://wiki.blender.org/index.php/Dev:2.5/Source/Render/Cycles/LightPaths
2011-09-01 15:53:36 +00:00
pack.prim_visibility.resize(prim_index_size);
Cycles render engine, initial commit. This is the engine itself, blender modifications and build instructions will follow later. Cycles uses code from some great open source projects, many thanks them: * BVH building and traversal code from NVidia's "Understanding the Efficiency of Ray Traversal on GPUs": http://code.google.com/p/understanding-the-efficiency-of-ray-traversal-on-gpus/ * Open Shading Language for a large part of the shading system: http://code.google.com/p/openshadinglanguage/ * Blender for procedural textures and a few other nodes. * Approximate Catmull Clark subdivision from NVidia Mesh tools: http://code.google.com/p/nvidia-mesh-tools/ * Sobol direction vectors from: http://web.maths.unsw.edu.au/~fkuo/sobol/ * Film response functions from: http://www.cs.columbia.edu/CAVE/software/softlib/dorf.php
2011-04-27 11:58:34 +00:00
pack.tri_woop.resize(tri_woop_size);
pack.nodes.resize(nodes_size);
pack.object_node.resize(objects.size());
Cycles: some tweaks to silence msvc assertions in debug mode.
2011-10-03 15:31:45 +00:00
int *pack_prim_index = (pack.prim_index.size())? &pack.prim_index[0]: NULL;
Cycles code refactor: changes to make adding new primitive types easier.
2014-03-29 12:03:46 +00:00
int *pack_prim_type = (pack.prim_type.size())? &pack.prim_type[0]: NULL;
Cycles: some tweaks to silence msvc assertions in debug mode.
2011-10-03 15:31:45 +00:00
int *pack_prim_object = (pack.prim_object.size())? &pack.prim_object[0]: NULL;
uint *pack_prim_visibility = (pack.prim_visibility.size())? &pack.prim_visibility[0]: NULL;
float4 *pack_tri_woop = (pack.tri_woop.size())? &pack.tri_woop[0]: NULL;
int4 *pack_nodes = (pack.nodes.size())? &pack.nodes[0]: NULL;
Cycles render engine, initial commit. This is the engine itself, blender modifications and build instructions will follow later. Cycles uses code from some great open source projects, many thanks them: * BVH building and traversal code from NVidia's "Understanding the Efficiency of Ray Traversal on GPUs": http://code.google.com/p/understanding-the-efficiency-of-ray-traversal-on-gpus/ * Open Shading Language for a large part of the shading system: http://code.google.com/p/openshadinglanguage/ * Blender for procedural textures and a few other nodes. * Approximate Catmull Clark subdivision from NVidia Mesh tools: http://code.google.com/p/nvidia-mesh-tools/ * Sobol direction vectors from: http://web.maths.unsw.edu.au/~fkuo/sobol/ * Film response functions from: http://www.cs.columbia.edu/CAVE/software/softlib/dorf.php
2011-04-27 11:58:34 +00:00
/* merge */
foreach(Object *ob, objects) {
Mesh *mesh = ob->mesh;
/* if mesh transform is applied, that means it's already in the top
style cleanup: block comments
2012-06-09 17:22:52 +00:00
* level BVH, and we don't need to merge it in */
Cycles render engine, initial commit. This is the engine itself, blender modifications and build instructions will follow later. Cycles uses code from some great open source projects, many thanks them: * BVH building and traversal code from NVidia's "Understanding the Efficiency of Ray Traversal on GPUs": http://code.google.com/p/understanding-the-efficiency-of-ray-traversal-on-gpus/ * Open Shading Language for a large part of the shading system: http://code.google.com/p/openshadinglanguage/ * Blender for procedural textures and a few other nodes. * Approximate Catmull Clark subdivision from NVidia Mesh tools: http://code.google.com/p/nvidia-mesh-tools/ * Sobol direction vectors from: http://web.maths.unsw.edu.au/~fkuo/sobol/ * Film response functions from: http://www.cs.columbia.edu/CAVE/software/softlib/dorf.php
2011-04-27 11:58:34 +00:00
if(mesh->transform_applied) {
pack.object_node[object_offset++] = 0;
continue;
}
Cycles: * Fix crash in light path node * Fix struct alignment issue for cuda * Fix issue with instances taking up too much memory * Fix issue with ray visibility working incorrect on some objects * Enable OpenCL always and remove option, it has no dependencies so may as well * Refuse to load kernel if OpenCL version < 1.1, recent drivers are needed * Better error handling for OpenCL device * 3D views with rendered draw mode will now revert to wireframe on file load
2011-09-02 14:55:06 +00:00
/* if mesh already added once, don't add it again, but used set
style cleanup: block comments
2012-06-09 17:22:52 +00:00
* node offset for this object */
Cycles: * Fix crash in light path node * Fix struct alignment issue for cuda * Fix issue with instances taking up too much memory * Fix issue with ray visibility working incorrect on some objects * Enable OpenCL always and remove option, it has no dependencies so may as well * Refuse to load kernel if OpenCL version < 1.1, recent drivers are needed * Better error handling for OpenCL device * 3D views with rendered draw mode will now revert to wireframe on file load
2011-09-02 14:55:06 +00:00
map<Mesh*, int>::iterator it = mesh_map.find(mesh);
if(mesh_map.find(mesh) != mesh_map.end()) {
int noffset = it->second;
pack.object_node[object_offset++] = noffset;
continue;
}
Cycles render engine, initial commit. This is the engine itself, blender modifications and build instructions will follow later. Cycles uses code from some great open source projects, many thanks them: * BVH building and traversal code from NVidia's "Understanding the Efficiency of Ray Traversal on GPUs": http://code.google.com/p/understanding-the-efficiency-of-ray-traversal-on-gpus/ * Open Shading Language for a large part of the shading system: http://code.google.com/p/openshadinglanguage/ * Blender for procedural textures and a few other nodes. * Approximate Catmull Clark subdivision from NVidia Mesh tools: http://code.google.com/p/nvidia-mesh-tools/ * Sobol direction vectors from: http://web.maths.unsw.edu.au/~fkuo/sobol/ * Film response functions from: http://www.cs.columbia.edu/CAVE/software/softlib/dorf.php
2011-04-27 11:58:34 +00:00
BVH *bvh = mesh->bvh;
int noffset = nodes_offset/nsize;
int mesh_tri_offset = mesh->tri_offset;
Cycles Hair: refactoring to store curves with the index of the first key and the number of keys in the curve, rather than curve segments with the indices of two keys. ShaderData.segment now stores the segment number in the curve.
2013-01-03 12:09:09 +00:00
int mesh_curve_offset = mesh->curve_offset;
Cycles render engine, initial commit. This is the engine itself, blender modifications and build instructions will follow later. Cycles uses code from some great open source projects, many thanks them: * BVH building and traversal code from NVidia's "Understanding the Efficiency of Ray Traversal on GPUs": http://code.google.com/p/understanding-the-efficiency-of-ray-traversal-on-gpus/ * Open Shading Language for a large part of the shading system: http://code.google.com/p/openshadinglanguage/ * Blender for procedural textures and a few other nodes. * Approximate Catmull Clark subdivision from NVidia Mesh tools: http://code.google.com/p/nvidia-mesh-tools/ * Sobol direction vectors from: http://web.maths.unsw.edu.au/~fkuo/sobol/ * Film response functions from: http://www.cs.columbia.edu/CAVE/software/softlib/dorf.php
2011-04-27 11:58:34 +00:00
/* fill in node indexes for instances */
Code tweak removing comment, the fix here is indeed correct.
2012-07-31 15:05:11 +00:00
if((bvh->pack.is_leaf.size() != 0) && bvh->pack.is_leaf[0])
Cycles render engine, initial commit. This is the engine itself, blender modifications and build instructions will follow later. Cycles uses code from some great open source projects, many thanks them: * BVH building and traversal code from NVidia's "Understanding the Efficiency of Ray Traversal on GPUs": http://code.google.com/p/understanding-the-efficiency-of-ray-traversal-on-gpus/ * Open Shading Language for a large part of the shading system: http://code.google.com/p/openshadinglanguage/ * Blender for procedural textures and a few other nodes. * Approximate Catmull Clark subdivision from NVidia Mesh tools: http://code.google.com/p/nvidia-mesh-tools/ * Sobol direction vectors from: http://web.maths.unsw.edu.au/~fkuo/sobol/ * Film response functions from: http://www.cs.columbia.edu/CAVE/software/softlib/dorf.php
2011-04-27 11:58:34 +00:00
pack.object_node[object_offset++] = -noffset-1;
Code tweak removing comment, the fix here is indeed correct.
2012-07-31 15:05:11 +00:00
else
Cycles render engine, initial commit. This is the engine itself, blender modifications and build instructions will follow later. Cycles uses code from some great open source projects, many thanks them: * BVH building and traversal code from NVidia's "Understanding the Efficiency of Ray Traversal on GPUs": http://code.google.com/p/understanding-the-efficiency-of-ray-traversal-on-gpus/ * Open Shading Language for a large part of the shading system: http://code.google.com/p/openshadinglanguage/ * Blender for procedural textures and a few other nodes. * Approximate Catmull Clark subdivision from NVidia Mesh tools: http://code.google.com/p/nvidia-mesh-tools/ * Sobol direction vectors from: http://web.maths.unsw.edu.au/~fkuo/sobol/ * Film response functions from: http://www.cs.columbia.edu/CAVE/software/softlib/dorf.php
2011-04-27 11:58:34 +00:00
pack.object_node[object_offset++] = noffset;
Cycles: * Fix crash in light path node * Fix struct alignment issue for cuda * Fix issue with instances taking up too much memory * Fix issue with ray visibility working incorrect on some objects * Enable OpenCL always and remove option, it has no dependencies so may as well * Refuse to load kernel if OpenCL version < 1.1, recent drivers are needed * Better error handling for OpenCL device * 3D views with rendered draw mode will now revert to wireframe on file load
2011-09-02 14:55:06 +00:00
mesh_map[mesh] = pack.object_node[object_offset-1];
Cycles render engine, initial commit. This is the engine itself, blender modifications and build instructions will follow later. Cycles uses code from some great open source projects, many thanks them: * BVH building and traversal code from NVidia's "Understanding the Efficiency of Ray Traversal on GPUs": http://code.google.com/p/understanding-the-efficiency-of-ray-traversal-on-gpus/ * Open Shading Language for a large part of the shading system: http://code.google.com/p/openshadinglanguage/ * Blender for procedural textures and a few other nodes. * Approximate Catmull Clark subdivision from NVidia Mesh tools: http://code.google.com/p/nvidia-mesh-tools/ * Sobol direction vectors from: http://web.maths.unsw.edu.au/~fkuo/sobol/ * Film response functions from: http://www.cs.columbia.edu/CAVE/software/softlib/dorf.php
2011-04-27 11:58:34 +00:00
/* merge primitive and object indexes */
Cycles: some tweaks to silence msvc assertions in debug mode.
2011-10-03 15:31:45 +00:00
if(bvh->pack.prim_index.size()) {
Cycles render engine, initial commit. This is the engine itself, blender modifications and build instructions will follow later. Cycles uses code from some great open source projects, many thanks them: * BVH building and traversal code from NVidia's "Understanding the Efficiency of Ray Traversal on GPUs": http://code.google.com/p/understanding-the-efficiency-of-ray-traversal-on-gpus/ * Open Shading Language for a large part of the shading system: http://code.google.com/p/openshadinglanguage/ * Blender for procedural textures and a few other nodes. * Approximate Catmull Clark subdivision from NVidia Mesh tools: http://code.google.com/p/nvidia-mesh-tools/ * Sobol direction vectors from: http://web.maths.unsw.edu.au/~fkuo/sobol/ * Film response functions from: http://www.cs.columbia.edu/CAVE/software/softlib/dorf.php
2011-04-27 11:58:34 +00:00
size_t bvh_prim_index_size = bvh->pack.prim_index.size();
int *bvh_prim_index = &bvh->pack.prim_index[0];
Cycles code refactor: changes to make adding new primitive types easier.
2014-03-29 12:03:46 +00:00
int *bvh_prim_type = &bvh->pack.prim_type[0];
Cycles: * Add max diffuse/glossy/transmission bounces * Add separate min/max for transparent depth * Updated/added some presets that use these options * Add ray visibility options for objects, to hide them from camera/diffuse/glossy/transmission/shadow rays * Is singular ray output for light path node Details here: http://wiki.blender.org/index.php/Dev:2.5/Source/Render/Cycles/LightPaths
2011-09-01 15:53:36 +00:00
uint *bvh_prim_visibility = &bvh->pack.prim_visibility[0];
Cycles render engine, initial commit. This is the engine itself, blender modifications and build instructions will follow later. Cycles uses code from some great open source projects, many thanks them: * BVH building and traversal code from NVidia's "Understanding the Efficiency of Ray Traversal on GPUs": http://code.google.com/p/understanding-the-efficiency-of-ray-traversal-on-gpus/ * Open Shading Language for a large part of the shading system: http://code.google.com/p/openshadinglanguage/ * Blender for procedural textures and a few other nodes. * Approximate Catmull Clark subdivision from NVidia Mesh tools: http://code.google.com/p/nvidia-mesh-tools/ * Sobol direction vectors from: http://web.maths.unsw.edu.au/~fkuo/sobol/ * Film response functions from: http://www.cs.columbia.edu/CAVE/software/softlib/dorf.php
2011-04-27 11:58:34 +00:00
for(size_t i = 0; i < bvh_prim_index_size; i++) {
Cycles code refactor: changes to make adding new primitive types easier.
2014-03-29 12:03:46 +00:00
if(bvh->pack.prim_type[i] & PRIMITIVE_ALL_CURVE)
New feature Patch [#33445] - Experimental Cycles Hair Rendering (CPU only) This patch allows hair data to be exported to cycles and introduces a new line segment primitive to render with. The UI appears under the particle tab and there is a new hair info node available. It is only available under the experimental feature set and for cpu rendering.
2012-12-28 14:21:30 +00:00
pack_prim_index[pack_prim_index_offset] = bvh_prim_index[i] + mesh_curve_offset;
else
pack_prim_index[pack_prim_index_offset] = bvh_prim_index[i] + mesh_tri_offset;
Cycles code refactor: changes to make adding new primitive types easier.
2014-03-29 12:03:46 +00:00
pack_prim_type[pack_prim_index_offset] = bvh_prim_type[i];
Cycles: * Fix crash in light path node * Fix struct alignment issue for cuda * Fix issue with instances taking up too much memory * Fix issue with ray visibility working incorrect on some objects * Enable OpenCL always and remove option, it has no dependencies so may as well * Refuse to load kernel if OpenCL version < 1.1, recent drivers are needed * Better error handling for OpenCL device * 3D views with rendered draw mode will now revert to wireframe on file load
2011-09-02 14:55:06 +00:00
pack_prim_visibility[pack_prim_index_offset] = bvh_prim_visibility[i];
Cycles render engine, initial commit. This is the engine itself, blender modifications and build instructions will follow later. Cycles uses code from some great open source projects, many thanks them: * BVH building and traversal code from NVidia's "Understanding the Efficiency of Ray Traversal on GPUs": http://code.google.com/p/understanding-the-efficiency-of-ray-traversal-on-gpus/ * Open Shading Language for a large part of the shading system: http://code.google.com/p/openshadinglanguage/ * Blender for procedural textures and a few other nodes. * Approximate Catmull Clark subdivision from NVidia Mesh tools: http://code.google.com/p/nvidia-mesh-tools/ * Sobol direction vectors from: http://web.maths.unsw.edu.au/~fkuo/sobol/ * Film response functions from: http://www.cs.columbia.edu/CAVE/software/softlib/dorf.php
2011-04-27 11:58:34 +00:00
pack_prim_object[pack_prim_index_offset] = 0; // unused for instances
pack_prim_index_offset++;
}
}
/* merge triangle intersection data */
Cycles: some tweaks to silence msvc assertions in debug mode.
2011-10-03 15:31:45 +00:00
if(bvh->pack.tri_woop.size()) {
New feature Patch [#33445] - Experimental Cycles Hair Rendering (CPU only) This patch allows hair data to be exported to cycles and introduces a new line segment primitive to render with. The UI appears under the particle tab and there is a new hair info node available. It is only available under the experimental feature set and for cpu rendering.
2012-12-28 14:21:30 +00:00
memcpy(pack_tri_woop + pack_tri_woop_offset, &bvh->pack.tri_woop[0],
Cycles render engine, initial commit. This is the engine itself, blender modifications and build instructions will follow later. Cycles uses code from some great open source projects, many thanks them: * BVH building and traversal code from NVidia's "Understanding the Efficiency of Ray Traversal on GPUs": http://code.google.com/p/understanding-the-efficiency-of-ray-traversal-on-gpus/ * Open Shading Language for a large part of the shading system: http://code.google.com/p/openshadinglanguage/ * Blender for procedural textures and a few other nodes. * Approximate Catmull Clark subdivision from NVidia Mesh tools: http://code.google.com/p/nvidia-mesh-tools/ * Sobol direction vectors from: http://web.maths.unsw.edu.au/~fkuo/sobol/ * Film response functions from: http://www.cs.columbia.edu/CAVE/software/softlib/dorf.php
2011-04-27 11:58:34 +00:00
bvh->pack.tri_woop.size()*sizeof(float4));
pack_tri_woop_offset += bvh->pack.tri_woop.size();
}
/* merge nodes */
Fix some (quite harmless) use of uninitialized memory reported by valgrind.
2013-02-12 13:48:04 +00:00
if(bvh->pack.nodes.size()) {
Cleanup: Typo fixes for comments.
2014-12-25 01:42:06 +00:00
/* For QBVH we're packing a child bbox into 6 float4,
* and for regular BVH they're packed into 3 float4.
Cycles: Make it more clear offsets in BVH construction Previously offsets were calculated based on the BVH node size, which is wrong and real PITA in cases when some extra data is to be added into (or removed from) the node. Now use offsets which are not calculated form the node size.
2014-12-21 17:35:51 +00:00
*/
size_t nsize_bbox = (use_qbvh)? 6: 3;
Cycles render engine, initial commit. This is the engine itself, blender modifications and build instructions will follow later. Cycles uses code from some great open source projects, many thanks them: * BVH building and traversal code from NVidia's "Understanding the Efficiency of Ray Traversal on GPUs": http://code.google.com/p/understanding-the-efficiency-of-ray-traversal-on-gpus/ * Open Shading Language for a large part of the shading system: http://code.google.com/p/openshadinglanguage/ * Blender for procedural textures and a few other nodes. * Approximate Catmull Clark subdivision from NVidia Mesh tools: http://code.google.com/p/nvidia-mesh-tools/ * Sobol direction vectors from: http://web.maths.unsw.edu.au/~fkuo/sobol/ * Film response functions from: http://www.cs.columbia.edu/CAVE/software/softlib/dorf.php
2011-04-27 11:58:34 +00:00
int4 *bvh_nodes = &bvh->pack.nodes[0];
size_t bvh_nodes_size = bvh->pack.nodes.size();
fix for rare crash in cycles BVH tree packing.
2012-08-22 11:18:37 +00:00
int *bvh_is_leaf = (bvh->pack.is_leaf.size() != 0) ? &bvh->pack.is_leaf[0] : NULL;
Cycles render engine, initial commit. This is the engine itself, blender modifications and build instructions will follow later. Cycles uses code from some great open source projects, many thanks them: * BVH building and traversal code from NVidia's "Understanding the Efficiency of Ray Traversal on GPUs": http://code.google.com/p/understanding-the-efficiency-of-ray-traversal-on-gpus/ * Open Shading Language for a large part of the shading system: http://code.google.com/p/openshadinglanguage/ * Blender for procedural textures and a few other nodes. * Approximate Catmull Clark subdivision from NVidia Mesh tools: http://code.google.com/p/nvidia-mesh-tools/ * Sobol direction vectors from: http://web.maths.unsw.edu.au/~fkuo/sobol/ * Film response functions from: http://www.cs.columbia.edu/CAVE/software/softlib/dorf.php
2011-04-27 11:58:34 +00:00
for(size_t i = 0, j = 0; i < bvh_nodes_size; i+=nsize, j++) {
memcpy(pack_nodes + pack_nodes_offset, bvh_nodes + i, nsize_bbox*sizeof(int4));
/* modify offsets into arrays */
int4 data = bvh_nodes[i + nsize_bbox];
fix for rare crash in cycles BVH tree packing.
2012-08-22 11:18:37 +00:00
if(bvh_is_leaf && bvh_is_leaf[j]) {
New feature Patch [#33445] - Experimental Cycles Hair Rendering (CPU only) This patch allows hair data to be exported to cycles and introduces a new line segment primitive to render with. The UI appears under the particle tab and there is a new hair info node available. It is only available under the experimental feature set and for cpu rendering.
2012-12-28 14:21:30 +00:00
data.x += prim_offset;
data.y += prim_offset;
Cycles render engine, initial commit. This is the engine itself, blender modifications and build instructions will follow later. Cycles uses code from some great open source projects, many thanks them: * BVH building and traversal code from NVidia's "Understanding the Efficiency of Ray Traversal on GPUs": http://code.google.com/p/understanding-the-efficiency-of-ray-traversal-on-gpus/ * Open Shading Language for a large part of the shading system: http://code.google.com/p/openshadinglanguage/ * Blender for procedural textures and a few other nodes. * Approximate Catmull Clark subdivision from NVidia Mesh tools: http://code.google.com/p/nvidia-mesh-tools/ * Sobol direction vectors from: http://web.maths.unsw.edu.au/~fkuo/sobol/ * Film response functions from: http://www.cs.columbia.edu/CAVE/software/softlib/dorf.php
2011-04-27 11:58:34 +00:00
}
else {
data.x += (data.x < 0)? -noffset: noffset;
data.y += (data.y < 0)? -noffset: noffset;
if(use_qbvh) {
data.z += (data.z < 0)? -noffset: noffset;
data.w += (data.w < 0)? -noffset: noffset;
}
}
pack_nodes[pack_nodes_offset + nsize_bbox] = data;
Cleanup: Typo fixes for comments.
2014-12-25 01:42:06 +00:00
/* Usually this copies nothing, but we better
* be prepared for possible node size extension.
Cycles: Correction for node tail copy on packing BVH This is harmless for now because tail of the node is zero in there, but better to fix it early so in the case of extending BVH nodes this code doesn't give issues.
2014-12-24 21:49:27 +00:00
*/
memcpy(&pack_nodes[pack_nodes_offset + nsize_bbox+1],
&bvh_nodes[i + nsize_bbox+1],
sizeof(int4) * (nsize - (nsize_bbox+1)));
Cycles render engine, initial commit. This is the engine itself, blender modifications and build instructions will follow later. Cycles uses code from some great open source projects, many thanks them: * BVH building and traversal code from NVidia's "Understanding the Efficiency of Ray Traversal on GPUs": http://code.google.com/p/understanding-the-efficiency-of-ray-traversal-on-gpus/ * Open Shading Language for a large part of the shading system: http://code.google.com/p/openshadinglanguage/ * Blender for procedural textures and a few other nodes. * Approximate Catmull Clark subdivision from NVidia Mesh tools: http://code.google.com/p/nvidia-mesh-tools/ * Sobol direction vectors from: http://web.maths.unsw.edu.au/~fkuo/sobol/ * Film response functions from: http://www.cs.columbia.edu/CAVE/software/softlib/dorf.php
2011-04-27 11:58:34 +00:00
pack_nodes_offset += nsize;
}
}
nodes_offset += bvh->pack.nodes.size();
New feature Patch [#33445] - Experimental Cycles Hair Rendering (CPU only) This patch allows hair data to be exported to cycles and introduces a new line segment primitive to render with. The UI appears under the particle tab and there is a new hair info node available. It is only available under the experimental feature set and for cpu rendering.
2012-12-28 14:21:30 +00:00
prim_offset += bvh->pack.prim_index.size();
Cycles render engine, initial commit. This is the engine itself, blender modifications and build instructions will follow later. Cycles uses code from some great open source projects, many thanks them: * BVH building and traversal code from NVidia's "Understanding the Efficiency of Ray Traversal on GPUs": http://code.google.com/p/understanding-the-efficiency-of-ray-traversal-on-gpus/ * Open Shading Language for a large part of the shading system: http://code.google.com/p/openshadinglanguage/ * Blender for procedural textures and a few other nodes. * Approximate Catmull Clark subdivision from NVidia Mesh tools: http://code.google.com/p/nvidia-mesh-tools/ * Sobol direction vectors from: http://web.maths.unsw.edu.au/~fkuo/sobol/ * Film response functions from: http://www.cs.columbia.edu/CAVE/software/softlib/dorf.php
2011-04-27 11:58:34 +00:00
}
}
/* Regular BVH */
RegularBVH::RegularBVH(const BVHParams& params_, const vector<Object*>& objects_)
: BVH(params_, objects_)
{
}
void RegularBVH::pack_leaf(const BVHStackEntry& e, const LeafNode *leaf)
{
Cycles: BVH params option to split leaf node by primitive types The idea of this change is make it possible to split leaf nodes by primitive type, making leaf containing primitives of the same type. This would become handy when working on a single ray to multiple triangles intersection code, plus with careful implementation it might give some extra benefits on BVH traversal code by avoiding primitive type fetch and check for each primitive in the node. But that's a bit tricky to have benefits on this change only because depth of BVH increases. This option is not exposed to the interface at all and not used even secretly, the commit is only needed to help working further in this direction without messing around with local patches and worrying of them running out of date.
2015-01-09 22:41:04 +00:00
if(leaf->num_triangles() == 1 && pack.prim_index[leaf->m_lo] == -1) {
Cycles render engine, initial commit. This is the engine itself, blender modifications and build instructions will follow later. Cycles uses code from some great open source projects, many thanks them: * BVH building and traversal code from NVidia's "Understanding the Efficiency of Ray Traversal on GPUs": http://code.google.com/p/understanding-the-efficiency-of-ray-traversal-on-gpus/ * Open Shading Language for a large part of the shading system: http://code.google.com/p/openshadinglanguage/ * Blender for procedural textures and a few other nodes. * Approximate Catmull Clark subdivision from NVidia Mesh tools: http://code.google.com/p/nvidia-mesh-tools/ * Sobol direction vectors from: http://web.maths.unsw.edu.au/~fkuo/sobol/ * Film response functions from: http://www.cs.columbia.edu/CAVE/software/softlib/dorf.php
2011-04-27 11:58:34 +00:00
/* object */
Cycles: BVH params option to split leaf node by primitive types The idea of this change is make it possible to split leaf nodes by primitive type, making leaf containing primitives of the same type. This would become handy when working on a single ray to multiple triangles intersection code, plus with careful implementation it might give some extra benefits on BVH traversal code by avoiding primitive type fetch and check for each primitive in the node. But that's a bit tricky to have benefits on this change only because depth of BVH increases. This option is not exposed to the interface at all and not used even secretly, the commit is only needed to help working further in this direction without messing around with local patches and worrying of them running out of date.
2015-01-09 22:41:04 +00:00
pack_node(e.idx, leaf->m_bounds, leaf->m_bounds, ~(leaf->m_lo), 0,
leaf->m_visibility, leaf->m_visibility);
}
else {
Cycles: Fix crash after recent BVH changes on empty BVH trees It's apparently not nice to access 0th element of zero-size vector in C++.
2015-01-12 14:11:32 +00:00
int prim_type = leaf->num_triangles() ? pack.prim_type[leaf->m_lo] : 0;
Cycles: BVH params option to split leaf node by primitive types The idea of this change is make it possible to split leaf nodes by primitive type, making leaf containing primitives of the same type. This would become handy when working on a single ray to multiple triangles intersection code, plus with careful implementation it might give some extra benefits on BVH traversal code by avoiding primitive type fetch and check for each primitive in the node. But that's a bit tricky to have benefits on this change only because depth of BVH increases. This option is not exposed to the interface at all and not used even secretly, the commit is only needed to help working further in this direction without messing around with local patches and worrying of them running out of date.
2015-01-09 22:41:04 +00:00
/* Triangle/curve primitive leaf. */
pack_node(e.idx, leaf->m_bounds, leaf->m_bounds,
leaf->m_lo, leaf->m_hi,
leaf->m_visibility,
Cycles: Fix crash after recent BVH changes on empty BVH trees It's apparently not nice to access 0th element of zero-size vector in C++.
2015-01-12 14:11:32 +00:00
prim_type);
Cycles: BVH params option to split leaf node by primitive types The idea of this change is make it possible to split leaf nodes by primitive type, making leaf containing primitives of the same type. This would become handy when working on a single ray to multiple triangles intersection code, plus with careful implementation it might give some extra benefits on BVH traversal code by avoiding primitive type fetch and check for each primitive in the node. But that's a bit tricky to have benefits on this change only because depth of BVH increases. This option is not exposed to the interface at all and not used even secretly, the commit is only needed to help working further in this direction without messing around with local patches and worrying of them running out of date.
2015-01-09 22:41:04 +00:00
}
Cycles render engine, initial commit. This is the engine itself, blender modifications and build instructions will follow later. Cycles uses code from some great open source projects, many thanks them: * BVH building and traversal code from NVidia's "Understanding the Efficiency of Ray Traversal on GPUs": http://code.google.com/p/understanding-the-efficiency-of-ray-traversal-on-gpus/ * Open Shading Language for a large part of the shading system: http://code.google.com/p/openshadinglanguage/ * Blender for procedural textures and a few other nodes. * Approximate Catmull Clark subdivision from NVidia Mesh tools: http://code.google.com/p/nvidia-mesh-tools/ * Sobol direction vectors from: http://web.maths.unsw.edu.au/~fkuo/sobol/ * Film response functions from: http://www.cs.columbia.edu/CAVE/software/softlib/dorf.php
2011-04-27 11:58:34 +00:00
}
void RegularBVH::pack_inner(const BVHStackEntry& e, const BVHStackEntry& e0, const BVHStackEntry& e1)
{
Cycles: * Add max diffuse/glossy/transmission bounces * Add separate min/max for transparent depth * Updated/added some presets that use these options * Add ray visibility options for objects, to hide them from camera/diffuse/glossy/transmission/shadow rays * Is singular ray output for light path node Details here: http://wiki.blender.org/index.php/Dev:2.5/Source/Render/Cycles/LightPaths
2011-09-01 15:53:36 +00:00
pack_node(e.idx, e0.node->m_bounds, e1.node->m_bounds, e0.encodeIdx(), e1.encodeIdx(), e0.node->m_visibility, e1.node->m_visibility);
Cycles render engine, initial commit. This is the engine itself, blender modifications and build instructions will follow later. Cycles uses code from some great open source projects, many thanks them: * BVH building and traversal code from NVidia's "Understanding the Efficiency of Ray Traversal on GPUs": http://code.google.com/p/understanding-the-efficiency-of-ray-traversal-on-gpus/ * Open Shading Language for a large part of the shading system: http://code.google.com/p/openshadinglanguage/ * Blender for procedural textures and a few other nodes. * Approximate Catmull Clark subdivision from NVidia Mesh tools: http://code.google.com/p/nvidia-mesh-tools/ * Sobol direction vectors from: http://web.maths.unsw.edu.au/~fkuo/sobol/ * Film response functions from: http://www.cs.columbia.edu/CAVE/software/softlib/dorf.php
2011-04-27 11:58:34 +00:00
}
Cycles: * Add max diffuse/glossy/transmission bounces * Add separate min/max for transparent depth * Updated/added some presets that use these options * Add ray visibility options for objects, to hide them from camera/diffuse/glossy/transmission/shadow rays * Is singular ray output for light path node Details here: http://wiki.blender.org/index.php/Dev:2.5/Source/Render/Cycles/LightPaths
2011-09-01 15:53:36 +00:00
void RegularBVH::pack_node(int idx, const BoundBox& b0, const BoundBox& b1, int c0, int c1, uint visibility0, uint visibility1)
Cycles render engine, initial commit. This is the engine itself, blender modifications and build instructions will follow later. Cycles uses code from some great open source projects, many thanks them: * BVH building and traversal code from NVidia's "Understanding the Efficiency of Ray Traversal on GPUs": http://code.google.com/p/understanding-the-efficiency-of-ray-traversal-on-gpus/ * Open Shading Language for a large part of the shading system: http://code.google.com/p/openshadinglanguage/ * Blender for procedural textures and a few other nodes. * Approximate Catmull Clark subdivision from NVidia Mesh tools: http://code.google.com/p/nvidia-mesh-tools/ * Sobol direction vectors from: http://web.maths.unsw.edu.au/~fkuo/sobol/ * Film response functions from: http://www.cs.columbia.edu/CAVE/software/softlib/dorf.php
2011-04-27 11:58:34 +00:00
{
int4 data[BVH_NODE_SIZE] =
{
Cycles: optimization for BVH traveral on CPU's with SSE3, using code from Embree. On the BMW scene, this gives roughly a 10% speedup overall with clang/gcc, and 30% speedup with visual studio (2008). It turns out visual studio was optimizing the existing code quite poorly compared to pretty good autovectorization by clang/gcc, but hand written SSE code also gives a smaller speed boost there. This code isn't enabled when using the hair minimum width feature yet, need to make that work with the SSE code still.
2013-06-18 09:36:06 +00:00
make_int4(__float_as_int(b0.min.x), __float_as_int(b1.min.x), __float_as_int(b0.max.x), __float_as_int(b1.max.x)),
make_int4(__float_as_int(b0.min.y), __float_as_int(b1.min.y), __float_as_int(b0.max.y), __float_as_int(b1.max.y)),
make_int4(__float_as_int(b0.min.z), __float_as_int(b1.min.z), __float_as_int(b0.max.z), __float_as_int(b1.max.z)),
Cycles: * Add max diffuse/glossy/transmission bounces * Add separate min/max for transparent depth * Updated/added some presets that use these options * Add ray visibility options for objects, to hide them from camera/diffuse/glossy/transmission/shadow rays * Is singular ray output for light path node Details here: http://wiki.blender.org/index.php/Dev:2.5/Source/Render/Cycles/LightPaths
2011-09-01 15:53:36 +00:00
make_int4(c0, c1, visibility0, visibility1)
Cycles render engine, initial commit. This is the engine itself, blender modifications and build instructions will follow later. Cycles uses code from some great open source projects, many thanks them: * BVH building and traversal code from NVidia's "Understanding the Efficiency of Ray Traversal on GPUs": http://code.google.com/p/understanding-the-efficiency-of-ray-traversal-on-gpus/ * Open Shading Language for a large part of the shading system: http://code.google.com/p/openshadinglanguage/ * Blender for procedural textures and a few other nodes. * Approximate Catmull Clark subdivision from NVidia Mesh tools: http://code.google.com/p/nvidia-mesh-tools/ * Sobol direction vectors from: http://web.maths.unsw.edu.au/~fkuo/sobol/ * Film response functions from: http://www.cs.columbia.edu/CAVE/software/softlib/dorf.php
2011-04-27 11:58:34 +00:00
};
memcpy(&pack.nodes[idx * BVH_NODE_SIZE], data, sizeof(int4)*BVH_NODE_SIZE);
}
void RegularBVH::pack_nodes(const array<int>& prims, const BVHNode *root)
{
size_t node_size = root->getSubtreeSize(BVH_STAT_NODE_COUNT);
/* resize arrays */
pack.nodes.clear();
pack.is_leaf.clear();
pack.is_leaf.resize(node_size);
/* for top level BVH, first merge existing BVH's so we know the offsets */
if(params.top_level)
pack_instances(node_size*BVH_NODE_SIZE);
else
pack.nodes.resize(node_size*BVH_NODE_SIZE);
int nextNodeIdx = 0;
vector<BVHStackEntry> stack;
Cycles: attempt to fix internal compile error with some visual studio builds
2013-06-18 13:19:16 +00:00
stack.reserve(BVHParams::MAX_DEPTH*2);
Cycles render engine, initial commit. This is the engine itself, blender modifications and build instructions will follow later. Cycles uses code from some great open source projects, many thanks them: * BVH building and traversal code from NVidia's "Understanding the Efficiency of Ray Traversal on GPUs": http://code.google.com/p/understanding-the-efficiency-of-ray-traversal-on-gpus/ * Open Shading Language for a large part of the shading system: http://code.google.com/p/openshadinglanguage/ * Blender for procedural textures and a few other nodes. * Approximate Catmull Clark subdivision from NVidia Mesh tools: http://code.google.com/p/nvidia-mesh-tools/ * Sobol direction vectors from: http://web.maths.unsw.edu.au/~fkuo/sobol/ * Film response functions from: http://www.cs.columbia.edu/CAVE/software/softlib/dorf.php
2011-04-27 11:58:34 +00:00
stack.push_back(BVHStackEntry(root, nextNodeIdx++));
while(stack.size()) {
BVHStackEntry e = stack.back();
stack.pop_back();
pack.is_leaf[e.idx] = e.node->is_leaf();
if(e.node->is_leaf()) {
/* leaf node */
const LeafNode* leaf = reinterpret_cast<const LeafNode*>(e.node);
pack_leaf(e, leaf);
}
else {
/* innner node */
stack.push_back(BVHStackEntry(e.node->get_child(0), nextNodeIdx++));
stack.push_back(BVHStackEntry(e.node->get_child(1), nextNodeIdx++));
pack_inner(e, stack[stack.size()-2], stack[stack.size()-1]);
}
}
/* root index to start traversal at, to handle case of single leaf node */
pack.root_index = (pack.is_leaf[0])? -1: 0;
}
void RegularBVH::refit_nodes()
{
assert(!params.top_level);
Cycles: merging features from tomato branch. === BVH build time optimizations === * BVH building was multithreaded. Not all building is multithreaded, packing and the initial bounding/splitting is still single threaded, but recursive splitting is, which was the main bottleneck. * Object splitting now uses binning rather than sorting of all elements, using code from the Embree raytracer from Intel. http://software.intel.com/en-us/articles/embree-photo-realistic-ray-tracing-kernels/ * Other small changes to avoid allocations, pack memory more tightly, avoid some unnecessary operations, ... These optimizations do not work yet when Spatial Splits are enabled, for that more work is needed. There's also other optimizations still needed, in particular for the case of many low poly objects, the packing step and node memory allocation. BVH raytracing time should remain about the same, but BVH build time should be significantly reduced, test here show speedup of about 5x to 10x on a dual core and 5x to 25x on an 8-core machine, depending on the scene. === Threads === Centralized task scheduler for multithreading, which is basically the CPU device threading code wrapped into something reusable. Basic idea is that there is a single TaskScheduler that keeps a pool of threads, one for each core. Other places in the code can then create a TaskPool that they can drop Tasks in to be executed by the scheduler, and wait for them to complete or cancel them early. === Normal ==== Added a Normal output to the texture coordinate node. This currently gives the object space normal, which is the same under object animation. In the future this might become a "generated" normal so it's also stable for deforming objects, but for now it's already useful for non-deforming objects. === Render Layers === Per render layer Samples control, leaving it to 0 will use the common scene setting. Environment pass will now render environment even if film is set to transparent. Exclude Layers" added. Scene layers (all object that influence the render, directly or indirectly) are shared between all render layers. However sometimes it's useful to leave out some object influence for a particular render layer. That's what this option allows you to do. === Filter Glossy === When using a value higher than 0.0, this will blur glossy reflections after blurry bounces, to reduce noise at the cost of accuracy. 1.0 is a good starting value to tweak. Some light paths have a low probability of being found while contributing much light to the pixel. As a result these light paths will be found in some pixels and not in others, causing fireflies. An example of such a difficult path might be a small light that is causing a small specular highlight on a sharp glossy material, which we are seeing through a rough glossy material. With path tracing it is difficult to find the specular highlight, but if we increase the roughness on the material the highlight gets bigger and softer, and so easier to find. Often this blurring will be hardly noticeable, because we are seeing it through a blurry material anyway, but there are also cases where this will lead to a loss of detail in lighting.
2012-04-28 08:53:59 +00:00
BoundBox bbox = BoundBox::empty;
Cycles: * Add max diffuse/glossy/transmission bounces * Add separate min/max for transparent depth * Updated/added some presets that use these options * Add ray visibility options for objects, to hide them from camera/diffuse/glossy/transmission/shadow rays * Is singular ray output for light path node Details here: http://wiki.blender.org/index.php/Dev:2.5/Source/Render/Cycles/LightPaths
2011-09-01 15:53:36 +00:00
uint visibility = 0;
refit_node(0, (pack.is_leaf[0])? true: false, bbox, visibility);
Cycles render engine, initial commit. This is the engine itself, blender modifications and build instructions will follow later. Cycles uses code from some great open source projects, many thanks them: * BVH building and traversal code from NVidia's "Understanding the Efficiency of Ray Traversal on GPUs": http://code.google.com/p/understanding-the-efficiency-of-ray-traversal-on-gpus/ * Open Shading Language for a large part of the shading system: http://code.google.com/p/openshadinglanguage/ * Blender for procedural textures and a few other nodes. * Approximate Catmull Clark subdivision from NVidia Mesh tools: http://code.google.com/p/nvidia-mesh-tools/ * Sobol direction vectors from: http://web.maths.unsw.edu.au/~fkuo/sobol/ * Film response functions from: http://www.cs.columbia.edu/CAVE/software/softlib/dorf.php
2011-04-27 11:58:34 +00:00
}
Cycles: * Add max diffuse/glossy/transmission bounces * Add separate min/max for transparent depth * Updated/added some presets that use these options * Add ray visibility options for objects, to hide them from camera/diffuse/glossy/transmission/shadow rays * Is singular ray output for light path node Details here: http://wiki.blender.org/index.php/Dev:2.5/Source/Render/Cycles/LightPaths
2011-09-01 15:53:36 +00:00
void RegularBVH::refit_node(int idx, bool leaf, BoundBox& bbox, uint& visibility)
Cycles render engine, initial commit. This is the engine itself, blender modifications and build instructions will follow later. Cycles uses code from some great open source projects, many thanks them: * BVH building and traversal code from NVidia's "Understanding the Efficiency of Ray Traversal on GPUs": http://code.google.com/p/understanding-the-efficiency-of-ray-traversal-on-gpus/ * Open Shading Language for a large part of the shading system: http://code.google.com/p/openshadinglanguage/ * Blender for procedural textures and a few other nodes. * Approximate Catmull Clark subdivision from NVidia Mesh tools: http://code.google.com/p/nvidia-mesh-tools/ * Sobol direction vectors from: http://web.maths.unsw.edu.au/~fkuo/sobol/ * Film response functions from: http://www.cs.columbia.edu/CAVE/software/softlib/dorf.php
2011-04-27 11:58:34 +00:00
{
Cycles: Replace magic constant in the code with actual node size
2014-12-21 17:14:58 +00:00
int4 *data = &pack.nodes[idx*BVH_NODE_SIZE];
Cycles render engine, initial commit. This is the engine itself, blender modifications and build instructions will follow later. Cycles uses code from some great open source projects, many thanks them: * BVH building and traversal code from NVidia's "Understanding the Efficiency of Ray Traversal on GPUs": http://code.google.com/p/understanding-the-efficiency-of-ray-traversal-on-gpus/ * Open Shading Language for a large part of the shading system: http://code.google.com/p/openshadinglanguage/ * Blender for procedural textures and a few other nodes. * Approximate Catmull Clark subdivision from NVidia Mesh tools: http://code.google.com/p/nvidia-mesh-tools/ * Sobol direction vectors from: http://web.maths.unsw.edu.au/~fkuo/sobol/ * Film response functions from: http://www.cs.columbia.edu/CAVE/software/softlib/dorf.php
2011-04-27 11:58:34 +00:00
int c0 = data[3].x;
int c1 = data[3].y;
if(leaf) {
/* refit leaf node */
New feature Patch [#33445] - Experimental Cycles Hair Rendering (CPU only) This patch allows hair data to be exported to cycles and introduces a new line segment primitive to render with. The UI appears under the particle tab and there is a new hair info node available. It is only available under the experimental feature set and for cpu rendering.
2012-12-28 14:21:30 +00:00
for(int prim = c0; prim < c1; prim++) {
int pidx = pack.prim_index[prim];
int tob = pack.prim_object[prim];
Cycles render engine, initial commit. This is the engine itself, blender modifications and build instructions will follow later. Cycles uses code from some great open source projects, many thanks them: * BVH building and traversal code from NVidia's "Understanding the Efficiency of Ray Traversal on GPUs": http://code.google.com/p/understanding-the-efficiency-of-ray-traversal-on-gpus/ * Open Shading Language for a large part of the shading system: http://code.google.com/p/openshadinglanguage/ * Blender for procedural textures and a few other nodes. * Approximate Catmull Clark subdivision from NVidia Mesh tools: http://code.google.com/p/nvidia-mesh-tools/ * Sobol direction vectors from: http://web.maths.unsw.edu.au/~fkuo/sobol/ * Film response functions from: http://www.cs.columbia.edu/CAVE/software/softlib/dorf.php
2011-04-27 11:58:34 +00:00
Object *ob = objects[tob];
New feature Patch [#33445] - Experimental Cycles Hair Rendering (CPU only) This patch allows hair data to be exported to cycles and introduces a new line segment primitive to render with. The UI appears under the particle tab and there is a new hair info node available. It is only available under the experimental feature set and for cpu rendering.
2012-12-28 14:21:30 +00:00
if(pidx == -1) {
Cycles render engine, initial commit. This is the engine itself, blender modifications and build instructions will follow later. Cycles uses code from some great open source projects, many thanks them: * BVH building and traversal code from NVidia's "Understanding the Efficiency of Ray Traversal on GPUs": http://code.google.com/p/understanding-the-efficiency-of-ray-traversal-on-gpus/ * Open Shading Language for a large part of the shading system: http://code.google.com/p/openshadinglanguage/ * Blender for procedural textures and a few other nodes. * Approximate Catmull Clark subdivision from NVidia Mesh tools: http://code.google.com/p/nvidia-mesh-tools/ * Sobol direction vectors from: http://web.maths.unsw.edu.au/~fkuo/sobol/ * Film response functions from: http://www.cs.columbia.edu/CAVE/software/softlib/dorf.php
2011-04-27 11:58:34 +00:00
/* object instance */
bbox.grow(ob->bounds);
}
else {
New feature Patch [#33445] - Experimental Cycles Hair Rendering (CPU only) This patch allows hair data to be exported to cycles and introduces a new line segment primitive to render with. The UI appears under the particle tab and there is a new hair info node available. It is only available under the experimental feature set and for cpu rendering.
2012-12-28 14:21:30 +00:00
/* primitives */
Cycles render engine, initial commit. This is the engine itself, blender modifications and build instructions will follow later. Cycles uses code from some great open source projects, many thanks them: * BVH building and traversal code from NVidia's "Understanding the Efficiency of Ray Traversal on GPUs": http://code.google.com/p/understanding-the-efficiency-of-ray-traversal-on-gpus/ * Open Shading Language for a large part of the shading system: http://code.google.com/p/openshadinglanguage/ * Blender for procedural textures and a few other nodes. * Approximate Catmull Clark subdivision from NVidia Mesh tools: http://code.google.com/p/nvidia-mesh-tools/ * Sobol direction vectors from: http://web.maths.unsw.edu.au/~fkuo/sobol/ * Film response functions from: http://www.cs.columbia.edu/CAVE/software/softlib/dorf.php
2011-04-27 11:58:34 +00:00
const Mesh *mesh = ob->mesh;
Cycles code refactor: changes to make adding new primitive types easier.
2014-03-29 12:03:46 +00:00
if(pack.prim_type[prim] & PRIMITIVE_ALL_CURVE) {
Cycles Hair: refactoring to store curves with the index of the first key and the number of keys in the curve, rather than curve segments with the indices of two keys. ShaderData.segment now stores the segment number in the curve.
2013-01-03 12:09:09 +00:00
/* curves */
int str_offset = (params.top_level)? mesh->curve_offset: 0;
Cycles code refactor: change curve key to float4 for easier storage as attribute.
2014-03-29 12:03:46 +00:00
const Mesh::Curve& curve = mesh->curves[pidx - str_offset];
int k = PRIMITIVE_UNPACK_SEGMENT(pack.prim_type[prim]);
curve.bounds_grow(k, &mesh->curve_keys[0], bbox);
Cycles Hair: Strand Minimum Pixel Size Code is added to restrict the pixel size of strands in cycles. It works best with ribbon primitives and a preset for these is included. It uses distance dependent expansion of the strands and then stochastic strand removal to give a fading. To prevent a slowdown for triangle mesh objects in the BVH an extra visibility flag has been added. It is also only applied for camera rays. The strand width settings are also changed, so that the particle size is not included in the width calculation. Instead there is a separate particle system parameter for width scaling.
2013-04-15 21:38:31 +00:00
visibility |= PATH_RAY_CURVE;
Cycles: add support for curve deformation motion blur.
2014-03-29 12:03:47 +00:00
/* motion curves */
if(mesh->use_motion_blur) {
Attribute *attr = mesh->curve_attributes.find(ATTR_STD_MOTION_VERTEX_POSITION);
if(attr) {
size_t mesh_size = mesh->curve_keys.size();
size_t steps = mesh->motion_steps - 1;
float4 *key_steps = attr->data_float4();
for (size_t i = 0; i < steps; i++)
curve.bounds_grow(k, key_steps + i*mesh_size, bbox);
}
}
New feature Patch [#33445] - Experimental Cycles Hair Rendering (CPU only) This patch allows hair data to be exported to cycles and introduces a new line segment primitive to render with. The UI appears under the particle tab and there is a new hair info node available. It is only available under the experimental feature set and for cpu rendering.
2012-12-28 14:21:30 +00:00
}
else {
/* triangles */
int tri_offset = (params.top_level)? mesh->tri_offset: 0;
Cycles code refactor: change curve key to float4 for easier storage as attribute.
2014-03-29 12:03:46 +00:00
const Mesh::Triangle& triangle = mesh->triangles[pidx - tri_offset];
New feature Patch [#33445] - Experimental Cycles Hair Rendering (CPU only) This patch allows hair data to be exported to cycles and introduces a new line segment primitive to render with. The UI appears under the particle tab and there is a new hair info node available. It is only available under the experimental feature set and for cpu rendering.
2012-12-28 14:21:30 +00:00
const float3 *vpos = &mesh->verts[0];
Cycles code refactor: change curve key to float4 for easier storage as attribute.
2014-03-29 12:03:46 +00:00
triangle.bounds_grow(vpos, bbox);
Cycles: add support for mesh deformation motion blur.
2014-03-29 12:03:47 +00:00
/* motion triangles */
if(mesh->use_motion_blur) {
Attribute *attr = mesh->attributes.find(ATTR_STD_MOTION_VERTEX_POSITION);
if(attr) {
size_t mesh_size = mesh->verts.size();
size_t steps = mesh->motion_steps - 1;
float3 *vert_steps = attr->data_float3();
for (size_t i = 0; i < steps; i++)
triangle.bounds_grow(vert_steps + i*mesh_size, bbox);
}
}
New feature Patch [#33445] - Experimental Cycles Hair Rendering (CPU only) This patch allows hair data to be exported to cycles and introduces a new line segment primitive to render with. The UI appears under the particle tab and there is a new hair info node available. It is only available under the experimental feature set and for cpu rendering.
2012-12-28 14:21:30 +00:00
}
Cycles render engine, initial commit. This is the engine itself, blender modifications and build instructions will follow later. Cycles uses code from some great open source projects, many thanks them: * BVH building and traversal code from NVidia's "Understanding the Efficiency of Ray Traversal on GPUs": http://code.google.com/p/understanding-the-efficiency-of-ray-traversal-on-gpus/ * Open Shading Language for a large part of the shading system: http://code.google.com/p/openshadinglanguage/ * Blender for procedural textures and a few other nodes. * Approximate Catmull Clark subdivision from NVidia Mesh tools: http://code.google.com/p/nvidia-mesh-tools/ * Sobol direction vectors from: http://web.maths.unsw.edu.au/~fkuo/sobol/ * Film response functions from: http://www.cs.columbia.edu/CAVE/software/softlib/dorf.php
2011-04-27 11:58:34 +00:00
}
Cycles: * Add max diffuse/glossy/transmission bounces * Add separate min/max for transparent depth * Updated/added some presets that use these options * Add ray visibility options for objects, to hide them from camera/diffuse/glossy/transmission/shadow rays * Is singular ray output for light path node Details here: http://wiki.blender.org/index.php/Dev:2.5/Source/Render/Cycles/LightPaths
2011-09-01 15:53:36 +00:00
visibility |= ob->visibility;
Cycles render engine, initial commit. This is the engine itself, blender modifications and build instructions will follow later. Cycles uses code from some great open source projects, many thanks them: * BVH building and traversal code from NVidia's "Understanding the Efficiency of Ray Traversal on GPUs": http://code.google.com/p/understanding-the-efficiency-of-ray-traversal-on-gpus/ * Open Shading Language for a large part of the shading system: http://code.google.com/p/openshadinglanguage/ * Blender for procedural textures and a few other nodes. * Approximate Catmull Clark subdivision from NVidia Mesh tools: http://code.google.com/p/nvidia-mesh-tools/ * Sobol direction vectors from: http://web.maths.unsw.edu.au/~fkuo/sobol/ * Film response functions from: http://www.cs.columbia.edu/CAVE/software/softlib/dorf.php
2011-04-27 11:58:34 +00:00
}
Cycles: BVH params option to split leaf node by primitive types The idea of this change is make it possible to split leaf nodes by primitive type, making leaf containing primitives of the same type. This would become handy when working on a single ray to multiple triangles intersection code, plus with careful implementation it might give some extra benefits on BVH traversal code by avoiding primitive type fetch and check for each primitive in the node. But that's a bit tricky to have benefits on this change only because depth of BVH increases. This option is not exposed to the interface at all and not used even secretly, the commit is only needed to help working further in this direction without messing around with local patches and worrying of them running out of date.
2015-01-09 22:41:04 +00:00
pack_node(idx, bbox, bbox, c0, c1, visibility, data[3].w);
Cycles render engine, initial commit. This is the engine itself, blender modifications and build instructions will follow later. Cycles uses code from some great open source projects, many thanks them: * BVH building and traversal code from NVidia's "Understanding the Efficiency of Ray Traversal on GPUs": http://code.google.com/p/understanding-the-efficiency-of-ray-traversal-on-gpus/ * Open Shading Language for a large part of the shading system: http://code.google.com/p/openshadinglanguage/ * Blender for procedural textures and a few other nodes. * Approximate Catmull Clark subdivision from NVidia Mesh tools: http://code.google.com/p/nvidia-mesh-tools/ * Sobol direction vectors from: http://web.maths.unsw.edu.au/~fkuo/sobol/ * Film response functions from: http://www.cs.columbia.edu/CAVE/software/softlib/dorf.php
2011-04-27 11:58:34 +00:00
}
else {
/* refit inner node, set bbox from children */
Cycles: merging features from tomato branch. === BVH build time optimizations === * BVH building was multithreaded. Not all building is multithreaded, packing and the initial bounding/splitting is still single threaded, but recursive splitting is, which was the main bottleneck. * Object splitting now uses binning rather than sorting of all elements, using code from the Embree raytracer from Intel. http://software.intel.com/en-us/articles/embree-photo-realistic-ray-tracing-kernels/ * Other small changes to avoid allocations, pack memory more tightly, avoid some unnecessary operations, ... These optimizations do not work yet when Spatial Splits are enabled, for that more work is needed. There's also other optimizations still needed, in particular for the case of many low poly objects, the packing step and node memory allocation. BVH raytracing time should remain about the same, but BVH build time should be significantly reduced, test here show speedup of about 5x to 10x on a dual core and 5x to 25x on an 8-core machine, depending on the scene. === Threads === Centralized task scheduler for multithreading, which is basically the CPU device threading code wrapped into something reusable. Basic idea is that there is a single TaskScheduler that keeps a pool of threads, one for each core. Other places in the code can then create a TaskPool that they can drop Tasks in to be executed by the scheduler, and wait for them to complete or cancel them early. === Normal ==== Added a Normal output to the texture coordinate node. This currently gives the object space normal, which is the same under object animation. In the future this might become a "generated" normal so it's also stable for deforming objects, but for now it's already useful for non-deforming objects. === Render Layers === Per render layer Samples control, leaving it to 0 will use the common scene setting. Environment pass will now render environment even if film is set to transparent. Exclude Layers" added. Scene layers (all object that influence the render, directly or indirectly) are shared between all render layers. However sometimes it's useful to leave out some object influence for a particular render layer. That's what this option allows you to do. === Filter Glossy === When using a value higher than 0.0, this will blur glossy reflections after blurry bounces, to reduce noise at the cost of accuracy. 1.0 is a good starting value to tweak. Some light paths have a low probability of being found while contributing much light to the pixel. As a result these light paths will be found in some pixels and not in others, causing fireflies. An example of such a difficult path might be a small light that is causing a small specular highlight on a sharp glossy material, which we are seeing through a rough glossy material. With path tracing it is difficult to find the specular highlight, but if we increase the roughness on the material the highlight gets bigger and softer, and so easier to find. Often this blurring will be hardly noticeable, because we are seeing it through a blurry material anyway, but there are also cases where this will lead to a loss of detail in lighting.
2012-04-28 08:53:59 +00:00
BoundBox bbox0 = BoundBox::empty, bbox1 = BoundBox::empty;
Cycles: * Add max diffuse/glossy/transmission bounces * Add separate min/max for transparent depth * Updated/added some presets that use these options * Add ray visibility options for objects, to hide them from camera/diffuse/glossy/transmission/shadow rays * Is singular ray output for light path node Details here: http://wiki.blender.org/index.php/Dev:2.5/Source/Render/Cycles/LightPaths
2011-09-01 15:53:36 +00:00
uint visibility0 = 0, visibility1 = 0;
Cycles render engine, initial commit. This is the engine itself, blender modifications and build instructions will follow later. Cycles uses code from some great open source projects, many thanks them: * BVH building and traversal code from NVidia's "Understanding the Efficiency of Ray Traversal on GPUs": http://code.google.com/p/understanding-the-efficiency-of-ray-traversal-on-gpus/ * Open Shading Language for a large part of the shading system: http://code.google.com/p/openshadinglanguage/ * Blender for procedural textures and a few other nodes. * Approximate Catmull Clark subdivision from NVidia Mesh tools: http://code.google.com/p/nvidia-mesh-tools/ * Sobol direction vectors from: http://web.maths.unsw.edu.au/~fkuo/sobol/ * Film response functions from: http://www.cs.columbia.edu/CAVE/software/softlib/dorf.php
2011-04-27 11:58:34 +00:00
Cycles: * Add max diffuse/glossy/transmission bounces * Add separate min/max for transparent depth * Updated/added some presets that use these options * Add ray visibility options for objects, to hide them from camera/diffuse/glossy/transmission/shadow rays * Is singular ray output for light path node Details here: http://wiki.blender.org/index.php/Dev:2.5/Source/Render/Cycles/LightPaths
2011-09-01 15:53:36 +00:00
refit_node((c0 < 0)? -c0-1: c0, (c0 < 0), bbox0, visibility0);
refit_node((c1 < 0)? -c1-1: c1, (c1 < 0), bbox1, visibility1);
pack_node(idx, bbox0, bbox1, c0, c1, visibility0, visibility1);
Cycles render engine, initial commit. This is the engine itself, blender modifications and build instructions will follow later. Cycles uses code from some great open source projects, many thanks them: * BVH building and traversal code from NVidia's "Understanding the Efficiency of Ray Traversal on GPUs": http://code.google.com/p/understanding-the-efficiency-of-ray-traversal-on-gpus/ * Open Shading Language for a large part of the shading system: http://code.google.com/p/openshadinglanguage/ * Blender for procedural textures and a few other nodes. * Approximate Catmull Clark subdivision from NVidia Mesh tools: http://code.google.com/p/nvidia-mesh-tools/ * Sobol direction vectors from: http://web.maths.unsw.edu.au/~fkuo/sobol/ * Film response functions from: http://www.cs.columbia.edu/CAVE/software/softlib/dorf.php
2011-04-27 11:58:34 +00:00
bbox.grow(bbox0);
bbox.grow(bbox1);
Cycles: * Add max diffuse/glossy/transmission bounces * Add separate min/max for transparent depth * Updated/added some presets that use these options * Add ray visibility options for objects, to hide them from camera/diffuse/glossy/transmission/shadow rays * Is singular ray output for light path node Details here: http://wiki.blender.org/index.php/Dev:2.5/Source/Render/Cycles/LightPaths
2011-09-01 15:53:36 +00:00
visibility = visibility0|visibility1;
Cycles render engine, initial commit. This is the engine itself, blender modifications and build instructions will follow later. Cycles uses code from some great open source projects, many thanks them: * BVH building and traversal code from NVidia's "Understanding the Efficiency of Ray Traversal on GPUs": http://code.google.com/p/understanding-the-efficiency-of-ray-traversal-on-gpus/ * Open Shading Language for a large part of the shading system: http://code.google.com/p/openshadinglanguage/ * Blender for procedural textures and a few other nodes. * Approximate Catmull Clark subdivision from NVidia Mesh tools: http://code.google.com/p/nvidia-mesh-tools/ * Sobol direction vectors from: http://web.maths.unsw.edu.au/~fkuo/sobol/ * Film response functions from: http://www.cs.columbia.edu/CAVE/software/softlib/dorf.php
2011-04-27 11:58:34 +00:00
}
}
/* QBVH */
QBVH::QBVH(const BVHParams& params_, const vector<Object*>& objects_)
: BVH(params_, objects_)
{
params.use_qbvh = true;
}
void QBVH::pack_leaf(const BVHStackEntry& e, const LeafNode *leaf)
{
float4 data[BVH_QNODE_SIZE];
memset(data, 0, sizeof(data));
if(leaf->num_triangles() == 1 && pack.prim_index[leaf->m_lo] == -1) {
/* object */
data[6].x = __int_as_float(~(leaf->m_lo));
data[6].y = __int_as_float(0);
}
else {
/* triangle */
data[6].x = __int_as_float(leaf->m_lo);
data[6].y = __int_as_float(leaf->m_hi);
}
Cycles: Optimize storage of QBVH node by one float4 The idea is to store visibility flags for leaf nodes only since visibility check for inner nodes costs too much for QBVH hence it is not optimal to perform. Leaf QBVH nodes have plenty of space to store all sort of flags, so we can make nodes one element smaller, saving noticeable amount of memory.
2014-12-22 19:58:19 +00:00
data[6].z = __uint_as_float(leaf->m_visibility);
Cycles: Fix crash after recent BVH changes on empty BVH trees It's apparently not nice to access 0th element of zero-size vector in C++.
2015-01-12 14:11:32 +00:00
if(leaf->num_triangles() != 0) {
data[6].w = __uint_as_float(pack.prim_type[leaf->m_lo]);
}
Cycles render engine, initial commit. This is the engine itself, blender modifications and build instructions will follow later. Cycles uses code from some great open source projects, many thanks them: * BVH building and traversal code from NVidia's "Understanding the Efficiency of Ray Traversal on GPUs": http://code.google.com/p/understanding-the-efficiency-of-ray-traversal-on-gpus/ * Open Shading Language for a large part of the shading system: http://code.google.com/p/openshadinglanguage/ * Blender for procedural textures and a few other nodes. * Approximate Catmull Clark subdivision from NVidia Mesh tools: http://code.google.com/p/nvidia-mesh-tools/ * Sobol direction vectors from: http://web.maths.unsw.edu.au/~fkuo/sobol/ * Film response functions from: http://www.cs.columbia.edu/CAVE/software/softlib/dorf.php
2011-04-27 11:58:34 +00:00
memcpy(&pack.nodes[e.idx * BVH_QNODE_SIZE], data, sizeof(float4)*BVH_QNODE_SIZE);
}
void QBVH::pack_inner(const BVHStackEntry& e, const BVHStackEntry *en, int num)
{
float4 data[BVH_QNODE_SIZE];
for(int i = 0; i < num; i++) {
float3 bb_min = en[i].node->m_bounds.min;
float3 bb_max = en[i].node->m_bounds.max;
data[0][i] = bb_min.x;
data[1][i] = bb_max.x;
data[2][i] = bb_min.y;
data[3][i] = bb_max.y;
data[4][i] = bb_min.z;
data[5][i] = bb_max.z;
data[6][i] = __int_as_float(en[i].encodeIdx());
}
for(int i = num; i < 4; i++) {
Cycles: Store proper empty boundbox for missing child nodes in QBVH The idea is to make sure those childs would never be intersected with a ray in order to make it so kernel never worries about number of child nodes.
2014-12-24 19:14:12 +00:00
/* We store BB which would never be recorded as intersection
* so kernel might safely assume there are always 4 child nodes.
*/
data[0][i] = FLT_MAX;
data[1][i] = -FLT_MAX;
Cycles render engine, initial commit. This is the engine itself, blender modifications and build instructions will follow later. Cycles uses code from some great open source projects, many thanks them: * BVH building and traversal code from NVidia's "Understanding the Efficiency of Ray Traversal on GPUs": http://code.google.com/p/understanding-the-efficiency-of-ray-traversal-on-gpus/ * Open Shading Language for a large part of the shading system: http://code.google.com/p/openshadinglanguage/ * Blender for procedural textures and a few other nodes. * Approximate Catmull Clark subdivision from NVidia Mesh tools: http://code.google.com/p/nvidia-mesh-tools/ * Sobol direction vectors from: http://web.maths.unsw.edu.au/~fkuo/sobol/ * Film response functions from: http://www.cs.columbia.edu/CAVE/software/softlib/dorf.php
2011-04-27 11:58:34 +00:00
Cycles: Store proper empty boundbox for missing child nodes in QBVH The idea is to make sure those childs would never be intersected with a ray in order to make it so kernel never worries about number of child nodes.
2014-12-24 19:14:12 +00:00
data[2][i] = FLT_MAX;
data[3][i] = -FLT_MAX;
data[4][i] = FLT_MAX;
data[5][i] = -FLT_MAX;
Cycles render engine, initial commit. This is the engine itself, blender modifications and build instructions will follow later. Cycles uses code from some great open source projects, many thanks them: * BVH building and traversal code from NVidia's "Understanding the Efficiency of Ray Traversal on GPUs": http://code.google.com/p/understanding-the-efficiency-of-ray-traversal-on-gpus/ * Open Shading Language for a large part of the shading system: http://code.google.com/p/openshadinglanguage/ * Blender for procedural textures and a few other nodes. * Approximate Catmull Clark subdivision from NVidia Mesh tools: http://code.google.com/p/nvidia-mesh-tools/ * Sobol direction vectors from: http://web.maths.unsw.edu.au/~fkuo/sobol/ * Film response functions from: http://www.cs.columbia.edu/CAVE/software/softlib/dorf.php
2011-04-27 11:58:34 +00:00
data[6][i] = __int_as_float(0);
}
memcpy(&pack.nodes[e.idx * BVH_QNODE_SIZE], data, sizeof(float4)*BVH_QNODE_SIZE);
}
/* Quad SIMD Nodes */
void QBVH::pack_nodes(const array<int>& prims, const BVHNode *root)
{
Cycles: Use proper node counter to allocate QBVH nodes Before all the nodes were counted and allocated, leading to situations when bunch of allocated memory is not used because reasonable amount of nodes are simply ignored.
2014-12-17 16:49:35 +00:00
size_t node_size = root->getSubtreeSize(BVH_STAT_QNODE_COUNT);
Cycles render engine, initial commit. This is the engine itself, blender modifications and build instructions will follow later. Cycles uses code from some great open source projects, many thanks them: * BVH building and traversal code from NVidia's "Understanding the Efficiency of Ray Traversal on GPUs": http://code.google.com/p/understanding-the-efficiency-of-ray-traversal-on-gpus/ * Open Shading Language for a large part of the shading system: http://code.google.com/p/openshadinglanguage/ * Blender for procedural textures and a few other nodes. * Approximate Catmull Clark subdivision from NVidia Mesh tools: http://code.google.com/p/nvidia-mesh-tools/ * Sobol direction vectors from: http://web.maths.unsw.edu.au/~fkuo/sobol/ * Film response functions from: http://www.cs.columbia.edu/CAVE/software/softlib/dorf.php
2011-04-27 11:58:34 +00:00
/* resize arrays */
pack.nodes.clear();
pack.is_leaf.clear();
pack.is_leaf.resize(node_size);
/* for top level BVH, first merge existing BVH's so we know the offsets */
if(params.top_level)
pack_instances(node_size*BVH_QNODE_SIZE);
else
pack.nodes.resize(node_size*BVH_QNODE_SIZE);
int nextNodeIdx = 0;
vector<BVHStackEntry> stack;
Cycles: attempt to fix internal compile error with some visual studio builds
2013-06-18 13:19:16 +00:00
stack.reserve(BVHParams::MAX_DEPTH*2);
Cycles render engine, initial commit. This is the engine itself, blender modifications and build instructions will follow later. Cycles uses code from some great open source projects, many thanks them: * BVH building and traversal code from NVidia's "Understanding the Efficiency of Ray Traversal on GPUs": http://code.google.com/p/understanding-the-efficiency-of-ray-traversal-on-gpus/ * Open Shading Language for a large part of the shading system: http://code.google.com/p/openshadinglanguage/ * Blender for procedural textures and a few other nodes. * Approximate Catmull Clark subdivision from NVidia Mesh tools: http://code.google.com/p/nvidia-mesh-tools/ * Sobol direction vectors from: http://web.maths.unsw.edu.au/~fkuo/sobol/ * Film response functions from: http://www.cs.columbia.edu/CAVE/software/softlib/dorf.php
2011-04-27 11:58:34 +00:00
stack.push_back(BVHStackEntry(root, nextNodeIdx++));
while(stack.size()) {
BVHStackEntry e = stack.back();
stack.pop_back();
pack.is_leaf[e.idx] = e.node->is_leaf();
if(e.node->is_leaf()) {
/* leaf node */
const LeafNode* leaf = reinterpret_cast<const LeafNode*>(e.node);
pack_leaf(e, leaf);
}
else {
/* inner node */
const BVHNode *node = e.node;
const BVHNode *node0 = node->get_child(0);
const BVHNode *node1 = node->get_child(1);
/* collect nodes */
const BVHNode *nodes[4];
int numnodes = 0;
if(node0->is_leaf()) {
nodes[numnodes++] = node0;
}
else {
nodes[numnodes++] = node0->get_child(0);
nodes[numnodes++] = node0->get_child(1);
}
if(node1->is_leaf()) {
nodes[numnodes++] = node1;
}
else {
nodes[numnodes++] = node1->get_child(0);
nodes[numnodes++] = node1->get_child(1);
}
/* push entries on the stack */
for(int i = 0; i < numnodes; i++)
stack.push_back(BVHStackEntry(nodes[i], nextNodeIdx++));
/* set node */
pack_inner(e, &stack[stack.size()-numnodes], numnodes);
}
}
/* root index to start traversal at, to handle case of single leaf node */
pack.root_index = (pack.is_leaf[0])? -1: 0;
}
void QBVH::refit_nodes()
{
Cycles: Implement refit_nodes for QBVH Title says it all, quite straightforward implementation. Would only mention that there's a bit of code duplication around packing node into pack.nodes. Trying to de-duplicate it ends up in quite hairy code (like functions with loads of arguments some of which could be NULL in certain circumstances etc..). Leaving solving this duplication for later.
2014-12-17 17:46:02 +00:00
assert(!params.top_level);
BoundBox bbox = BoundBox::empty;
uint visibility = 0;
refit_node(0, (pack.is_leaf[0])? true: false, bbox, visibility);
Cycles render engine, initial commit. This is the engine itself, blender modifications and build instructions will follow later. Cycles uses code from some great open source projects, many thanks them: * BVH building and traversal code from NVidia's "Understanding the Efficiency of Ray Traversal on GPUs": http://code.google.com/p/understanding-the-efficiency-of-ray-traversal-on-gpus/ * Open Shading Language for a large part of the shading system: http://code.google.com/p/openshadinglanguage/ * Blender for procedural textures and a few other nodes. * Approximate Catmull Clark subdivision from NVidia Mesh tools: http://code.google.com/p/nvidia-mesh-tools/ * Sobol direction vectors from: http://web.maths.unsw.edu.au/~fkuo/sobol/ * Film response functions from: http://www.cs.columbia.edu/CAVE/software/softlib/dorf.php
2011-04-27 11:58:34 +00:00
}
Cycles: Implement refit_nodes for QBVH Title says it all, quite straightforward implementation. Would only mention that there's a bit of code duplication around packing node into pack.nodes. Trying to de-duplicate it ends up in quite hairy code (like functions with loads of arguments some of which could be NULL in certain circumstances etc..). Leaving solving this duplication for later.
2014-12-17 17:46:02 +00:00
void QBVH::refit_node(int idx, bool leaf, BoundBox& bbox, uint& visibility)
{
int4 *data = &pack.nodes[idx*BVH_QNODE_SIZE];
int4 c = data[6];
if(leaf) {
/* Refit leaf node. */
for(int prim = c.x; prim < c.y; prim++) {
int pidx = pack.prim_index[prim];
int tob = pack.prim_object[prim];
Object *ob = objects[tob];
if(pidx == -1) {
/* Object instance. */
bbox.grow(ob->bounds);
}
else {
/* Primitives. */
const Mesh *mesh = ob->mesh;
if(pack.prim_type[prim] & PRIMITIVE_ALL_CURVE) {
/* Curves. */
int str_offset = (params.top_level)? mesh->curve_offset: 0;
const Mesh::Curve& curve = mesh->curves[pidx - str_offset];
int k = PRIMITIVE_UNPACK_SEGMENT(pack.prim_type[prim]);
curve.bounds_grow(k, &mesh->curve_keys[0], bbox);
visibility |= PATH_RAY_CURVE;
/* Motion curves. */
if(mesh->use_motion_blur) {
Attribute *attr = mesh->curve_attributes.find(ATTR_STD_MOTION_VERTEX_POSITION);
if(attr) {
size_t mesh_size = mesh->curve_keys.size();
size_t steps = mesh->motion_steps - 1;
float4 *key_steps = attr->data_float4();
for (size_t i = 0; i < steps; i++)
curve.bounds_grow(k, key_steps + i*mesh_size, bbox);
}
}
}
else {
/* Triangles. */
int tri_offset = (params.top_level)? mesh->tri_offset: 0;
const Mesh::Triangle& triangle = mesh->triangles[pidx - tri_offset];
const float3 *vpos = &mesh->verts[0];
Cycles render engine, initial commit. This is the engine itself, blender modifications and build instructions will follow later. Cycles uses code from some great open source projects, many thanks them: * BVH building and traversal code from NVidia's "Understanding the Efficiency of Ray Traversal on GPUs": http://code.google.com/p/understanding-the-efficiency-of-ray-traversal-on-gpus/ * Open Shading Language for a large part of the shading system: http://code.google.com/p/openshadinglanguage/ * Blender for procedural textures and a few other nodes. * Approximate Catmull Clark subdivision from NVidia Mesh tools: http://code.google.com/p/nvidia-mesh-tools/ * Sobol direction vectors from: http://web.maths.unsw.edu.au/~fkuo/sobol/ * Film response functions from: http://www.cs.columbia.edu/CAVE/software/softlib/dorf.php
2011-04-27 11:58:34 +00:00
Cycles: Implement refit_nodes for QBVH Title says it all, quite straightforward implementation. Would only mention that there's a bit of code duplication around packing node into pack.nodes. Trying to de-duplicate it ends up in quite hairy code (like functions with loads of arguments some of which could be NULL in certain circumstances etc..). Leaving solving this duplication for later.
2014-12-17 17:46:02 +00:00
triangle.bounds_grow(vpos, bbox);
/* Motion triangles. */
if(mesh->use_motion_blur) {
Attribute *attr = mesh->attributes.find(ATTR_STD_MOTION_VERTEX_POSITION);
if(attr) {
size_t mesh_size = mesh->verts.size();
size_t steps = mesh->motion_steps - 1;
float3 *vert_steps = attr->data_float3();
for (size_t i = 0; i < steps; i++)
triangle.bounds_grow(vert_steps + i*mesh_size, bbox);
}
}
}
}
visibility |= ob->visibility;
}
/* TODO(sergey): This is actually a copy of pack_leaf(),
* but this chunk of code only knows actual data and has
* no idea about BVHNode.
*
* Would be nice to de-duplicate code, but trying to make
cleanup: style
2015-01-02 08:18:57 +00:00
* making code more general ends up in much nastier code
Cycles: Implement refit_nodes for QBVH Title says it all, quite straightforward implementation. Would only mention that there's a bit of code duplication around packing node into pack.nodes. Trying to de-duplicate it ends up in quite hairy code (like functions with loads of arguments some of which could be NULL in certain circumstances etc..). Leaving solving this duplication for later.
2014-12-17 17:46:02 +00:00
* in my opinion so far.
*
* Same applies to the inner nodes case below.
*/
float4 leaf_data[BVH_QNODE_SIZE];
memset(leaf_data, 0, sizeof(leaf_data));
leaf_data[6].x = __int_as_float(c.x);
leaf_data[6].y = __int_as_float(c.y);
Cycles: Optimize storage of QBVH node by one float4 The idea is to store visibility flags for leaf nodes only since visibility check for inner nodes costs too much for QBVH hence it is not optimal to perform. Leaf QBVH nodes have plenty of space to store all sort of flags, so we can make nodes one element smaller, saving noticeable amount of memory.
2014-12-22 19:58:19 +00:00
leaf_data[6].z = __uint_as_float(visibility);
Cycles: Fix QBVH refit nodes not setting primitive type properly
2015-01-13 21:16:45 +00:00
leaf_data[6].w = __uint_as_float(c.w);
Cycles: Implement refit_nodes for QBVH Title says it all, quite straightforward implementation. Would only mention that there's a bit of code duplication around packing node into pack.nodes. Trying to de-duplicate it ends up in quite hairy code (like functions with loads of arguments some of which could be NULL in certain circumstances etc..). Leaving solving this duplication for later.
2014-12-17 17:46:02 +00:00
memcpy(&pack.nodes[idx * BVH_QNODE_SIZE],
leaf_data,
sizeof(float4)*BVH_QNODE_SIZE);
}
else {
/* Refit inner node, set bbox from children. */
BoundBox child_bbox[4] = {BoundBox::empty,
BoundBox::empty,
BoundBox::empty,
BoundBox::empty};
uint child_visibility[4] = {0};
int num_nodes = 0;
for(int i = 0; i < 4; ++i) {
if(c[i] != 0) {
refit_node((c[i] < 0)? -c[i]-1: c[i], (c[i] < 0),
child_bbox[i], child_visibility[i]);
++num_nodes;
bbox.grow(child_bbox[i]);
visibility |= child_visibility[i];
}
}
float4 inner_data[BVH_QNODE_SIZE];
for(int i = 0; i < 4; ++i) {
float3 bb_min = child_bbox[i].min;
float3 bb_max = child_bbox[i].max;
inner_data[0][i] = bb_min.x;
inner_data[1][i] = bb_max.x;
inner_data[2][i] = bb_min.y;
inner_data[3][i] = bb_max.y;
inner_data[4][i] = bb_min.z;
inner_data[5][i] = bb_max.z;
inner_data[6][i] = __int_as_float(c[i]);
}
memcpy(&pack.nodes[idx * BVH_QNODE_SIZE],
inner_data,
sizeof(float4)*BVH_QNODE_SIZE);
}
}
CCL_NAMESPACE_END
Reference in New Issue Copy Permalink