sandey/blender
1
0
Fork 0
forked from bartvdbraak/blender
Code Pull Requests Activity
5475314f49
blender/intern/cycles/bvh/bvh_split.cpp

555 lines
18 KiB
C++
Raw Normal View History

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
/*
* 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.
*/
Cycles: Make all #include statements relative to cycles source directory The idea is to make include statements more explicit and obvious where the file is coming from, additionally reducing chance of wrong header being picked up. For example, it was not obvious whether bvh.h was refferring to builder or traversal, whenter node.h is a generic graph node or a shader node and cases like that. Surely this might look obvious for the active developers, but after some time of not touching the code it becomes less obvious where file is coming from. This was briefly mentioned in T50824 and seems @brecht is fine with such explicitness, but need to agree with all active developers before committing this. Please note that this patch is lacking changes related on GPU/OpenCL support. This will be solved if/when we all agree this is a good idea to move forward. Reviewers: brecht, lukasstockner97, maiself, nirved, dingto, juicyfruit, swerner Reviewed By: lukasstockner97, maiself, nirved, dingto Subscribers: brecht Differential Revision: https://developer.blender.org/D2586
2017-03-28 18:39:14 +00:00
#include "bvh/bvh_split.h"
Cycles: Header cleanup in BVH: move self header to be the first one This makes us more sure that header files are more self-sufficient.
2017-04-13 09:20:33 +00:00
#include "bvh/bvh_build.h"
Cycles: Make all #include statements relative to cycles source directory The idea is to make include statements more explicit and obvious where the file is coming from, additionally reducing chance of wrong header being picked up. For example, it was not obvious whether bvh.h was refferring to builder or traversal, whenter node.h is a generic graph node or a shader node and cases like that. Surely this might look obvious for the active developers, but after some time of not touching the code it becomes less obvious where file is coming from. This was briefly mentioned in T50824 and seems @brecht is fine with such explicitness, but need to agree with all active developers before committing this. Please note that this patch is lacking changes related on GPU/OpenCL support. This will be solved if/when we all agree this is a good idea to move forward. Reviewers: brecht, lukasstockner97, maiself, nirved, dingto, juicyfruit, swerner Reviewed By: lukasstockner97, maiself, nirved, dingto Subscribers: brecht Differential Revision: https://developer.blender.org/D2586
2017-03-28 18:39:14 +00:00
#include "bvh/bvh_sort.h"
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
Cycles: Make all #include statements relative to cycles source directory The idea is to make include statements more explicit and obvious where the file is coming from, additionally reducing chance of wrong header being picked up. For example, it was not obvious whether bvh.h was refferring to builder or traversal, whenter node.h is a generic graph node or a shader node and cases like that. Surely this might look obvious for the active developers, but after some time of not touching the code it becomes less obvious where file is coming from. This was briefly mentioned in T50824 and seems @brecht is fine with such explicitness, but need to agree with all active developers before committing this. Please note that this patch is lacking changes related on GPU/OpenCL support. This will be solved if/when we all agree this is a good idea to move forward. Reviewers: brecht, lukasstockner97, maiself, nirved, dingto, juicyfruit, swerner Reviewed By: lukasstockner97, maiself, nirved, dingto Subscribers: brecht Differential Revision: https://developer.blender.org/D2586
2017-03-28 18:39:14 +00:00
#include "render/mesh.h"
#include "render/object.h"
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
Cycles: Make all #include statements relative to cycles source directory The idea is to make include statements more explicit and obvious where the file is coming from, additionally reducing chance of wrong header being picked up. For example, it was not obvious whether bvh.h was refferring to builder or traversal, whenter node.h is a generic graph node or a shader node and cases like that. Surely this might look obvious for the active developers, but after some time of not touching the code it becomes less obvious where file is coming from. This was briefly mentioned in T50824 and seems @brecht is fine with such explicitness, but need to agree with all active developers before committing this. Please note that this patch is lacking changes related on GPU/OpenCL support. This will be solved if/when we all agree this is a good idea to move forward. Reviewers: brecht, lukasstockner97, maiself, nirved, dingto, juicyfruit, swerner Reviewed By: lukasstockner97, maiself, nirved, dingto Subscribers: brecht Differential Revision: https://developer.blender.org/D2586
2017-03-28 18:39:14 +00:00
#include "util/util_algorithm.h"
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
CCL_NAMESPACE_BEGIN
/* Object Split */
Cycles: Temporarily revert index sort commit for spatial split There are in fact some missing parts to it (Split BVH builder should be creating bins from result of Object Split constructor). Doable, but need to quickly fix issue for the studio here, easier to revert for now.
2016-04-01 15:44:08 +00:00
BVHObjectSplit::BVHObjectSplit(BVHBuild *builder,
Cycles: Wrap spatial split storage into own structure This has following advantages: - Localizes all the run-time storage into a single structure, which could easily be extended further. - Storage could be created per-thread, so once builder is threaded we wouldn't have any conflicts between threads. - Global nature of the storage avoids memory re-allocation on the runtime, keeping builder as fast as possible. Currently it's just API changes, which don't affect user at all.
2016-02-21 14:39:02 +00:00
BVHSpatialStorage *storage,
const BVHRange& range,
Cycles: Make spatial split BVH multi-threaded The title actually covers it all, This commit exploits all the work being done in previous changes to make it possible to build spatial splits in threads. Works quite nicely, but has a downside of some extra memory usage. In practice it doesn't seem to be a huge problem and that we can always look into later if it becomes a real showstopper. In practice it shows some nice speedup: - BMW27 scene takes 3 now (used to be 4) - Agent shot takes 5 sec (used to be 80) Such non-linear speedup is most likely coming from much less amount of heap re-allocations. A a downside, there's a bit of extra memory used by BVH arrays. From the tests amount of extra memory is below 0.001% so far, so it's not that bad at all. Reviewers: brecht, juicyfruit, dingto, lukasstockner97 Differential Revision: https://developer.blender.org/D1820
2016-04-04 12:43:21 +00:00
vector<BVHReference> *references,
Cycles: Implement unaligned nodes BVH builder This is a special builder type which is allowed to orient nodes to strands direction, hence minimizing their surface area in comparison with axis-aligned nodes. Such nodes are much more efficient for hair rendering. Implementation of BVH builder is based on Embree, and generally idea there is to calculate axis-aligned SAH and oriented SAH and if SAH of oriented node is smaller than axis-aligned SAH we create unaligned node. We store both aligned and unaligned nodes in the same tree (which seems to be different from what Embree is doing) so we don't have any any extra calculations needed to set up hair ray for BVH traversal, hence avoiding any possible negative effect of this new BVH nodes type. This new builder is currently not in use, still need to make BVH traversal code aware of unaligned nodes.
2016-07-07 10:18:57 +00:00
float nodeSAH,
const BVHUnaligned *unaligned_heuristic,
const Transform *aligned_space)
Cycles: Wrap spatial split storage into own structure This has following advantages: - Localizes all the run-time storage into a single structure, which could easily be extended further. - Storage could be created per-thread, so once builder is threaded we wouldn't have any conflicts between threads. - Global nature of the storage avoids memory re-allocation on the runtime, keeping builder as fast as possible. Currently it's just API changes, which don't affect user at all.
2016-02-21 14:39:02 +00:00
: sah(FLT_MAX),
dim(0),
num_left(0),
left_bounds(BoundBox::empty),
right_bounds(BoundBox::empty),
Cycles: Make spatial split BVH multi-threaded The title actually covers it all, This commit exploits all the work being done in previous changes to make it possible to build spatial splits in threads. Works quite nicely, but has a downside of some extra memory usage. In practice it doesn't seem to be a huge problem and that we can always look into later if it becomes a real showstopper. In practice it shows some nice speedup: - BMW27 scene takes 3 now (used to be 4) - Agent shot takes 5 sec (used to be 80) Such non-linear speedup is most likely coming from much less amount of heap re-allocations. A a downside, there's a bit of extra memory used by BVH arrays. From the tests amount of extra memory is below 0.001% so far, so it's not that bad at all. Reviewers: brecht, juicyfruit, dingto, lukasstockner97 Differential Revision: https://developer.blender.org/D1820
2016-04-04 12:43:21 +00:00
storage_(storage),
Cycles: Implement unaligned nodes BVH builder This is a special builder type which is allowed to orient nodes to strands direction, hence minimizing their surface area in comparison with axis-aligned nodes. Such nodes are much more efficient for hair rendering. Implementation of BVH builder is based on Embree, and generally idea there is to calculate axis-aligned SAH and oriented SAH and if SAH of oriented node is smaller than axis-aligned SAH we create unaligned node. We store both aligned and unaligned nodes in the same tree (which seems to be different from what Embree is doing) so we don't have any any extra calculations needed to set up hair ray for BVH traversal, hence avoiding any possible negative effect of this new BVH nodes type. This new builder is currently not in use, still need to make BVH traversal code aware of unaligned nodes.
2016-07-07 10:18:57 +00:00
references_(references),
unaligned_heuristic_(unaligned_heuristic),
aligned_space_(aligned_space)
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
{
Cycles: Make spatial split BVH multi-threaded The title actually covers it all, This commit exploits all the work being done in previous changes to make it possible to build spatial splits in threads. Works quite nicely, but has a downside of some extra memory usage. In practice it doesn't seem to be a huge problem and that we can always look into later if it becomes a real showstopper. In practice it shows some nice speedup: - BMW27 scene takes 3 now (used to be 4) - Agent shot takes 5 sec (used to be 80) Such non-linear speedup is most likely coming from much less amount of heap re-allocations. A a downside, there's a bit of extra memory used by BVH arrays. From the tests amount of extra memory is below 0.001% so far, so it's not that bad at all. Reviewers: brecht, juicyfruit, dingto, lukasstockner97 Differential Revision: https://developer.blender.org/D1820
2016-04-04 12:43:21 +00:00
const BVHReference *ref_ptr = &references_->at(range.start());
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
float min_sah = FLT_MAX;
Cycles: Move vector re-allocation out of loops
2016-04-25 10:25:30 +00:00
storage_->right_bounds.resize(range.size());
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
for(int dim = 0; dim < 3; dim++) {
Cycles: Temporarily revert index sort commit for spatial split There are in fact some missing parts to it (Split BVH builder should be creating bins from result of Object Split constructor). Doable, but need to quickly fix issue for the studio here, easier to revert for now.
2016-04-01 15:44:08 +00:00
/* Sort references. */
bvh_reference_sort(range.start(),
range.end(),
Cycles: Make spatial split BVH multi-threaded The title actually covers it all, This commit exploits all the work being done in previous changes to make it possible to build spatial splits in threads. Works quite nicely, but has a downside of some extra memory usage. In practice it doesn't seem to be a huge problem and that we can always look into later if it becomes a real showstopper. In practice it shows some nice speedup: - BMW27 scene takes 3 now (used to be 4) - Agent shot takes 5 sec (used to be 80) Such non-linear speedup is most likely coming from much less amount of heap re-allocations. A a downside, there's a bit of extra memory used by BVH arrays. From the tests amount of extra memory is below 0.001% so far, so it's not that bad at all. Reviewers: brecht, juicyfruit, dingto, lukasstockner97 Differential Revision: https://developer.blender.org/D1820
2016-04-04 12:43:21 +00:00
&references_->at(0),
Cycles: Implement unaligned nodes BVH builder This is a special builder type which is allowed to orient nodes to strands direction, hence minimizing their surface area in comparison with axis-aligned nodes. Such nodes are much more efficient for hair rendering. Implementation of BVH builder is based on Embree, and generally idea there is to calculate axis-aligned SAH and oriented SAH and if SAH of oriented node is smaller than axis-aligned SAH we create unaligned node. We store both aligned and unaligned nodes in the same tree (which seems to be different from what Embree is doing) so we don't have any any extra calculations needed to set up hair ray for BVH traversal, hence avoiding any possible negative effect of this new BVH nodes type. This new builder is currently not in use, still need to make BVH traversal code aware of unaligned nodes.
2016-07-07 10:18:57 +00:00
dim,
unaligned_heuristic_,
aligned_space_);
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
/* sweep right to left and determine bounds. */
BoundBox right_bounds = BoundBox::empty;
for(int i = range.size() - 1; i > 0; i--) {
Cycles: Implement unaligned nodes BVH builder This is a special builder type which is allowed to orient nodes to strands direction, hence minimizing their surface area in comparison with axis-aligned nodes. Such nodes are much more efficient for hair rendering. Implementation of BVH builder is based on Embree, and generally idea there is to calculate axis-aligned SAH and oriented SAH and if SAH of oriented node is smaller than axis-aligned SAH we create unaligned node. We store both aligned and unaligned nodes in the same tree (which seems to be different from what Embree is doing) so we don't have any any extra calculations needed to set up hair ray for BVH traversal, hence avoiding any possible negative effect of this new BVH nodes type. This new builder is currently not in use, still need to make BVH traversal code aware of unaligned nodes.
2016-07-07 10:18:57 +00:00
BoundBox prim_bounds = get_prim_bounds(ref_ptr[i]);
right_bounds.grow(prim_bounds);
Cycles: Name cleanup and some comments in BVH code
2016-02-22 15:33:42 +00:00
storage_->right_bounds[i - 1] = right_bounds;
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
}
/* sweep left to right and select lowest SAH. */
BoundBox left_bounds = BoundBox::empty;
for(int i = 1; i < range.size(); i++) {
Cycles: Implement unaligned nodes BVH builder This is a special builder type which is allowed to orient nodes to strands direction, hence minimizing their surface area in comparison with axis-aligned nodes. Such nodes are much more efficient for hair rendering. Implementation of BVH builder is based on Embree, and generally idea there is to calculate axis-aligned SAH and oriented SAH and if SAH of oriented node is smaller than axis-aligned SAH we create unaligned node. We store both aligned and unaligned nodes in the same tree (which seems to be different from what Embree is doing) so we don't have any any extra calculations needed to set up hair ray for BVH traversal, hence avoiding any possible negative effect of this new BVH nodes type. This new builder is currently not in use, still need to make BVH traversal code aware of unaligned nodes.
2016-07-07 10:18:57 +00:00
BoundBox prim_bounds = get_prim_bounds(ref_ptr[i - 1]);
left_bounds.grow(prim_bounds);
Cycles: Name cleanup and some comments in BVH code
2016-02-22 15:33:42 +00:00
right_bounds = storage_->right_bounds[i - 1];
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
float sah = nodeSAH +
Cycles: Temporarily revert index sort commit for spatial split There are in fact some missing parts to it (Split BVH builder should be creating bins from result of Object Split constructor). Doable, but need to quickly fix issue for the studio here, easier to revert for now.
2016-04-01 15:44:08 +00:00
left_bounds.safe_area() * builder->params.primitive_cost(i) +
right_bounds.safe_area() * builder->params.primitive_cost(range.size() - i);
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
if(sah < min_sah) {
min_sah = sah;
this->sah = sah;
this->dim = dim;
this->num_left = i;
this->left_bounds = left_bounds;
this->right_bounds = right_bounds;
}
}
}
}
Cycles: Make spatial split BVH multi-threaded The title actually covers it all, This commit exploits all the work being done in previous changes to make it possible to build spatial splits in threads. Works quite nicely, but has a downside of some extra memory usage. In practice it doesn't seem to be a huge problem and that we can always look into later if it becomes a real showstopper. In practice it shows some nice speedup: - BMW27 scene takes 3 now (used to be 4) - Agent shot takes 5 sec (used to be 80) Such non-linear speedup is most likely coming from much less amount of heap re-allocations. A a downside, there's a bit of extra memory used by BVH arrays. From the tests amount of extra memory is below 0.001% so far, so it's not that bad at all. Reviewers: brecht, juicyfruit, dingto, lukasstockner97 Differential Revision: https://developer.blender.org/D1820
2016-04-04 12:43:21 +00:00
void BVHObjectSplit::split(BVHRange& left,
BVHRange& right,
const BVHRange& range)
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
{
Cycles: Implement unaligned nodes BVH builder This is a special builder type which is allowed to orient nodes to strands direction, hence minimizing their surface area in comparison with axis-aligned nodes. Such nodes are much more efficient for hair rendering. Implementation of BVH builder is based on Embree, and generally idea there is to calculate axis-aligned SAH and oriented SAH and if SAH of oriented node is smaller than axis-aligned SAH we create unaligned node. We store both aligned and unaligned nodes in the same tree (which seems to be different from what Embree is doing) so we don't have any any extra calculations needed to set up hair ray for BVH traversal, hence avoiding any possible negative effect of this new BVH nodes type. This new builder is currently not in use, still need to make BVH traversal code aware of unaligned nodes.
2016-07-07 10:18:57 +00:00
assert(references_->size() > 0);
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
/* sort references according to split */
Cycles: Make spatial split BVH multi-threaded The title actually covers it all, This commit exploits all the work being done in previous changes to make it possible to build spatial splits in threads. Works quite nicely, but has a downside of some extra memory usage. In practice it doesn't seem to be a huge problem and that we can always look into later if it becomes a real showstopper. In practice it shows some nice speedup: - BMW27 scene takes 3 now (used to be 4) - Agent shot takes 5 sec (used to be 80) Such non-linear speedup is most likely coming from much less amount of heap re-allocations. A a downside, there's a bit of extra memory used by BVH arrays. From the tests amount of extra memory is below 0.001% so far, so it's not that bad at all. Reviewers: brecht, juicyfruit, dingto, lukasstockner97 Differential Revision: https://developer.blender.org/D1820
2016-04-04 12:43:21 +00:00
bvh_reference_sort(range.start(),
range.end(),
&references_->at(0),
Cycles: Implement unaligned nodes BVH builder This is a special builder type which is allowed to orient nodes to strands direction, hence minimizing their surface area in comparison with axis-aligned nodes. Such nodes are much more efficient for hair rendering. Implementation of BVH builder is based on Embree, and generally idea there is to calculate axis-aligned SAH and oriented SAH and if SAH of oriented node is smaller than axis-aligned SAH we create unaligned node. We store both aligned and unaligned nodes in the same tree (which seems to be different from what Embree is doing) so we don't have any any extra calculations needed to set up hair ray for BVH traversal, hence avoiding any possible negative effect of this new BVH nodes type. This new builder is currently not in use, still need to make BVH traversal code aware of unaligned nodes.
2016-07-07 10:18:57 +00:00
this->dim,
unaligned_heuristic_,
aligned_space_);
BoundBox effective_left_bounds, effective_right_bounds;
const int num_right = range.size() - this->num_left;
if(aligned_space_ == NULL) {
effective_left_bounds = left_bounds;
effective_right_bounds = right_bounds;
}
else {
effective_left_bounds = BoundBox::empty;
effective_right_bounds = BoundBox::empty;
for(int i = 0; i < this->num_left; ++i) {
BoundBox prim_boundbox = references_->at(range.start() + i).bounds();
effective_left_bounds.grow(prim_boundbox);
}
for(int i = 0; i < num_right; ++i) {
BoundBox prim_boundbox = references_->at(range.start() + this->num_left + i).bounds();
effective_right_bounds.grow(prim_boundbox);
}
}
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
/* split node ranges */
Cycles: Implement unaligned nodes BVH builder This is a special builder type which is allowed to orient nodes to strands direction, hence minimizing their surface area in comparison with axis-aligned nodes. Such nodes are much more efficient for hair rendering. Implementation of BVH builder is based on Embree, and generally idea there is to calculate axis-aligned SAH and oriented SAH and if SAH of oriented node is smaller than axis-aligned SAH we create unaligned node. We store both aligned and unaligned nodes in the same tree (which seems to be different from what Embree is doing) so we don't have any any extra calculations needed to set up hair ray for BVH traversal, hence avoiding any possible negative effect of this new BVH nodes type. This new builder is currently not in use, still need to make BVH traversal code aware of unaligned nodes.
2016-07-07 10:18:57 +00:00
left = BVHRange(effective_left_bounds, range.start(), this->num_left);
right = BVHRange(effective_right_bounds, left.end(), num_right);
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
}
/* Spatial Split */
Cycles: Pass BVH builder by const reference to spatial splitters
2016-02-21 15:36:57 +00:00
BVHSpatialSplit::BVHSpatialSplit(const BVHBuild& builder,
Cycles: Wrap spatial split storage into own structure This has following advantages: - Localizes all the run-time storage into a single structure, which could easily be extended further. - Storage could be created per-thread, so once builder is threaded we wouldn't have any conflicts between threads. - Global nature of the storage avoids memory re-allocation on the runtime, keeping builder as fast as possible. Currently it's just API changes, which don't affect user at all.
2016-02-21 14:39:02 +00:00
BVHSpatialStorage *storage,
const BVHRange& range,
Cycles: Make spatial split BVH multi-threaded The title actually covers it all, This commit exploits all the work being done in previous changes to make it possible to build spatial splits in threads. Works quite nicely, but has a downside of some extra memory usage. In practice it doesn't seem to be a huge problem and that we can always look into later if it becomes a real showstopper. In practice it shows some nice speedup: - BMW27 scene takes 3 now (used to be 4) - Agent shot takes 5 sec (used to be 80) Such non-linear speedup is most likely coming from much less amount of heap re-allocations. A a downside, there's a bit of extra memory used by BVH arrays. From the tests amount of extra memory is below 0.001% so far, so it's not that bad at all. Reviewers: brecht, juicyfruit, dingto, lukasstockner97 Differential Revision: https://developer.blender.org/D1820
2016-04-04 12:43:21 +00:00
vector<BVHReference> *references,
Cycles: Implement unaligned nodes BVH builder This is a special builder type which is allowed to orient nodes to strands direction, hence minimizing their surface area in comparison with axis-aligned nodes. Such nodes are much more efficient for hair rendering. Implementation of BVH builder is based on Embree, and generally idea there is to calculate axis-aligned SAH and oriented SAH and if SAH of oriented node is smaller than axis-aligned SAH we create unaligned node. We store both aligned and unaligned nodes in the same tree (which seems to be different from what Embree is doing) so we don't have any any extra calculations needed to set up hair ray for BVH traversal, hence avoiding any possible negative effect of this new BVH nodes type. This new builder is currently not in use, still need to make BVH traversal code aware of unaligned nodes.
2016-07-07 10:18:57 +00:00
float nodeSAH,
const BVHUnaligned *unaligned_heuristic,
const Transform *aligned_space)
Cycles: Wrap spatial split storage into own structure This has following advantages: - Localizes all the run-time storage into a single structure, which could easily be extended further. - Storage could be created per-thread, so once builder is threaded we wouldn't have any conflicts between threads. - Global nature of the storage avoids memory re-allocation on the runtime, keeping builder as fast as possible. Currently it's just API changes, which don't affect user at all.
2016-02-21 14:39:02 +00:00
: sah(FLT_MAX),
dim(0),
pos(0.0f),
Cycles: Make spatial split BVH multi-threaded The title actually covers it all, This commit exploits all the work being done in previous changes to make it possible to build spatial splits in threads. Works quite nicely, but has a downside of some extra memory usage. In practice it doesn't seem to be a huge problem and that we can always look into later if it becomes a real showstopper. In practice it shows some nice speedup: - BMW27 scene takes 3 now (used to be 4) - Agent shot takes 5 sec (used to be 80) Such non-linear speedup is most likely coming from much less amount of heap re-allocations. A a downside, there's a bit of extra memory used by BVH arrays. From the tests amount of extra memory is below 0.001% so far, so it's not that bad at all. Reviewers: brecht, juicyfruit, dingto, lukasstockner97 Differential Revision: https://developer.blender.org/D1820
2016-04-04 12:43:21 +00:00
storage_(storage),
Cycles: Implement unaligned nodes BVH builder This is a special builder type which is allowed to orient nodes to strands direction, hence minimizing their surface area in comparison with axis-aligned nodes. Such nodes are much more efficient for hair rendering. Implementation of BVH builder is based on Embree, and generally idea there is to calculate axis-aligned SAH and oriented SAH and if SAH of oriented node is smaller than axis-aligned SAH we create unaligned node. We store both aligned and unaligned nodes in the same tree (which seems to be different from what Embree is doing) so we don't have any any extra calculations needed to set up hair ray for BVH traversal, hence avoiding any possible negative effect of this new BVH nodes type. This new builder is currently not in use, still need to make BVH traversal code aware of unaligned nodes.
2016-07-07 10:18:57 +00:00
references_(references),
unaligned_heuristic_(unaligned_heuristic),
aligned_space_(aligned_space)
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
{
/* initialize bins. */
Cycles: Implement unaligned nodes BVH builder This is a special builder type which is allowed to orient nodes to strands direction, hence minimizing their surface area in comparison with axis-aligned nodes. Such nodes are much more efficient for hair rendering. Implementation of BVH builder is based on Embree, and generally idea there is to calculate axis-aligned SAH and oriented SAH and if SAH of oriented node is smaller than axis-aligned SAH we create unaligned node. We store both aligned and unaligned nodes in the same tree (which seems to be different from what Embree is doing) so we don't have any any extra calculations needed to set up hair ray for BVH traversal, hence avoiding any possible negative effect of this new BVH nodes type. This new builder is currently not in use, still need to make BVH traversal code aware of unaligned nodes.
2016-07-07 10:18:57 +00:00
BoundBox range_bounds;
if(aligned_space == NULL) {
range_bounds = range.bounds();
}
else {
range_bounds = unaligned_heuristic->compute_aligned_boundbox(
range,
&references->at(0),
*aligned_space);
}
float3 origin = range_bounds.min;
float3 binSize = (range_bounds.max - origin) * (1.0f / (float)BVHParams::NUM_SPATIAL_BINS);
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
float3 invBinSize = 1.0f / binSize;
for(int dim = 0; dim < 3; dim++) {
for(int i = 0; i < BVHParams::NUM_SPATIAL_BINS; i++) {
Cycles: Name cleanup and some comments in BVH code
2016-02-22 15:33:42 +00:00
BVHSpatialBin& bin = storage_->bins[dim][i];
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
bin.bounds = BoundBox::empty;
bin.enter = 0;
bin.exit = 0;
}
}
/* chop references into bins. */
for(unsigned int refIdx = range.start(); refIdx < range.end(); refIdx++) {
Cycles: Make spatial split BVH multi-threaded The title actually covers it all, This commit exploits all the work being done in previous changes to make it possible to build spatial splits in threads. Works quite nicely, but has a downside of some extra memory usage. In practice it doesn't seem to be a huge problem and that we can always look into later if it becomes a real showstopper. In practice it shows some nice speedup: - BMW27 scene takes 3 now (used to be 4) - Agent shot takes 5 sec (used to be 80) Such non-linear speedup is most likely coming from much less amount of heap re-allocations. A a downside, there's a bit of extra memory used by BVH arrays. From the tests amount of extra memory is below 0.001% so far, so it's not that bad at all. Reviewers: brecht, juicyfruit, dingto, lukasstockner97 Differential Revision: https://developer.blender.org/D1820
2016-04-04 12:43:21 +00:00
const BVHReference& ref = references_->at(refIdx);
Cycles: Implement unaligned nodes BVH builder This is a special builder type which is allowed to orient nodes to strands direction, hence minimizing their surface area in comparison with axis-aligned nodes. Such nodes are much more efficient for hair rendering. Implementation of BVH builder is based on Embree, and generally idea there is to calculate axis-aligned SAH and oriented SAH and if SAH of oriented node is smaller than axis-aligned SAH we create unaligned node. We store both aligned and unaligned nodes in the same tree (which seems to be different from what Embree is doing) so we don't have any any extra calculations needed to set up hair ray for BVH traversal, hence avoiding any possible negative effect of this new BVH nodes type. This new builder is currently not in use, still need to make BVH traversal code aware of unaligned nodes.
2016-07-07 10:18:57 +00:00
BoundBox prim_bounds = get_prim_bounds(ref);
float3 firstBinf = (prim_bounds.min - origin) * invBinSize;
float3 lastBinf = (prim_bounds.max - origin) * invBinSize;
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
int3 firstBin = make_int3((int)firstBinf.x, (int)firstBinf.y, (int)firstBinf.z);
int3 lastBin = make_int3((int)lastBinf.x, (int)lastBinf.y, (int)lastBinf.z);
firstBin = clamp(firstBin, 0, BVHParams::NUM_SPATIAL_BINS - 1);
lastBin = clamp(lastBin, firstBin, BVHParams::NUM_SPATIAL_BINS - 1);
for(int dim = 0; dim < 3; dim++) {
Cycles: Implement unaligned nodes BVH builder This is a special builder type which is allowed to orient nodes to strands direction, hence minimizing their surface area in comparison with axis-aligned nodes. Such nodes are much more efficient for hair rendering. Implementation of BVH builder is based on Embree, and generally idea there is to calculate axis-aligned SAH and oriented SAH and if SAH of oriented node is smaller than axis-aligned SAH we create unaligned node. We store both aligned and unaligned nodes in the same tree (which seems to be different from what Embree is doing) so we don't have any any extra calculations needed to set up hair ray for BVH traversal, hence avoiding any possible negative effect of this new BVH nodes type. This new builder is currently not in use, still need to make BVH traversal code aware of unaligned nodes.
2016-07-07 10:18:57 +00:00
BVHReference currRef(get_prim_bounds(ref),
ref.prim_index(),
ref.prim_object(),
ref.prim_type());
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
for(int i = firstBin[dim]; i < lastBin[dim]; i++) {
BVHReference leftRef, rightRef;
split_reference(builder, leftRef, rightRef, currRef, dim, origin[dim] + binSize[dim] * (float)(i + 1));
Cycles: Name cleanup and some comments in BVH code
2016-02-22 15:33:42 +00:00
storage_->bins[dim][i].bounds.grow(leftRef.bounds());
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
currRef = rightRef;
}
Cycles: Name cleanup and some comments in BVH code
2016-02-22 15:33:42 +00:00
storage_->bins[dim][lastBin[dim]].bounds.grow(currRef.bounds());
storage_->bins[dim][firstBin[dim]].enter++;
storage_->bins[dim][lastBin[dim]].exit++;
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
}
}
/* select best split plane. */
Cycles: Move vector re-allocation out of loops
2016-04-25 10:25:30 +00:00
storage_->right_bounds.resize(BVHParams::NUM_SPATIAL_BINS);
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
for(int dim = 0; dim < 3; dim++) {
/* sweep right to left and determine bounds. */
BoundBox right_bounds = BoundBox::empty;
for(int i = BVHParams::NUM_SPATIAL_BINS - 1; i > 0; i--) {
Cycles: Name cleanup and some comments in BVH code
2016-02-22 15:33:42 +00:00
right_bounds.grow(storage_->bins[dim][i].bounds);
storage_->right_bounds[i - 1] = right_bounds;
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
}
/* sweep left to right and select lowest SAH. */
BoundBox left_bounds = BoundBox::empty;
int leftNum = 0;
int rightNum = range.size();
for(int i = 1; i < BVHParams::NUM_SPATIAL_BINS; i++) {
Cycles: Name cleanup and some comments in BVH code
2016-02-22 15:33:42 +00:00
left_bounds.grow(storage_->bins[dim][i - 1].bounds);
leftNum += storage_->bins[dim][i - 1].enter;
rightNum -= storage_->bins[dim][i - 1].exit;
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
float sah = nodeSAH +
Cycles: Pass BVH builder by const reference to spatial splitters
2016-02-21 15:36:57 +00:00
left_bounds.safe_area() * builder.params.primitive_cost(leftNum) +
Cycles: Name cleanup and some comments in BVH code
2016-02-22 15:33:42 +00:00
storage_->right_bounds[i - 1].safe_area() * builder.params.primitive_cost(rightNum);
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
if(sah < this->sah) {
this->sah = sah;
this->dim = dim;
this->pos = origin[dim] + binSize[dim] * (float)i;
}
}
}
}
Cycles: Make spatial split BVH multi-threaded The title actually covers it all, This commit exploits all the work being done in previous changes to make it possible to build spatial splits in threads. Works quite nicely, but has a downside of some extra memory usage. In practice it doesn't seem to be a huge problem and that we can always look into later if it becomes a real showstopper. In practice it shows some nice speedup: - BMW27 scene takes 3 now (used to be 4) - Agent shot takes 5 sec (used to be 80) Such non-linear speedup is most likely coming from much less amount of heap re-allocations. A a downside, there's a bit of extra memory used by BVH arrays. From the tests amount of extra memory is below 0.001% so far, so it's not that bad at all. Reviewers: brecht, juicyfruit, dingto, lukasstockner97 Differential Revision: https://developer.blender.org/D1820
2016-04-04 12:43:21 +00:00
void BVHSpatialSplit::split(BVHBuild *builder,
BVHRange& left,
BVHRange& right,
const BVHRange& range)
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
{
/* Categorize references and compute bounds.
*
* Left-hand side: [left_start, left_end[
* Uncategorized/split: [left_end, right_start[
* Right-hand side: [right_start, refs.size()[ */
Cycles: Make spatial split BVH multi-threaded The title actually covers it all, This commit exploits all the work being done in previous changes to make it possible to build spatial splits in threads. Works quite nicely, but has a downside of some extra memory usage. In practice it doesn't seem to be a huge problem and that we can always look into later if it becomes a real showstopper. In practice it shows some nice speedup: - BMW27 scene takes 3 now (used to be 4) - Agent shot takes 5 sec (used to be 80) Such non-linear speedup is most likely coming from much less amount of heap re-allocations. A a downside, there's a bit of extra memory used by BVH arrays. From the tests amount of extra memory is below 0.001% so far, so it's not that bad at all. Reviewers: brecht, juicyfruit, dingto, lukasstockner97 Differential Revision: https://developer.blender.org/D1820
2016-04-04 12:43:21 +00:00
vector<BVHReference>& refs = *references_;
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
int left_start = range.start();
int left_end = left_start;
int right_start = range.end();
int right_end = range.end();
BoundBox left_bounds = BoundBox::empty;
BoundBox right_bounds = BoundBox::empty;
for(int i = left_end; i < right_start; i++) {
Cycles: Implement unaligned nodes BVH builder This is a special builder type which is allowed to orient nodes to strands direction, hence minimizing their surface area in comparison with axis-aligned nodes. Such nodes are much more efficient for hair rendering. Implementation of BVH builder is based on Embree, and generally idea there is to calculate axis-aligned SAH and oriented SAH and if SAH of oriented node is smaller than axis-aligned SAH we create unaligned node. We store both aligned and unaligned nodes in the same tree (which seems to be different from what Embree is doing) so we don't have any any extra calculations needed to set up hair ray for BVH traversal, hence avoiding any possible negative effect of this new BVH nodes type. This new builder is currently not in use, still need to make BVH traversal code aware of unaligned nodes.
2016-07-07 10:18:57 +00:00
BoundBox prim_bounds = get_prim_bounds(refs[i]);
if(prim_bounds.max[this->dim] <= this->pos) {
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
/* entirely on the left-hand side */
Cycles: Implement unaligned nodes BVH builder This is a special builder type which is allowed to orient nodes to strands direction, hence minimizing their surface area in comparison with axis-aligned nodes. Such nodes are much more efficient for hair rendering. Implementation of BVH builder is based on Embree, and generally idea there is to calculate axis-aligned SAH and oriented SAH and if SAH of oriented node is smaller than axis-aligned SAH we create unaligned node. We store both aligned and unaligned nodes in the same tree (which seems to be different from what Embree is doing) so we don't have any any extra calculations needed to set up hair ray for BVH traversal, hence avoiding any possible negative effect of this new BVH nodes type. This new builder is currently not in use, still need to make BVH traversal code aware of unaligned nodes.
2016-07-07 10:18:57 +00:00
left_bounds.grow(prim_bounds);
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
swap(refs[i], refs[left_end++]);
}
Cycles: Implement unaligned nodes BVH builder This is a special builder type which is allowed to orient nodes to strands direction, hence minimizing their surface area in comparison with axis-aligned nodes. Such nodes are much more efficient for hair rendering. Implementation of BVH builder is based on Embree, and generally idea there is to calculate axis-aligned SAH and oriented SAH and if SAH of oriented node is smaller than axis-aligned SAH we create unaligned node. We store both aligned and unaligned nodes in the same tree (which seems to be different from what Embree is doing) so we don't have any any extra calculations needed to set up hair ray for BVH traversal, hence avoiding any possible negative effect of this new BVH nodes type. This new builder is currently not in use, still need to make BVH traversal code aware of unaligned nodes.
2016-07-07 10:18:57 +00:00
else if(prim_bounds.min[this->dim] >= this->pos) {
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
/* entirely on the right-hand side */
Cycles: Implement unaligned nodes BVH builder This is a special builder type which is allowed to orient nodes to strands direction, hence minimizing their surface area in comparison with axis-aligned nodes. Such nodes are much more efficient for hair rendering. Implementation of BVH builder is based on Embree, and generally idea there is to calculate axis-aligned SAH and oriented SAH and if SAH of oriented node is smaller than axis-aligned SAH we create unaligned node. We store both aligned and unaligned nodes in the same tree (which seems to be different from what Embree is doing) so we don't have any any extra calculations needed to set up hair ray for BVH traversal, hence avoiding any possible negative effect of this new BVH nodes type. This new builder is currently not in use, still need to make BVH traversal code aware of unaligned nodes.
2016-07-07 10:18:57 +00:00
right_bounds.grow(prim_bounds);
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
swap(refs[i--], refs[--right_start]);
}
}
Cycles: Speedup of Spatial BVH split code Avoid memmove() happening on every insert of duplicated node to the references list. Temporary pre-allocated vector is used for new references which is then being inserted into actual array in one go later. Gives around 4x speedup building spatially split BVH for the grass field in the cassette player shot from Gooseberry.
2015-08-23 13:03:00 +00:00
/* Duplicate or unsplit references intersecting both sides.
*
* Duplication happens into a temporary pre-allocated vector in order to
* reduce number of memmove() calls happening in vector.insert().
*/
Cycles: Name cleanup and some comments in BVH code
2016-02-22 15:33:42 +00:00
vector<BVHReference>& new_refs = storage_->new_references;
Cycles: Avoid per-split memory allocation for the new references list
2016-02-21 15:46:16 +00:00
new_refs.clear();
Cycles: Speedup of Spatial BVH split code Avoid memmove() happening on every insert of duplicated node to the references list. Temporary pre-allocated vector is used for new references which is then being inserted into actual array in one go later. Gives around 4x speedup building spatially split BVH for the grass field in the cassette player shot from Gooseberry.
2015-08-23 13:03:00 +00:00
new_refs.reserve(right_start - left_end);
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
while(left_end < right_start) {
/* split reference. */
Cycles: Implement unaligned nodes BVH builder This is a special builder type which is allowed to orient nodes to strands direction, hence minimizing their surface area in comparison with axis-aligned nodes. Such nodes are much more efficient for hair rendering. Implementation of BVH builder is based on Embree, and generally idea there is to calculate axis-aligned SAH and oriented SAH and if SAH of oriented node is smaller than axis-aligned SAH we create unaligned node. We store both aligned and unaligned nodes in the same tree (which seems to be different from what Embree is doing) so we don't have any any extra calculations needed to set up hair ray for BVH traversal, hence avoiding any possible negative effect of this new BVH nodes type. This new builder is currently not in use, still need to make BVH traversal code aware of unaligned nodes.
2016-07-07 10:18:57 +00:00
BVHReference curr_ref(get_prim_bounds(refs[left_end]),
refs[left_end].prim_index(),
refs[left_end].prim_object(),
refs[left_end].prim_type());
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
BVHReference lref, rref;
Cycles: Implement unaligned nodes BVH builder This is a special builder type which is allowed to orient nodes to strands direction, hence minimizing their surface area in comparison with axis-aligned nodes. Such nodes are much more efficient for hair rendering. Implementation of BVH builder is based on Embree, and generally idea there is to calculate axis-aligned SAH and oriented SAH and if SAH of oriented node is smaller than axis-aligned SAH we create unaligned node. We store both aligned and unaligned nodes in the same tree (which seems to be different from what Embree is doing) so we don't have any any extra calculations needed to set up hair ray for BVH traversal, hence avoiding any possible negative effect of this new BVH nodes type. This new builder is currently not in use, still need to make BVH traversal code aware of unaligned nodes.
2016-07-07 10:18:57 +00:00
split_reference(*builder, lref, rref, curr_ref, this->dim, this->pos);
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
/* compute SAH for duplicate/unsplit candidates. */
BoundBox lub = left_bounds; // Unsplit to left: new left-hand bounds.
BoundBox rub = right_bounds; // Unsplit to right: new right-hand bounds.
BoundBox ldb = left_bounds; // Duplicate: new left-hand bounds.
BoundBox rdb = right_bounds; // Duplicate: new right-hand bounds.
Cycles: Implement unaligned nodes BVH builder This is a special builder type which is allowed to orient nodes to strands direction, hence minimizing their surface area in comparison with axis-aligned nodes. Such nodes are much more efficient for hair rendering. Implementation of BVH builder is based on Embree, and generally idea there is to calculate axis-aligned SAH and oriented SAH and if SAH of oriented node is smaller than axis-aligned SAH we create unaligned node. We store both aligned and unaligned nodes in the same tree (which seems to be different from what Embree is doing) so we don't have any any extra calculations needed to set up hair ray for BVH traversal, hence avoiding any possible negative effect of this new BVH nodes type. This new builder is currently not in use, still need to make BVH traversal code aware of unaligned nodes.
2016-07-07 10:18:57 +00:00
lub.grow(curr_ref.bounds());
rub.grow(curr_ref.bounds());
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
ldb.grow(lref.bounds());
rdb.grow(rref.bounds());
Cycles: optimization for hair BVH build, allow max 2 hair curves per leaf. This gives me 14% reduction in render time for koro_final.blend.
2014-04-22 14:54:02 +00:00
float lac = builder->params.primitive_cost(left_end - left_start);
float rac = builder->params.primitive_cost(right_end - right_start);
float lbc = builder->params.primitive_cost(left_end - left_start + 1);
float rbc = builder->params.primitive_cost(right_end - right_start + 1);
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
float unsplitLeftSAH = lub.safe_area() * lbc + right_bounds.safe_area() * rac;
float unsplitRightSAH = left_bounds.safe_area() * lac + rub.safe_area() * rbc;
float duplicateSAH = ldb.safe_area() * lbc + rdb.safe_area() * rbc;
float minSAH = min(min(unsplitLeftSAH, unsplitRightSAH), duplicateSAH);
if(minSAH == unsplitLeftSAH) {
/* unsplit to left */
left_bounds = lub;
left_end++;
}
else if(minSAH == unsplitRightSAH) {
/* unsplit to right */
right_bounds = rub;
swap(refs[left_end], refs[--right_start]);
}
else {
/* duplicate */
left_bounds = ldb;
right_bounds = rdb;
refs[left_end++] = lref;
Cycles: Speedup of Spatial BVH split code Avoid memmove() happening on every insert of duplicated node to the references list. Temporary pre-allocated vector is used for new references which is then being inserted into actual array in one go later. Gives around 4x speedup building spatially split BVH for the grass field in the cassette player shot from Gooseberry.
2015-08-23 13:03:00 +00:00
new_refs.push_back(rref);
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
right_end++;
}
}
Cycles: Speedup of Spatial BVH split code Avoid memmove() happening on every insert of duplicated node to the references list. Temporary pre-allocated vector is used for new references which is then being inserted into actual array in one go later. Gives around 4x speedup building spatially split BVH for the grass field in the cassette player shot from Gooseberry.
2015-08-23 13:03:00 +00:00
/* Insert duplicated references into actual array in one go. */
if(new_refs.size() != 0) {
Fix (harmless) assert in BVH spatial splits with Visual Studio debug builds.
2016-02-16 21:35:18 +00:00
refs.insert(refs.begin() + (right_end - new_refs.size()),
Cycles: Speedup of Spatial BVH split code Avoid memmove() happening on every insert of duplicated node to the references list. Temporary pre-allocated vector is used for new references which is then being inserted into actual array in one go later. Gives around 4x speedup building spatially split BVH for the grass field in the cassette player shot from Gooseberry.
2015-08-23 13:03:00 +00:00
new_refs.begin(),
new_refs.end());
}
Cycles: Implement unaligned nodes BVH builder This is a special builder type which is allowed to orient nodes to strands direction, hence minimizing their surface area in comparison with axis-aligned nodes. Such nodes are much more efficient for hair rendering. Implementation of BVH builder is based on Embree, and generally idea there is to calculate axis-aligned SAH and oriented SAH and if SAH of oriented node is smaller than axis-aligned SAH we create unaligned node. We store both aligned and unaligned nodes in the same tree (which seems to be different from what Embree is doing) so we don't have any any extra calculations needed to set up hair ray for BVH traversal, hence avoiding any possible negative effect of this new BVH nodes type. This new builder is currently not in use, still need to make BVH traversal code aware of unaligned nodes.
2016-07-07 10:18:57 +00:00
if(aligned_space_ != NULL) {
left_bounds = right_bounds = BoundBox::empty;
for(int i = left_start; i < left_end - left_start; ++i) {
BoundBox prim_boundbox = references_->at(i).bounds();
left_bounds.grow(prim_boundbox);
}
for(int i = right_start; i < right_end - right_start; ++i) {
BoundBox prim_boundbox = references_->at(i).bounds();
right_bounds.grow(prim_boundbox);
}
}
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
left = BVHRange(left_bounds, left_start, left_end - left_start);
right = BVHRange(right_bounds, right_start, right_end - right_start);
}
Cycles: Make primitive split code easier for re-use by reference splitting function
2015-08-21 17:40:14 +00:00
void BVHSpatialSplit::split_triangle_primitive(const Mesh *mesh,
Cycles: Implementation of object reference nodes spatial split This commit implements object reference node spatial split making it possible to use spatial split for top-level BVH. The code is not in use yet because enabling spatial split on top level BVH is not coming for free and it needs to be investigated if it's worth in terms of improved render times.
2015-08-22 19:19:00 +00:00
const Transform *tfm,
Cycles: Make primitive split code easier for re-use by reference splitting function
2015-08-21 17:40:14 +00:00
int prim_index,
Cycles: Move primitive splitting into own functions This way it's easy to add more reference types allowed for splitting to the BVH reference split function without making this function too much big. This way it's possible to experiment with such features as splitting object instance references. So far should not be any functional changes.
2015-08-21 15:49:41 +00:00
int dim,
float pos,
BoundBox& left_bounds,
BoundBox& right_bounds)
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
{
Code refactor: modify mesh storage to use arrays rather than vectors, separate some arrays. Differential Revision: https://developer.blender.org/D2016
2016-05-07 22:09:08 +00:00
Mesh::Triangle t = mesh->get_triangle(prim_index);
Cycles: Move primitive splitting into own functions This way it's easy to add more reference types allowed for splitting to the BVH reference split function without making this function too much big. This way it's possible to experiment with such features as splitting object instance references. So far should not be any functional changes.
2015-08-21 15:49:41 +00:00
const float3 *verts = &mesh->verts[0];
Code refactor: modify mesh storage to use arrays rather than vectors, separate some arrays. Differential Revision: https://developer.blender.org/D2016
2016-05-07 22:09:08 +00:00
float3 v1 = tfm ? transform_point(tfm, verts[t.v[2]]) : verts[t.v[2]];
Cycles: Implement unaligned nodes BVH builder This is a special builder type which is allowed to orient nodes to strands direction, hence minimizing their surface area in comparison with axis-aligned nodes. Such nodes are much more efficient for hair rendering. Implementation of BVH builder is based on Embree, and generally idea there is to calculate axis-aligned SAH and oriented SAH and if SAH of oriented node is smaller than axis-aligned SAH we create unaligned node. We store both aligned and unaligned nodes in the same tree (which seems to be different from what Embree is doing) so we don't have any any extra calculations needed to set up hair ray for BVH traversal, hence avoiding any possible negative effect of this new BVH nodes type. This new builder is currently not in use, still need to make BVH traversal code aware of unaligned nodes.
2016-07-07 10:18:57 +00:00
v1 = get_unaligned_point(v1);
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: Move primitive splitting into own functions This way it's easy to add more reference types allowed for splitting to the BVH reference split function without making this function too much big. This way it's possible to experiment with such features as splitting object instance references. So far should not be any functional changes.
2015-08-21 15:49:41 +00:00
for(int i = 0; i < 3; i++) {
Cycles: Implementation of object reference nodes spatial split This commit implements object reference node spatial split making it possible to use spatial split for top-level BVH. The code is not in use yet because enabling spatial split on top level BVH is not coming for free and it needs to be investigated if it's worth in terms of improved render times.
2015-08-22 19:19:00 +00:00
float3 v0 = v1;
Code refactor: modify mesh storage to use arrays rather than vectors, separate some arrays. Differential Revision: https://developer.blender.org/D2016
2016-05-07 22:09:08 +00:00
int vindex = t.v[i];
Cycles: Implementation of object reference nodes spatial split This commit implements object reference node spatial split making it possible to use spatial split for top-level BVH. The code is not in use yet because enabling spatial split on top level BVH is not coming for free and it needs to be investigated if it's worth in terms of improved render times.
2015-08-22 19:19:00 +00:00
v1 = tfm ? transform_point(tfm, verts[vindex]) : verts[vindex];
Cycles: Implement unaligned nodes BVH builder This is a special builder type which is allowed to orient nodes to strands direction, hence minimizing their surface area in comparison with axis-aligned nodes. Such nodes are much more efficient for hair rendering. Implementation of BVH builder is based on Embree, and generally idea there is to calculate axis-aligned SAH and oriented SAH and if SAH of oriented node is smaller than axis-aligned SAH we create unaligned node. We store both aligned and unaligned nodes in the same tree (which seems to be different from what Embree is doing) so we don't have any any extra calculations needed to set up hair ray for BVH traversal, hence avoiding any possible negative effect of this new BVH nodes type. This new builder is currently not in use, still need to make BVH traversal code aware of unaligned nodes.
2016-07-07 10:18:57 +00:00
v1 = get_unaligned_point(v1);
Cycles: Implementation of object reference nodes spatial split This commit implements object reference node spatial split making it possible to use spatial split for top-level BVH. The code is not in use yet because enabling spatial split on top level BVH is not coming for free and it needs to be investigated if it's worth in terms of improved render times.
2015-08-22 19:19:00 +00:00
float v0p = v0[dim];
float v1p = v1[dim];
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
/* insert vertex to the boxes it belongs to. */
if(v0p <= pos)
Cycles: Implementation of object reference nodes spatial split This commit implements object reference node spatial split making it possible to use spatial split for top-level BVH. The code is not in use yet because enabling spatial split on top level BVH is not coming for free and it needs to be investigated if it's worth in terms of improved render times.
2015-08-22 19:19:00 +00:00
left_bounds.grow(v0);
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
if(v0p >= pos)
Cycles: Implementation of object reference nodes spatial split This commit implements object reference node spatial split making it possible to use spatial split for top-level BVH. The code is not in use yet because enabling spatial split on top level BVH is not coming for free and it needs to be investigated if it's worth in terms of improved render times.
2015-08-22 19:19:00 +00:00
right_bounds.grow(v0);
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: 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
/* edge intersects the plane => insert intersection to both boxes. */
if((v0p < pos && v1p > pos) || (v0p > pos && v1p < pos)) {
Cycles: Implementation of object reference nodes spatial split This commit implements object reference node spatial split making it possible to use spatial split for top-level BVH. The code is not in use yet because enabling spatial split on top level BVH is not coming for free and it needs to be investigated if it's worth in terms of improved render times.
2015-08-22 19:19:00 +00:00
float3 t = lerp(v0, v1, clamp((pos - v0p) / (v1p - v0p), 0.0f, 1.0f));
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
left_bounds.grow(t);
right_bounds.grow(t);
}
}
Cycles: Move primitive splitting into own functions This way it's easy to add more reference types allowed for splitting to the BVH reference split function without making this function too much big. This way it's possible to experiment with such features as splitting object instance references. So far should not be any functional changes.
2015-08-21 15:49:41 +00:00
}
Cycles: Make primitive split code easier for re-use by reference splitting function
2015-08-21 17:40:14 +00:00
void BVHSpatialSplit::split_curve_primitive(const Mesh *mesh,
Cycles: Implementation of object reference nodes spatial split This commit implements object reference node spatial split making it possible to use spatial split for top-level BVH. The code is not in use yet because enabling spatial split on top level BVH is not coming for free and it needs to be investigated if it's worth in terms of improved render times.
2015-08-22 19:19:00 +00:00
const Transform *tfm,
Cycles: Make primitive split code easier for re-use by reference splitting function
2015-08-21 17:40:14 +00:00
int prim_index,
int segment_index,
Cycles: Move primitive splitting into own functions This way it's easy to add more reference types allowed for splitting to the BVH reference split function without making this function too much big. This way it's possible to experiment with such features as splitting object instance references. So far should not be any functional changes.
2015-08-21 15:49:41 +00:00
int dim,
float pos,
BoundBox& left_bounds,
BoundBox& right_bounds)
{
/* curve split: NOTE - Currently ignores curve width and needs to be fixed.*/
Code refactor: modify mesh storage to use arrays rather than vectors, separate some arrays. Differential Revision: https://developer.blender.org/D2016
2016-05-07 22:09:08 +00:00
Mesh::Curve curve = mesh->get_curve(prim_index);
const int k0 = curve.first_key + segment_index;
Cycles: Move primitive splitting into own functions This way it's easy to add more reference types allowed for splitting to the BVH reference split function without making this function too much big. This way it's possible to experiment with such features as splitting object instance references. So far should not be any functional changes.
2015-08-21 15:49:41 +00:00
const int k1 = k0 + 1;
Code refactor: modify mesh storage to use arrays rather than vectors, separate some arrays. Differential Revision: https://developer.blender.org/D2016
2016-05-07 22:09:08 +00:00
float3 v0 = mesh->curve_keys[k0];
float3 v1 = mesh->curve_keys[k1];
Cycles: Implementation of object reference nodes spatial split This commit implements object reference node spatial split making it possible to use spatial split for top-level BVH. The code is not in use yet because enabling spatial split on top level BVH is not coming for free and it needs to be investigated if it's worth in terms of improved render times.
2015-08-22 19:19:00 +00:00
if(tfm != NULL) {
v0 = transform_point(tfm, v0);
v1 = transform_point(tfm, v1);
}
Cycles: Implement unaligned nodes BVH builder This is a special builder type which is allowed to orient nodes to strands direction, hence minimizing their surface area in comparison with axis-aligned nodes. Such nodes are much more efficient for hair rendering. Implementation of BVH builder is based on Embree, and generally idea there is to calculate axis-aligned SAH and oriented SAH and if SAH of oriented node is smaller than axis-aligned SAH we create unaligned node. We store both aligned and unaligned nodes in the same tree (which seems to be different from what Embree is doing) so we don't have any any extra calculations needed to set up hair ray for BVH traversal, hence avoiding any possible negative effect of this new BVH nodes type. This new builder is currently not in use, still need to make BVH traversal code aware of unaligned nodes.
2016-07-07 10:18:57 +00:00
v0 = get_unaligned_point(v0);
v1 = get_unaligned_point(v1);
Cycles: Move primitive splitting into own functions This way it's easy to add more reference types allowed for splitting to the BVH reference split function without making this function too much big. This way it's possible to experiment with such features as splitting object instance references. So far should not be any functional changes.
2015-08-21 15:49:41 +00:00
float v0p = v0[dim];
float v1p = v1[dim];
/* insert vertex to the boxes it belongs to. */
if(v0p <= pos)
left_bounds.grow(v0);
if(v0p >= pos)
right_bounds.grow(v0);
if(v1p <= pos)
left_bounds.grow(v1);
if(v1p >= pos)
right_bounds.grow(v1);
/* edge intersects the plane => insert intersection to both boxes. */
if((v0p < pos && v1p > pos) || (v0p > pos && v1p < pos)) {
float3 t = lerp(v0, v1, clamp((pos - v0p) / (v1p - v0p), 0.0f, 1.0f));
left_bounds.grow(t);
right_bounds.grow(t);
}
}
Cycles: Make primitive split code easier for re-use by reference splitting function
2015-08-21 17:40:14 +00:00
void BVHSpatialSplit::split_triangle_reference(const BVHReference& ref,
const Mesh *mesh,
int dim,
float pos,
BoundBox& left_bounds,
BoundBox& right_bounds)
{
split_triangle_primitive(mesh,
Cycles: Implementation of object reference nodes spatial split This commit implements object reference node spatial split making it possible to use spatial split for top-level BVH. The code is not in use yet because enabling spatial split on top level BVH is not coming for free and it needs to be investigated if it's worth in terms of improved render times.
2015-08-22 19:19:00 +00:00
NULL,
Cycles: Make primitive split code easier for re-use by reference splitting function
2015-08-21 17:40:14 +00:00
ref.prim_index(),
dim,
pos,
left_bounds,
right_bounds);
}
void BVHSpatialSplit::split_curve_reference(const BVHReference& ref,
const Mesh *mesh,
int dim,
float pos,
BoundBox& left_bounds,
BoundBox& right_bounds)
{
split_curve_primitive(mesh,
Cycles: Implementation of object reference nodes spatial split This commit implements object reference node spatial split making it possible to use spatial split for top-level BVH. The code is not in use yet because enabling spatial split on top level BVH is not coming for free and it needs to be investigated if it's worth in terms of improved render times.
2015-08-22 19:19:00 +00:00
NULL,
Cycles: Make primitive split code easier for re-use by reference splitting function
2015-08-21 17:40:14 +00:00
ref.prim_index(),
PRIMITIVE_UNPACK_SEGMENT(ref.prim_type()),
dim,
pos,
left_bounds,
right_bounds);
}
Cycles: Implementation of object reference nodes spatial split This commit implements object reference node spatial split making it possible to use spatial split for top-level BVH. The code is not in use yet because enabling spatial split on top level BVH is not coming for free and it needs to be investigated if it's worth in terms of improved render times.
2015-08-22 19:19:00 +00:00
void BVHSpatialSplit::split_object_reference(const Object *object,
int dim,
float pos,
BoundBox& left_bounds,
BoundBox& right_bounds)
{
Mesh *mesh = object->mesh;
Cycles: Fix regression introduced in c96a4c8 A few places still needed to be updated to use the new Mesh::num_triangles() method; wrong number from triangles.size() was causing crashes.
2016-06-03 00:57:04 +00:00
for(int tri_idx = 0; tri_idx < mesh->num_triangles(); ++tri_idx) {
Cycles: Implementation of object reference nodes spatial split This commit implements object reference node spatial split making it possible to use spatial split for top-level BVH. The code is not in use yet because enabling spatial split on top level BVH is not coming for free and it needs to be investigated if it's worth in terms of improved render times.
2015-08-22 19:19:00 +00:00
split_triangle_primitive(mesh,
&object->tfm,
tri_idx,
dim,
pos,
left_bounds,
right_bounds);
}
Code refactor: modify mesh storage to use arrays rather than vectors, separate some arrays. Differential Revision: https://developer.blender.org/D2016
2016-05-07 22:09:08 +00:00
for(int curve_idx = 0; curve_idx < mesh->num_curves(); ++curve_idx) {
Mesh::Curve curve = mesh->get_curve(curve_idx);
Cycles: Implementation of object reference nodes spatial split This commit implements object reference node spatial split making it possible to use spatial split for top-level BVH. The code is not in use yet because enabling spatial split on top level BVH is not coming for free and it needs to be investigated if it's worth in terms of improved render times.
2015-08-22 19:19:00 +00:00
for(int segment_idx = 0;
segment_idx < curve.num_keys - 1;
++segment_idx)
{
split_curve_primitive(mesh,
&object->tfm,
curve_idx,
segment_idx,
dim,
pos,
left_bounds,
right_bounds);
}
}
}
Cycles: Pass BVH builder by const reference to spatial splitters
2016-02-21 15:36:57 +00:00
void BVHSpatialSplit::split_reference(const BVHBuild& builder,
Cycles: Move primitive splitting into own functions This way it's easy to add more reference types allowed for splitting to the BVH reference split function without making this function too much big. This way it's possible to experiment with such features as splitting object instance references. So far should not be any functional changes.
2015-08-21 15:49:41 +00:00
BVHReference& left,
BVHReference& right,
const BVHReference& ref,
int dim,
float pos)
{
/* initialize boundboxes */
BoundBox left_bounds = BoundBox::empty;
BoundBox right_bounds = BoundBox::empty;
/* loop over vertices/edges. */
Cycles: Pass BVH builder by const reference to spatial splitters
2016-02-21 15:36:57 +00:00
const Object *ob = builder.objects[ref.prim_object()];
Cycles: Move primitive splitting into own functions This way it's easy to add more reference types allowed for splitting to the BVH reference split function without making this function too much big. This way it's possible to experiment with such features as splitting object instance references. So far should not be any functional changes.
2015-08-21 15:49:41 +00:00
const Mesh *mesh = ob->mesh;
if(ref.prim_type() & PRIMITIVE_ALL_TRIANGLE) {
split_triangle_reference(ref,
mesh,
dim,
pos,
left_bounds,
right_bounds);
}
else if(ref.prim_type() & PRIMITIVE_ALL_CURVE) {
split_curve_reference(ref,
mesh,
dim,
pos,
left_bounds,
right_bounds);
}
else {
Cycles: Implementation of object reference nodes spatial split This commit implements object reference node spatial split making it possible to use spatial split for top-level BVH. The code is not in use yet because enabling spatial split on top level BVH is not coming for free and it needs to be investigated if it's worth in terms of improved render times.
2015-08-22 19:19:00 +00:00
split_object_reference(ob,
dim,
pos,
left_bounds,
right_bounds);
Cycles: Move primitive splitting into own functions This way it's easy to add more reference types allowed for splitting to the BVH reference split function without making this function too much big. This way it's possible to experiment with such features as splitting object instance references. So far should not be any functional changes.
2015-08-21 15:49:41 +00:00
}
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
/* intersect with original bounds. */
left_bounds.max[dim] = pos;
right_bounds.min[dim] = pos;
Cycles: Implement unaligned nodes BVH builder This is a special builder type which is allowed to orient nodes to strands direction, hence minimizing their surface area in comparison with axis-aligned nodes. Such nodes are much more efficient for hair rendering. Implementation of BVH builder is based on Embree, and generally idea there is to calculate axis-aligned SAH and oriented SAH and if SAH of oriented node is smaller than axis-aligned SAH we create unaligned node. We store both aligned and unaligned nodes in the same tree (which seems to be different from what Embree is doing) so we don't have any any extra calculations needed to set up hair ray for BVH traversal, hence avoiding any possible negative effect of this new BVH nodes type. This new builder is currently not in use, still need to make BVH traversal code aware of unaligned nodes.
2016-07-07 10:18:57 +00:00
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
left_bounds.intersect(ref.bounds());
right_bounds.intersect(ref.bounds());
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
/* set references */
Cycles code refactor: changes to make adding new primitive types easier.
2014-03-29 12:03:46 +00:00
left = BVHReference(left_bounds, ref.prim_index(), ref.prim_object(), ref.prim_type());
right = BVHReference(right_bounds, ref.prim_index(), ref.prim_object(), ref.prim_type());
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
}
CCL_NAMESPACE_END
Copy Permalink