Cycles: OpenCL kernel split
This commit contains all the work related on the AMD megakernel split work
which was mainly done by Varun Sundar, George Kyriazis and Lenny Wang, plus
some help from Sergey Sharybin, Martijn Berger, Thomas Dinges and likely
someone else which we're forgetting to mention.
Currently only AMD cards are enabled for the new split kernel, but it is
possible to force split opencl kernel to be used by setting the following
environment variable: CYCLES_OPENCL_SPLIT_KERNEL_TEST=1.
Not all the features are supported yet, and that being said no motion blur,
camera blur, SSS and volumetrics for now. Also transparent shadows are
disabled on AMD device because of some compiler bug.
This kernel is also only implements regular path tracing and supporting
branched one will take a bit. Branched path tracing is exposed to the
interface still, which is a bit misleading and will be hidden there soon.
More feature will be enabled once they're ported to the split kernel and
tested.
Neither regular CPU nor CUDA has any difference, they're generating the
same exact code, which means no regressions/improvements there.
Based on the research paper:
https://research.nvidia.com/sites/default/files/publications/laine2013hpg_paper.pdf
Here's the documentation:
https://docs.google.com/document/d/1LuXW-CV-sVJkQaEGZlMJ86jZ8FmoPfecaMdR-oiWbUY/edit
Design discussion of the patch:
https://developer.blender.org/T44197
Differential Revision: https://developer.blender.org/D1200
2015-05-09 14:34:30 +00:00
|
|
|
/*
|
|
|
|
* Copyright 2011-2015 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.
|
|
|
|
*/
|
|
|
|
|
2017-02-22 13:10:02 +00:00
|
|
|
CCL_NAMESPACE_BEGIN
|
Cycles: OpenCL kernel split
This commit contains all the work related on the AMD megakernel split work
which was mainly done by Varun Sundar, George Kyriazis and Lenny Wang, plus
some help from Sergey Sharybin, Martijn Berger, Thomas Dinges and likely
someone else which we're forgetting to mention.
Currently only AMD cards are enabled for the new split kernel, but it is
possible to force split opencl kernel to be used by setting the following
environment variable: CYCLES_OPENCL_SPLIT_KERNEL_TEST=1.
Not all the features are supported yet, and that being said no motion blur,
camera blur, SSS and volumetrics for now. Also transparent shadows are
disabled on AMD device because of some compiler bug.
This kernel is also only implements regular path tracing and supporting
branched one will take a bit. Branched path tracing is exposed to the
interface still, which is a bit misleading and will be hidden there soon.
More feature will be enabled once they're ported to the split kernel and
tested.
Neither regular CPU nor CUDA has any difference, they're generating the
same exact code, which means no regressions/improvements there.
Based on the research paper:
https://research.nvidia.com/sites/default/files/publications/laine2013hpg_paper.pdf
Here's the documentation:
https://docs.google.com/document/d/1LuXW-CV-sVJkQaEGZlMJ86jZ8FmoPfecaMdR-oiWbUY/edit
Design discussion of the patch:
https://developer.blender.org/T44197
Differential Revision: https://developer.blender.org/D1200
2015-05-09 14:34:30 +00:00
|
|
|
|
2015-05-26 14:12:49 +00:00
|
|
|
/* Note on kernel_scene_intersect kernel.
|
Cycles: OpenCL kernel split
This commit contains all the work related on the AMD megakernel split work
which was mainly done by Varun Sundar, George Kyriazis and Lenny Wang, plus
some help from Sergey Sharybin, Martijn Berger, Thomas Dinges and likely
someone else which we're forgetting to mention.
Currently only AMD cards are enabled for the new split kernel, but it is
possible to force split opencl kernel to be used by setting the following
environment variable: CYCLES_OPENCL_SPLIT_KERNEL_TEST=1.
Not all the features are supported yet, and that being said no motion blur,
camera blur, SSS and volumetrics for now. Also transparent shadows are
disabled on AMD device because of some compiler bug.
This kernel is also only implements regular path tracing and supporting
branched one will take a bit. Branched path tracing is exposed to the
interface still, which is a bit misleading and will be hidden there soon.
More feature will be enabled once they're ported to the split kernel and
tested.
Neither regular CPU nor CUDA has any difference, they're generating the
same exact code, which means no regressions/improvements there.
Based on the research paper:
https://research.nvidia.com/sites/default/files/publications/laine2013hpg_paper.pdf
Here's the documentation:
https://docs.google.com/document/d/1LuXW-CV-sVJkQaEGZlMJ86jZ8FmoPfecaMdR-oiWbUY/edit
Design discussion of the patch:
https://developer.blender.org/T44197
Differential Revision: https://developer.blender.org/D1200
2015-05-09 14:34:30 +00:00
|
|
|
* This is the second kernel in the ray tracing logic. This is the first
|
|
|
|
* of the path iteration kernels. This kernel takes care of scene_intersect function.
|
|
|
|
*
|
|
|
|
* This kernel changes the ray_state of RAY_REGENERATED rays to RAY_ACTIVE.
|
|
|
|
* This kernel processes rays of ray state RAY_ACTIVE
|
|
|
|
* This kernel determines the rays that have hit the background and changes their ray state to RAY_HIT_BACKGROUND.
|
|
|
|
*
|
|
|
|
* The input and output are as follows,
|
|
|
|
*
|
2015-05-21 12:40:04 +00:00
|
|
|
* Ray_coop ---------------------------------------|--------- kernel_scene_intersect----------|--- PathState
|
|
|
|
* PathState_coop ---------------------------------| |--- Intersection
|
|
|
|
* ray_state --------------------------------------| |--- ray_state
|
|
|
|
* use_queues_flag --------------------------------| |
|
|
|
|
* QueueData(QUEUE_ACTIVE_AND_REGENERATED_RAYS) ---| |
|
2015-11-01 11:51:32 +00:00
|
|
|
* kg (globals) -----------------------------------| |
|
2015-05-21 12:40:04 +00:00
|
|
|
* rng_coop ---------------------------------------| |
|
|
|
|
* sw ---------------------------------------------| |
|
|
|
|
* sh ---------------------------------------------| |
|
|
|
|
* queuesize --------------------------------------| |
|
Cycles: OpenCL kernel split
This commit contains all the work related on the AMD megakernel split work
which was mainly done by Varun Sundar, George Kyriazis and Lenny Wang, plus
some help from Sergey Sharybin, Martijn Berger, Thomas Dinges and likely
someone else which we're forgetting to mention.
Currently only AMD cards are enabled for the new split kernel, but it is
possible to force split opencl kernel to be used by setting the following
environment variable: CYCLES_OPENCL_SPLIT_KERNEL_TEST=1.
Not all the features are supported yet, and that being said no motion blur,
camera blur, SSS and volumetrics for now. Also transparent shadows are
disabled on AMD device because of some compiler bug.
This kernel is also only implements regular path tracing and supporting
branched one will take a bit. Branched path tracing is exposed to the
interface still, which is a bit misleading and will be hidden there soon.
More feature will be enabled once they're ported to the split kernel and
tested.
Neither regular CPU nor CUDA has any difference, they're generating the
same exact code, which means no regressions/improvements there.
Based on the research paper:
https://research.nvidia.com/sites/default/files/publications/laine2013hpg_paper.pdf
Here's the documentation:
https://docs.google.com/document/d/1LuXW-CV-sVJkQaEGZlMJ86jZ8FmoPfecaMdR-oiWbUY/edit
Design discussion of the patch:
https://developer.blender.org/T44197
Differential Revision: https://developer.blender.org/D1200
2015-05-09 14:34:30 +00:00
|
|
|
*
|
|
|
|
* Note on Queues :
|
2015-05-21 12:40:04 +00:00
|
|
|
* Ideally we would want kernel_scene_intersect to work on queues.
|
2015-05-26 06:46:11 +00:00
|
|
|
* But during the very first time, the queues will be empty and hence we perform a direct mapping
|
Cycles: OpenCL kernel split
This commit contains all the work related on the AMD megakernel split work
which was mainly done by Varun Sundar, George Kyriazis and Lenny Wang, plus
some help from Sergey Sharybin, Martijn Berger, Thomas Dinges and likely
someone else which we're forgetting to mention.
Currently only AMD cards are enabled for the new split kernel, but it is
possible to force split opencl kernel to be used by setting the following
environment variable: CYCLES_OPENCL_SPLIT_KERNEL_TEST=1.
Not all the features are supported yet, and that being said no motion blur,
camera blur, SSS and volumetrics for now. Also transparent shadows are
disabled on AMD device because of some compiler bug.
This kernel is also only implements regular path tracing and supporting
branched one will take a bit. Branched path tracing is exposed to the
interface still, which is a bit misleading and will be hidden there soon.
More feature will be enabled once they're ported to the split kernel and
tested.
Neither regular CPU nor CUDA has any difference, they're generating the
same exact code, which means no regressions/improvements there.
Based on the research paper:
https://research.nvidia.com/sites/default/files/publications/laine2013hpg_paper.pdf
Here's the documentation:
https://docs.google.com/document/d/1LuXW-CV-sVJkQaEGZlMJ86jZ8FmoPfecaMdR-oiWbUY/edit
Design discussion of the patch:
https://developer.blender.org/T44197
Differential Revision: https://developer.blender.org/D1200
2015-05-09 14:34:30 +00:00
|
|
|
* between ray-index and thread-index; From the next time onward, the queue will be filled and
|
|
|
|
* we may start operating on queues.
|
|
|
|
*
|
|
|
|
* State of queue during the first time this kernel is called :
|
|
|
|
* QUEUE_ACTIVE_AND_REGENERATED_RAYS and QUEUE_HITBG_BUFF_UPDATE_TOREGEN_RAYS will be empty.before and after this kernel
|
|
|
|
*
|
|
|
|
* State of queues during other times this kernel is called :
|
|
|
|
* At entry,
|
|
|
|
* QUEUE_ACTIVE_AND_REGENERATED_RAYS will have a mix of RAY_ACTIVE, RAY_UPDATE_BUFFER and RAY_REGENERATED rays;
|
|
|
|
* QUEUE_HITBG_BUFF_UPDATE_TOREGEN_RAYS will be filled with RAY_TO_REGENERATE and RAY_UPDATE_BUFFER rays ;
|
|
|
|
* (The rays that are in the state RAY_UPDATE_BUFFER in both the queues are actually the same rays; These
|
|
|
|
* are the rays that were in RAY_ACTIVE state during the initial enqueue but on further processing
|
|
|
|
* , by different kernels, have turned into RAY_UPDATE_BUFFER rays. Since all kernel, even after fetching from
|
|
|
|
* QUEUE_ACTIVE_AND_REGENERATED_RAYS, proceed further based on ray state information, RAY_UPDATE_BUFFER rays
|
|
|
|
* being present in QUEUE_ACTIVE_AND_REGENERATED_RAYS does not cause any logical issues)
|
|
|
|
* At exit,
|
|
|
|
* QUEUE_ACTIVE_AND_REGENERATED_RAYS - All RAY_REGENERATED rays will have been converted to RAY_ACTIVE and
|
|
|
|
* Some rays in QUEUE_ACTIVE_AND_REGENERATED_RAYS queue will move to state RAY_HIT_BACKGROUND
|
|
|
|
* QUEUE_HITBF_BUFF_UPDATE_TOREGEN_RAYS - no change
|
|
|
|
*/
|
|
|
|
|
2017-02-22 13:10:02 +00:00
|
|
|
ccl_device void kernel_scene_intersect(KernelGlobals *kg)
|
Cycles: OpenCL kernel split
This commit contains all the work related on the AMD megakernel split work
which was mainly done by Varun Sundar, George Kyriazis and Lenny Wang, plus
some help from Sergey Sharybin, Martijn Berger, Thomas Dinges and likely
someone else which we're forgetting to mention.
Currently only AMD cards are enabled for the new split kernel, but it is
possible to force split opencl kernel to be used by setting the following
environment variable: CYCLES_OPENCL_SPLIT_KERNEL_TEST=1.
Not all the features are supported yet, and that being said no motion blur,
camera blur, SSS and volumetrics for now. Also transparent shadows are
disabled on AMD device because of some compiler bug.
This kernel is also only implements regular path tracing and supporting
branched one will take a bit. Branched path tracing is exposed to the
interface still, which is a bit misleading and will be hidden there soon.
More feature will be enabled once they're ported to the split kernel and
tested.
Neither regular CPU nor CUDA has any difference, they're generating the
same exact code, which means no regressions/improvements there.
Based on the research paper:
https://research.nvidia.com/sites/default/files/publications/laine2013hpg_paper.pdf
Here's the documentation:
https://docs.google.com/document/d/1LuXW-CV-sVJkQaEGZlMJ86jZ8FmoPfecaMdR-oiWbUY/edit
Design discussion of the patch:
https://developer.blender.org/T44197
Differential Revision: https://developer.blender.org/D1200
2015-05-09 14:34:30 +00:00
|
|
|
{
|
2017-02-22 13:10:02 +00:00
|
|
|
/* Fetch use_queues_flag */
|
|
|
|
ccl_local char local_use_queues_flag;
|
|
|
|
if(ccl_local_id(0) == 0 && ccl_local_id(1) == 0) {
|
|
|
|
local_use_queues_flag = *kernel_split_params.use_queues_flag;
|
|
|
|
}
|
|
|
|
ccl_barrier(CCL_LOCAL_MEM_FENCE);
|
|
|
|
|
|
|
|
int ray_index = ccl_global_id(1) * ccl_global_size(0) + ccl_global_id(0);
|
|
|
|
if(local_use_queues_flag) {
|
|
|
|
ray_index = get_ray_index(kg, ray_index,
|
|
|
|
QUEUE_ACTIVE_AND_REGENERATED_RAYS,
|
|
|
|
kernel_split_state.queue_data,
|
|
|
|
kernel_split_params.queue_size,
|
|
|
|
0);
|
|
|
|
|
|
|
|
if(ray_index == QUEUE_EMPTY_SLOT) {
|
|
|
|
return;
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
Cycles: OpenCL kernel split
This commit contains all the work related on the AMD megakernel split work
which was mainly done by Varun Sundar, George Kyriazis and Lenny Wang, plus
some help from Sergey Sharybin, Martijn Berger, Thomas Dinges and likely
someone else which we're forgetting to mention.
Currently only AMD cards are enabled for the new split kernel, but it is
possible to force split opencl kernel to be used by setting the following
environment variable: CYCLES_OPENCL_SPLIT_KERNEL_TEST=1.
Not all the features are supported yet, and that being said no motion blur,
camera blur, SSS and volumetrics for now. Also transparent shadows are
disabled on AMD device because of some compiler bug.
This kernel is also only implements regular path tracing and supporting
branched one will take a bit. Branched path tracing is exposed to the
interface still, which is a bit misleading and will be hidden there soon.
More feature will be enabled once they're ported to the split kernel and
tested.
Neither regular CPU nor CUDA has any difference, they're generating the
same exact code, which means no regressions/improvements there.
Based on the research paper:
https://research.nvidia.com/sites/default/files/publications/laine2013hpg_paper.pdf
Here's the documentation:
https://docs.google.com/document/d/1LuXW-CV-sVJkQaEGZlMJ86jZ8FmoPfecaMdR-oiWbUY/edit
Design discussion of the patch:
https://developer.blender.org/T44197
Differential Revision: https://developer.blender.org/D1200
2015-05-09 14:34:30 +00:00
|
|
|
/* All regenerated rays become active here */
|
2017-02-22 13:10:02 +00:00
|
|
|
if(IS_STATE(kernel_split_state.ray_state, ray_index, RAY_REGENERATED))
|
|
|
|
ASSIGN_RAY_STATE(kernel_split_state.ray_state, ray_index, RAY_ACTIVE);
|
Cycles: OpenCL kernel split
This commit contains all the work related on the AMD megakernel split work
which was mainly done by Varun Sundar, George Kyriazis and Lenny Wang, plus
some help from Sergey Sharybin, Martijn Berger, Thomas Dinges and likely
someone else which we're forgetting to mention.
Currently only AMD cards are enabled for the new split kernel, but it is
possible to force split opencl kernel to be used by setting the following
environment variable: CYCLES_OPENCL_SPLIT_KERNEL_TEST=1.
Not all the features are supported yet, and that being said no motion blur,
camera blur, SSS and volumetrics for now. Also transparent shadows are
disabled on AMD device because of some compiler bug.
This kernel is also only implements regular path tracing and supporting
branched one will take a bit. Branched path tracing is exposed to the
interface still, which is a bit misleading and will be hidden there soon.
More feature will be enabled once they're ported to the split kernel and
tested.
Neither regular CPU nor CUDA has any difference, they're generating the
same exact code, which means no regressions/improvements there.
Based on the research paper:
https://research.nvidia.com/sites/default/files/publications/laine2013hpg_paper.pdf
Here's the documentation:
https://docs.google.com/document/d/1LuXW-CV-sVJkQaEGZlMJ86jZ8FmoPfecaMdR-oiWbUY/edit
Design discussion of the patch:
https://developer.blender.org/T44197
Differential Revision: https://developer.blender.org/D1200
2015-05-09 14:34:30 +00:00
|
|
|
|
2017-02-22 13:10:02 +00:00
|
|
|
if(!IS_STATE(kernel_split_state.ray_state, ray_index, RAY_ACTIVE))
|
Cycles: OpenCL kernel split
This commit contains all the work related on the AMD megakernel split work
which was mainly done by Varun Sundar, George Kyriazis and Lenny Wang, plus
some help from Sergey Sharybin, Martijn Berger, Thomas Dinges and likely
someone else which we're forgetting to mention.
Currently only AMD cards are enabled for the new split kernel, but it is
possible to force split opencl kernel to be used by setting the following
environment variable: CYCLES_OPENCL_SPLIT_KERNEL_TEST=1.
Not all the features are supported yet, and that being said no motion blur,
camera blur, SSS and volumetrics for now. Also transparent shadows are
disabled on AMD device because of some compiler bug.
This kernel is also only implements regular path tracing and supporting
branched one will take a bit. Branched path tracing is exposed to the
interface still, which is a bit misleading and will be hidden there soon.
More feature will be enabled once they're ported to the split kernel and
tested.
Neither regular CPU nor CUDA has any difference, they're generating the
same exact code, which means no regressions/improvements there.
Based on the research paper:
https://research.nvidia.com/sites/default/files/publications/laine2013hpg_paper.pdf
Here's the documentation:
https://docs.google.com/document/d/1LuXW-CV-sVJkQaEGZlMJ86jZ8FmoPfecaMdR-oiWbUY/edit
Design discussion of the patch:
https://developer.blender.org/T44197
Differential Revision: https://developer.blender.org/D1200
2015-05-09 14:34:30 +00:00
|
|
|
return;
|
|
|
|
|
|
|
|
#ifdef __KERNEL_DEBUG__
|
2017-02-22 13:10:02 +00:00
|
|
|
DebugData *debug_data = &kernel_split_state.debug_data[ray_index];
|
Cycles: OpenCL kernel split
This commit contains all the work related on the AMD megakernel split work
which was mainly done by Varun Sundar, George Kyriazis and Lenny Wang, plus
some help from Sergey Sharybin, Martijn Berger, Thomas Dinges and likely
someone else which we're forgetting to mention.
Currently only AMD cards are enabled for the new split kernel, but it is
possible to force split opencl kernel to be used by setting the following
environment variable: CYCLES_OPENCL_SPLIT_KERNEL_TEST=1.
Not all the features are supported yet, and that being said no motion blur,
camera blur, SSS and volumetrics for now. Also transparent shadows are
disabled on AMD device because of some compiler bug.
This kernel is also only implements regular path tracing and supporting
branched one will take a bit. Branched path tracing is exposed to the
interface still, which is a bit misleading and will be hidden there soon.
More feature will be enabled once they're ported to the split kernel and
tested.
Neither regular CPU nor CUDA has any difference, they're generating the
same exact code, which means no regressions/improvements there.
Based on the research paper:
https://research.nvidia.com/sites/default/files/publications/laine2013hpg_paper.pdf
Here's the documentation:
https://docs.google.com/document/d/1LuXW-CV-sVJkQaEGZlMJ86jZ8FmoPfecaMdR-oiWbUY/edit
Design discussion of the patch:
https://developer.blender.org/T44197
Differential Revision: https://developer.blender.org/D1200
2015-05-09 14:34:30 +00:00
|
|
|
#endif
|
2017-03-08 14:42:26 +00:00
|
|
|
Intersection isect;
|
2017-02-22 13:10:02 +00:00
|
|
|
PathState state = kernel_split_state.path_state[ray_index];
|
|
|
|
Ray ray = kernel_split_state.ray[ray_index];
|
Cycles: OpenCL kernel split
This commit contains all the work related on the AMD megakernel split work
which was mainly done by Varun Sundar, George Kyriazis and Lenny Wang, plus
some help from Sergey Sharybin, Martijn Berger, Thomas Dinges and likely
someone else which we're forgetting to mention.
Currently only AMD cards are enabled for the new split kernel, but it is
possible to force split opencl kernel to be used by setting the following
environment variable: CYCLES_OPENCL_SPLIT_KERNEL_TEST=1.
Not all the features are supported yet, and that being said no motion blur,
camera blur, SSS and volumetrics for now. Also transparent shadows are
disabled on AMD device because of some compiler bug.
This kernel is also only implements regular path tracing and supporting
branched one will take a bit. Branched path tracing is exposed to the
interface still, which is a bit misleading and will be hidden there soon.
More feature will be enabled once they're ported to the split kernel and
tested.
Neither regular CPU nor CUDA has any difference, they're generating the
same exact code, which means no regressions/improvements there.
Based on the research paper:
https://research.nvidia.com/sites/default/files/publications/laine2013hpg_paper.pdf
Here's the documentation:
https://docs.google.com/document/d/1LuXW-CV-sVJkQaEGZlMJ86jZ8FmoPfecaMdR-oiWbUY/edit
Design discussion of the patch:
https://developer.blender.org/T44197
Differential Revision: https://developer.blender.org/D1200
2015-05-09 14:34:30 +00:00
|
|
|
|
|
|
|
/* intersect scene */
|
|
|
|
uint visibility = path_state_ray_visibility(kg, &state);
|
|
|
|
|
|
|
|
#ifdef __HAIR__
|
|
|
|
float difl = 0.0f, extmax = 0.0f;
|
|
|
|
uint lcg_state = 0;
|
2017-02-22 13:10:02 +00:00
|
|
|
RNG rng = kernel_split_state.rng[ray_index];
|
Cycles: OpenCL kernel split
This commit contains all the work related on the AMD megakernel split work
which was mainly done by Varun Sundar, George Kyriazis and Lenny Wang, plus
some help from Sergey Sharybin, Martijn Berger, Thomas Dinges and likely
someone else which we're forgetting to mention.
Currently only AMD cards are enabled for the new split kernel, but it is
possible to force split opencl kernel to be used by setting the following
environment variable: CYCLES_OPENCL_SPLIT_KERNEL_TEST=1.
Not all the features are supported yet, and that being said no motion blur,
camera blur, SSS and volumetrics for now. Also transparent shadows are
disabled on AMD device because of some compiler bug.
This kernel is also only implements regular path tracing and supporting
branched one will take a bit. Branched path tracing is exposed to the
interface still, which is a bit misleading and will be hidden there soon.
More feature will be enabled once they're ported to the split kernel and
tested.
Neither regular CPU nor CUDA has any difference, they're generating the
same exact code, which means no regressions/improvements there.
Based on the research paper:
https://research.nvidia.com/sites/default/files/publications/laine2013hpg_paper.pdf
Here's the documentation:
https://docs.google.com/document/d/1LuXW-CV-sVJkQaEGZlMJ86jZ8FmoPfecaMdR-oiWbUY/edit
Design discussion of the patch:
https://developer.blender.org/T44197
Differential Revision: https://developer.blender.org/D1200
2015-05-09 14:34:30 +00:00
|
|
|
|
|
|
|
if(kernel_data.bvh.have_curves) {
|
|
|
|
if((kernel_data.cam.resolution == 1) && (state.flag & PATH_RAY_CAMERA)) {
|
|
|
|
float3 pixdiff = ray.dD.dx + ray.dD.dy;
|
|
|
|
/*pixdiff = pixdiff - dot(pixdiff, ray.D)*ray.D;*/
|
|
|
|
difl = kernel_data.curve.minimum_width * len(pixdiff) * 0.5f;
|
|
|
|
}
|
|
|
|
|
|
|
|
extmax = kernel_data.curve.maximum_width;
|
|
|
|
lcg_state = lcg_state_init(&rng, &state, 0x51633e2d);
|
|
|
|
}
|
|
|
|
|
2017-03-08 14:42:26 +00:00
|
|
|
bool hit = scene_intersect(kg, ray, visibility, &isect, &lcg_state, difl, extmax);
|
Cycles: OpenCL kernel split
This commit contains all the work related on the AMD megakernel split work
which was mainly done by Varun Sundar, George Kyriazis and Lenny Wang, plus
some help from Sergey Sharybin, Martijn Berger, Thomas Dinges and likely
someone else which we're forgetting to mention.
Currently only AMD cards are enabled for the new split kernel, but it is
possible to force split opencl kernel to be used by setting the following
environment variable: CYCLES_OPENCL_SPLIT_KERNEL_TEST=1.
Not all the features are supported yet, and that being said no motion blur,
camera blur, SSS and volumetrics for now. Also transparent shadows are
disabled on AMD device because of some compiler bug.
This kernel is also only implements regular path tracing and supporting
branched one will take a bit. Branched path tracing is exposed to the
interface still, which is a bit misleading and will be hidden there soon.
More feature will be enabled once they're ported to the split kernel and
tested.
Neither regular CPU nor CUDA has any difference, they're generating the
same exact code, which means no regressions/improvements there.
Based on the research paper:
https://research.nvidia.com/sites/default/files/publications/laine2013hpg_paper.pdf
Here's the documentation:
https://docs.google.com/document/d/1LuXW-CV-sVJkQaEGZlMJ86jZ8FmoPfecaMdR-oiWbUY/edit
Design discussion of the patch:
https://developer.blender.org/T44197
Differential Revision: https://developer.blender.org/D1200
2015-05-09 14:34:30 +00:00
|
|
|
#else
|
2017-03-08 14:42:26 +00:00
|
|
|
bool hit = scene_intersect(kg, ray, visibility, &isect, NULL, 0.0f, 0.0f);
|
Cycles: OpenCL kernel split
This commit contains all the work related on the AMD megakernel split work
which was mainly done by Varun Sundar, George Kyriazis and Lenny Wang, plus
some help from Sergey Sharybin, Martijn Berger, Thomas Dinges and likely
someone else which we're forgetting to mention.
Currently only AMD cards are enabled for the new split kernel, but it is
possible to force split opencl kernel to be used by setting the following
environment variable: CYCLES_OPENCL_SPLIT_KERNEL_TEST=1.
Not all the features are supported yet, and that being said no motion blur,
camera blur, SSS and volumetrics for now. Also transparent shadows are
disabled on AMD device because of some compiler bug.
This kernel is also only implements regular path tracing and supporting
branched one will take a bit. Branched path tracing is exposed to the
interface still, which is a bit misleading and will be hidden there soon.
More feature will be enabled once they're ported to the split kernel and
tested.
Neither regular CPU nor CUDA has any difference, they're generating the
same exact code, which means no regressions/improvements there.
Based on the research paper:
https://research.nvidia.com/sites/default/files/publications/laine2013hpg_paper.pdf
Here's the documentation:
https://docs.google.com/document/d/1LuXW-CV-sVJkQaEGZlMJ86jZ8FmoPfecaMdR-oiWbUY/edit
Design discussion of the patch:
https://developer.blender.org/T44197
Differential Revision: https://developer.blender.org/D1200
2015-05-09 14:34:30 +00:00
|
|
|
#endif
|
2017-03-08 14:42:26 +00:00
|
|
|
kernel_split_state.isect[ray_index] = isect;
|
Cycles: OpenCL kernel split
This commit contains all the work related on the AMD megakernel split work
which was mainly done by Varun Sundar, George Kyriazis and Lenny Wang, plus
some help from Sergey Sharybin, Martijn Berger, Thomas Dinges and likely
someone else which we're forgetting to mention.
Currently only AMD cards are enabled for the new split kernel, but it is
possible to force split opencl kernel to be used by setting the following
environment variable: CYCLES_OPENCL_SPLIT_KERNEL_TEST=1.
Not all the features are supported yet, and that being said no motion blur,
camera blur, SSS and volumetrics for now. Also transparent shadows are
disabled on AMD device because of some compiler bug.
This kernel is also only implements regular path tracing and supporting
branched one will take a bit. Branched path tracing is exposed to the
interface still, which is a bit misleading and will be hidden there soon.
More feature will be enabled once they're ported to the split kernel and
tested.
Neither regular CPU nor CUDA has any difference, they're generating the
same exact code, which means no regressions/improvements there.
Based on the research paper:
https://research.nvidia.com/sites/default/files/publications/laine2013hpg_paper.pdf
Here's the documentation:
https://docs.google.com/document/d/1LuXW-CV-sVJkQaEGZlMJ86jZ8FmoPfecaMdR-oiWbUY/edit
Design discussion of the patch:
https://developer.blender.org/T44197
Differential Revision: https://developer.blender.org/D1200
2015-05-09 14:34:30 +00:00
|
|
|
|
|
|
|
#ifdef __KERNEL_DEBUG__
|
|
|
|
if(state.flag & PATH_RAY_CAMERA) {
|
2017-03-08 14:42:26 +00:00
|
|
|
debug_data->num_bvh_traversed_nodes += isect.num_traversed_nodes;
|
|
|
|
debug_data->num_bvh_traversed_instances += isect.num_traversed_instances;
|
|
|
|
debug_data->num_bvh_intersections += isect.num_intersections;
|
Cycles: OpenCL kernel split
This commit contains all the work related on the AMD megakernel split work
which was mainly done by Varun Sundar, George Kyriazis and Lenny Wang, plus
some help from Sergey Sharybin, Martijn Berger, Thomas Dinges and likely
someone else which we're forgetting to mention.
Currently only AMD cards are enabled for the new split kernel, but it is
possible to force split opencl kernel to be used by setting the following
environment variable: CYCLES_OPENCL_SPLIT_KERNEL_TEST=1.
Not all the features are supported yet, and that being said no motion blur,
camera blur, SSS and volumetrics for now. Also transparent shadows are
disabled on AMD device because of some compiler bug.
This kernel is also only implements regular path tracing and supporting
branched one will take a bit. Branched path tracing is exposed to the
interface still, which is a bit misleading and will be hidden there soon.
More feature will be enabled once they're ported to the split kernel and
tested.
Neither regular CPU nor CUDA has any difference, they're generating the
same exact code, which means no regressions/improvements there.
Based on the research paper:
https://research.nvidia.com/sites/default/files/publications/laine2013hpg_paper.pdf
Here's the documentation:
https://docs.google.com/document/d/1LuXW-CV-sVJkQaEGZlMJ86jZ8FmoPfecaMdR-oiWbUY/edit
Design discussion of the patch:
https://developer.blender.org/T44197
Differential Revision: https://developer.blender.org/D1200
2015-05-09 14:34:30 +00:00
|
|
|
}
|
2015-06-11 08:42:38 +00:00
|
|
|
debug_data->num_ray_bounces++;
|
Cycles: OpenCL kernel split
This commit contains all the work related on the AMD megakernel split work
which was mainly done by Varun Sundar, George Kyriazis and Lenny Wang, plus
some help from Sergey Sharybin, Martijn Berger, Thomas Dinges and likely
someone else which we're forgetting to mention.
Currently only AMD cards are enabled for the new split kernel, but it is
possible to force split opencl kernel to be used by setting the following
environment variable: CYCLES_OPENCL_SPLIT_KERNEL_TEST=1.
Not all the features are supported yet, and that being said no motion blur,
camera blur, SSS and volumetrics for now. Also transparent shadows are
disabled on AMD device because of some compiler bug.
This kernel is also only implements regular path tracing and supporting
branched one will take a bit. Branched path tracing is exposed to the
interface still, which is a bit misleading and will be hidden there soon.
More feature will be enabled once they're ported to the split kernel and
tested.
Neither regular CPU nor CUDA has any difference, they're generating the
same exact code, which means no regressions/improvements there.
Based on the research paper:
https://research.nvidia.com/sites/default/files/publications/laine2013hpg_paper.pdf
Here's the documentation:
https://docs.google.com/document/d/1LuXW-CV-sVJkQaEGZlMJ86jZ8FmoPfecaMdR-oiWbUY/edit
Design discussion of the patch:
https://developer.blender.org/T44197
Differential Revision: https://developer.blender.org/D1200
2015-05-09 14:34:30 +00:00
|
|
|
#endif
|
|
|
|
|
|
|
|
if(!hit) {
|
|
|
|
/* Change the state of rays that hit the background;
|
|
|
|
* These rays undergo special processing in the
|
2015-07-03 08:56:40 +00:00
|
|
|
* background_bufferUpdate kernel.
|
|
|
|
*/
|
2017-02-22 13:10:02 +00:00
|
|
|
ASSIGN_RAY_STATE(kernel_split_state.ray_state, ray_index, RAY_HIT_BACKGROUND);
|
Cycles: OpenCL kernel split
This commit contains all the work related on the AMD megakernel split work
which was mainly done by Varun Sundar, George Kyriazis and Lenny Wang, plus
some help from Sergey Sharybin, Martijn Berger, Thomas Dinges and likely
someone else which we're forgetting to mention.
Currently only AMD cards are enabled for the new split kernel, but it is
possible to force split opencl kernel to be used by setting the following
environment variable: CYCLES_OPENCL_SPLIT_KERNEL_TEST=1.
Not all the features are supported yet, and that being said no motion blur,
camera blur, SSS and volumetrics for now. Also transparent shadows are
disabled on AMD device because of some compiler bug.
This kernel is also only implements regular path tracing and supporting
branched one will take a bit. Branched path tracing is exposed to the
interface still, which is a bit misleading and will be hidden there soon.
More feature will be enabled once they're ported to the split kernel and
tested.
Neither regular CPU nor CUDA has any difference, they're generating the
same exact code, which means no regressions/improvements there.
Based on the research paper:
https://research.nvidia.com/sites/default/files/publications/laine2013hpg_paper.pdf
Here's the documentation:
https://docs.google.com/document/d/1LuXW-CV-sVJkQaEGZlMJ86jZ8FmoPfecaMdR-oiWbUY/edit
Design discussion of the patch:
https://developer.blender.org/T44197
Differential Revision: https://developer.blender.org/D1200
2015-05-09 14:34:30 +00:00
|
|
|
}
|
|
|
|
}
|
2017-02-22 13:10:02 +00:00
|
|
|
|
|
|
|
CCL_NAMESPACE_END
|
|
|
|
|