forked from bartvdbraak/blender
7f4479da42
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
161 lines
4.6 KiB
C
161 lines
4.6 KiB
C
/*
|
|
* Copyright 2011-2013 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.
|
|
*/
|
|
|
|
CCL_NAMESPACE_BEGIN
|
|
|
|
/* Attribute Node */
|
|
|
|
ccl_device void svm_node_attr_init(KernelGlobals *kg, ShaderData *sd,
|
|
uint4 node, NodeAttributeType *type,
|
|
NodeAttributeType *mesh_type, AttributeElement *elem, int *offset, uint *out_offset)
|
|
{
|
|
if(ccl_fetch(sd, object) != OBJECT_NONE) {
|
|
/* find attribute by unique id */
|
|
uint id = node.y;
|
|
uint attr_offset = ccl_fetch(sd, object)*kernel_data.bvh.attributes_map_stride;
|
|
#ifdef __HAIR__
|
|
attr_offset = (ccl_fetch(sd, type) & PRIMITIVE_ALL_CURVE)? attr_offset + ATTR_PRIM_CURVE: attr_offset;
|
|
#endif
|
|
uint4 attr_map = kernel_tex_fetch(__attributes_map, attr_offset);
|
|
|
|
while(attr_map.x != id) {
|
|
attr_offset += ATTR_PRIM_TYPES;
|
|
attr_map = kernel_tex_fetch(__attributes_map, attr_offset);
|
|
}
|
|
|
|
/* return result */
|
|
*elem = (AttributeElement)attr_map.y;
|
|
*offset = as_int(attr_map.z);
|
|
*mesh_type = (NodeAttributeType)attr_map.w;
|
|
}
|
|
else {
|
|
/* background */
|
|
*elem = ATTR_ELEMENT_NONE;
|
|
*offset = 0;
|
|
*mesh_type = (NodeAttributeType)node.w;
|
|
}
|
|
|
|
*out_offset = node.z;
|
|
*type = (NodeAttributeType)node.w;
|
|
}
|
|
|
|
ccl_device void svm_node_attr(KernelGlobals *kg, ShaderData *sd, float *stack, uint4 node)
|
|
{
|
|
NodeAttributeType type, mesh_type;
|
|
AttributeElement elem;
|
|
uint out_offset;
|
|
int offset;
|
|
|
|
svm_node_attr_init(kg, sd, node, &type, &mesh_type, &elem, &offset, &out_offset);
|
|
|
|
/* fetch and store attribute */
|
|
if(type == NODE_ATTR_FLOAT) {
|
|
if(mesh_type == NODE_ATTR_FLOAT) {
|
|
float f = primitive_attribute_float(kg, sd, elem, offset, NULL, NULL);
|
|
stack_store_float(stack, out_offset, f);
|
|
}
|
|
else {
|
|
float3 f = primitive_attribute_float3(kg, sd, elem, offset, NULL, NULL);
|
|
stack_store_float(stack, out_offset, average(f));
|
|
}
|
|
}
|
|
else {
|
|
if(mesh_type == NODE_ATTR_FLOAT3) {
|
|
float3 f = primitive_attribute_float3(kg, sd, elem, offset, NULL, NULL);
|
|
stack_store_float3(stack, out_offset, f);
|
|
}
|
|
else {
|
|
float f = primitive_attribute_float(kg, sd, elem, offset, NULL, NULL);
|
|
stack_store_float3(stack, out_offset, make_float3(f, f, f));
|
|
}
|
|
}
|
|
}
|
|
|
|
ccl_device void svm_node_attr_bump_dx(KernelGlobals *kg, ShaderData *sd, float *stack, uint4 node)
|
|
{
|
|
NodeAttributeType type, mesh_type;
|
|
AttributeElement elem;
|
|
uint out_offset;
|
|
int offset;
|
|
|
|
svm_node_attr_init(kg, sd, node, &type, &mesh_type, &elem, &offset, &out_offset);
|
|
|
|
/* fetch and store attribute */
|
|
if(type == NODE_ATTR_FLOAT) {
|
|
if(mesh_type == NODE_ATTR_FLOAT) {
|
|
float dx;
|
|
float f = primitive_attribute_float(kg, sd, elem, offset, &dx, NULL);
|
|
stack_store_float(stack, out_offset, f+dx);
|
|
}
|
|
else {
|
|
float3 dx;
|
|
float3 f = primitive_attribute_float3(kg, sd, elem, offset, &dx, NULL);
|
|
stack_store_float(stack, out_offset, average(f+dx));
|
|
}
|
|
}
|
|
else {
|
|
if(mesh_type == NODE_ATTR_FLOAT3) {
|
|
float3 dx;
|
|
float3 f = primitive_attribute_float3(kg, sd, elem, offset, &dx, NULL);
|
|
stack_store_float3(stack, out_offset, f+dx);
|
|
}
|
|
else {
|
|
float dx;
|
|
float f = primitive_attribute_float(kg, sd, elem, offset, &dx, NULL);
|
|
stack_store_float3(stack, out_offset, make_float3(f+dx, f+dx, f+dx));
|
|
}
|
|
}
|
|
}
|
|
|
|
ccl_device void svm_node_attr_bump_dy(KernelGlobals *kg, ShaderData *sd, float *stack, uint4 node)
|
|
{
|
|
NodeAttributeType type, mesh_type;
|
|
AttributeElement elem;
|
|
uint out_offset;
|
|
int offset;
|
|
|
|
svm_node_attr_init(kg, sd, node, &type, &mesh_type, &elem, &offset, &out_offset);
|
|
|
|
/* fetch and store attribute */
|
|
if(type == NODE_ATTR_FLOAT) {
|
|
if(mesh_type == NODE_ATTR_FLOAT) {
|
|
float dy;
|
|
float f = primitive_attribute_float(kg, sd, elem, offset, NULL, &dy);
|
|
stack_store_float(stack, out_offset, f+dy);
|
|
}
|
|
else {
|
|
float3 dy;
|
|
float3 f = primitive_attribute_float3(kg, sd, elem, offset, NULL, &dy);
|
|
stack_store_float(stack, out_offset, average(f+dy));
|
|
}
|
|
}
|
|
else {
|
|
if(mesh_type == NODE_ATTR_FLOAT3) {
|
|
float3 dy;
|
|
float3 f = primitive_attribute_float3(kg, sd, elem, offset, NULL, &dy);
|
|
stack_store_float3(stack, out_offset, f+dy);
|
|
}
|
|
else {
|
|
float dy;
|
|
float f = primitive_attribute_float(kg, sd, elem, offset, NULL, &dy);
|
|
stack_store_float3(stack, out_offset, make_float3(f+dy, f+dy, f+dy));
|
|
}
|
|
}
|
|
}
|
|
|
|
CCL_NAMESPACE_END
|
|
|