2011-04-27 11:58:34 +00:00
|
|
|
/*
|
2013-08-18 14:16:15 +00:00
|
|
|
* Copyright 2011-2013 Blender Foundation
|
2011-04-27 11:58:34 +00:00
|
|
|
*
|
2013-08-18 14:16:15 +00:00
|
|
|
* 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
|
2011-04-27 11:58:34 +00:00
|
|
|
*
|
2013-08-18 14:16:15 +00:00
|
|
|
* http://www.apache.org/licenses/LICENSE-2.0
|
2011-04-27 11:58:34 +00:00
|
|
|
*
|
2013-08-18 14:16:15 +00:00
|
|
|
* 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
|
2014-12-25 01:50:24 +00:00
|
|
|
* limitations under the License.
|
2011-04-27 11:58:34 +00:00
|
|
|
*/
|
|
|
|
|
|
|
|
/* CPU kernel entry points */
|
|
|
|
|
2016-03-25 15:09:05 +00:00
|
|
|
/* On x86-64, we can assume SSE2, so avoid the extra kernel and compile this
|
|
|
|
* one with SSE2 intrinsics.
|
|
|
|
*/
|
2015-12-30 12:54:02 +00:00
|
|
|
#if defined(__x86_64__) || defined(_M_X64)
|
2016-02-12 17:33:43 +00:00
|
|
|
# define __KERNEL_SSE2__
|
2015-12-30 12:54:02 +00:00
|
|
|
#endif
|
|
|
|
|
2016-03-25 15:09:05 +00:00
|
|
|
/* When building kernel for native machine detect kernel features from the flags
|
|
|
|
* set by compiler.
|
|
|
|
*/
|
|
|
|
#ifdef WITH_KERNEL_NATIVE
|
|
|
|
# ifdef __SSE2__
|
|
|
|
# ifndef __KERNEL_SSE2__
|
|
|
|
# define __KERNEL_SSE2__
|
|
|
|
# endif
|
|
|
|
# endif
|
|
|
|
# ifdef __SSE3__
|
|
|
|
# define __KERNEL_SSE3__
|
|
|
|
# endif
|
|
|
|
# ifdef __SSSE3__
|
|
|
|
# define __KERNEL_SSSE3__
|
|
|
|
# endif
|
|
|
|
# ifdef __SSE4_1__
|
|
|
|
# define __KERNEL_SSE41__
|
|
|
|
# endif
|
|
|
|
# ifdef __AVX__
|
2016-10-25 14:10:47 +00:00
|
|
|
# define __KERNEL_SSE__
|
2016-03-25 15:09:05 +00:00
|
|
|
# define __KERNEL_AVX__
|
|
|
|
# endif
|
|
|
|
# ifdef __AVX2__
|
2016-10-12 15:35:03 +00:00
|
|
|
# define __KERNEL_SSE__
|
2016-03-25 15:09:05 +00:00
|
|
|
# define __KERNEL_AVX2__
|
|
|
|
# endif
|
|
|
|
#endif
|
|
|
|
|
2015-12-30 12:54:02 +00:00
|
|
|
/* quiet unused define warnings */
|
|
|
|
#if defined(__KERNEL_SSE2__)
|
|
|
|
/* do nothing */
|
|
|
|
#endif
|
|
|
|
|
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 "kernel/kernel.h"
|
2015-12-30 12:54:02 +00:00
|
|
|
#define KERNEL_ARCH cpu
|
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 "kernel/kernels/cpu/kernel_cpu_impl.h"
|
2011-04-27 11:58:34 +00:00
|
|
|
|
|
|
|
CCL_NAMESPACE_BEGIN
|
|
|
|
|
|
|
|
/* Memory Copy */
|
|
|
|
|
|
|
|
void kernel_const_copy(KernelGlobals *kg, const char *name, void *host, size_t size)
|
|
|
|
{
|
|
|
|
if(strcmp(name, "__data") == 0)
|
|
|
|
memcpy(&kg->__data, host, size);
|
|
|
|
else
|
|
|
|
assert(0);
|
|
|
|
}
|
|
|
|
|
2015-07-21 19:58:19 +00:00
|
|
|
void kernel_tex_copy(KernelGlobals *kg,
|
|
|
|
const char *name,
|
|
|
|
device_ptr mem,
|
|
|
|
size_t width,
|
|
|
|
size_t height,
|
|
|
|
size_t depth,
|
|
|
|
InterpolationType interpolation,
|
2015-07-28 11:51:10 +00:00
|
|
|
ExtensionType extension)
|
2011-04-27 11:58:34 +00:00
|
|
|
{
|
2011-09-27 20:37:24 +00:00
|
|
|
if(0) {
|
2011-04-27 11:58:34 +00:00
|
|
|
}
|
2011-09-27 20:37:24 +00:00
|
|
|
|
|
|
|
#define KERNEL_TEX(type, ttype, tname) \
|
|
|
|
else if(strcmp(name, #tname) == 0) { \
|
|
|
|
kg->tname.data = (type*)mem; \
|
|
|
|
kg->tname.width = width; \
|
2011-04-27 11:58:34 +00:00
|
|
|
}
|
2011-09-27 20:37:24 +00:00
|
|
|
#define KERNEL_IMAGE_TEX(type, ttype, tname)
|
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 "kernel/kernel_textures.h"
|
2011-09-27 20:37:24 +00:00
|
|
|
|
2016-05-06 11:42:50 +00:00
|
|
|
else if(strstr(name, "__tex_image_float4")) {
|
2012-03-07 12:27:18 +00:00
|
|
|
texture_image_float4 *tex = NULL;
|
2016-05-06 11:42:50 +00:00
|
|
|
int id = atoi(name + strlen("__tex_image_float4_"));
|
Unlimited number of textures for Cycles
This patch allows for an unlimited number of textures in Cycles where the hardware allows. It replaces a number static arrays with dynamic arrays and changes the way the flat_slot indices are calculated. Eventually, I'd like to get to a point where there are only flat slots left and textures off all kinds are stored in a single array.
Note that the arrays in DeviceScene are changed from containing device_vector<T> objects to device_vector<T>* pointers. Ideally, I'd like to store objects, but dynamic resizing of a std:vector in pre-C++11 calls the copy constructor, which for a good reason is not implemented for device_vector. Once we require C++11 for Cycles builds, we can implement a move constructor for device_vector and store objects again.
The limits for CUDA Fermi hardware still apply.
Reviewers: tod_baudais, InsigMathK, dingto, #cycles
Reviewed By: dingto, #cycles
Subscribers: dingto, smellslikedonkey
Differential Revision: https://developer.blender.org/D2650
2017-04-27 07:34:51 +00:00
|
|
|
int array_index = kernel_tex_index(id);
|
2012-03-07 12:27:18 +00:00
|
|
|
|
Unlimited number of textures for Cycles
This patch allows for an unlimited number of textures in Cycles where the hardware allows. It replaces a number static arrays with dynamic arrays and changes the way the flat_slot indices are calculated. Eventually, I'd like to get to a point where there are only flat slots left and textures off all kinds are stored in a single array.
Note that the arrays in DeviceScene are changed from containing device_vector<T> objects to device_vector<T>* pointers. Ideally, I'd like to store objects, but dynamic resizing of a std:vector in pre-C++11 calls the copy constructor, which for a good reason is not implemented for device_vector. Once we require C++11 for Cycles builds, we can implement a move constructor for device_vector and store objects again.
The limits for CUDA Fermi hardware still apply.
Reviewers: tod_baudais, InsigMathK, dingto, #cycles
Reviewed By: dingto, #cycles
Subscribers: dingto, smellslikedonkey
Differential Revision: https://developer.blender.org/D2650
2017-04-27 07:34:51 +00:00
|
|
|
if(array_index >= 0) {
|
|
|
|
if(array_index >= kg->texture_float4_images.size()) {
|
|
|
|
kg->texture_float4_images.resize(array_index+1);
|
|
|
|
}
|
2016-05-06 11:42:50 +00:00
|
|
|
tex = &kg->texture_float4_images[array_index];
|
2012-03-07 12:27:18 +00:00
|
|
|
}
|
|
|
|
|
|
|
|
if(tex) {
|
|
|
|
tex->data = (float4*)mem;
|
2014-05-04 20:57:33 +00:00
|
|
|
tex->dimensions_set(width, height, depth);
|
2014-03-07 22:16:09 +00:00
|
|
|
tex->interpolation = interpolation;
|
2015-07-28 11:51:10 +00:00
|
|
|
tex->extension = extension;
|
2012-03-07 12:27:18 +00:00
|
|
|
}
|
|
|
|
}
|
2016-05-09 10:51:42 +00:00
|
|
|
else if(strstr(name, "__tex_image_float")) {
|
|
|
|
texture_image_float *tex = NULL;
|
|
|
|
int id = atoi(name + strlen("__tex_image_float_"));
|
Unlimited number of textures for Cycles
This patch allows for an unlimited number of textures in Cycles where the hardware allows. It replaces a number static arrays with dynamic arrays and changes the way the flat_slot indices are calculated. Eventually, I'd like to get to a point where there are only flat slots left and textures off all kinds are stored in a single array.
Note that the arrays in DeviceScene are changed from containing device_vector<T> objects to device_vector<T>* pointers. Ideally, I'd like to store objects, but dynamic resizing of a std:vector in pre-C++11 calls the copy constructor, which for a good reason is not implemented for device_vector. Once we require C++11 for Cycles builds, we can implement a move constructor for device_vector and store objects again.
The limits for CUDA Fermi hardware still apply.
Reviewers: tod_baudais, InsigMathK, dingto, #cycles
Reviewed By: dingto, #cycles
Subscribers: dingto, smellslikedonkey
Differential Revision: https://developer.blender.org/D2650
2017-04-27 07:34:51 +00:00
|
|
|
int array_index = kernel_tex_index(id);
|
|
|
|
|
|
|
|
if(array_index >= 0) {
|
|
|
|
if(array_index >= kg->texture_float_images.size()) {
|
|
|
|
kg->texture_float_images.resize(array_index+1);
|
|
|
|
}
|
2016-05-09 10:51:42 +00:00
|
|
|
tex = &kg->texture_float_images[array_index];
|
|
|
|
}
|
|
|
|
|
|
|
|
if(tex) {
|
|
|
|
tex->data = (float*)mem;
|
|
|
|
tex->dimensions_set(width, height, depth);
|
|
|
|
tex->interpolation = interpolation;
|
|
|
|
tex->extension = extension;
|
|
|
|
}
|
|
|
|
}
|
2016-05-09 00:22:01 +00:00
|
|
|
else if(strstr(name, "__tex_image_byte4")) {
|
2011-04-27 11:58:34 +00:00
|
|
|
texture_image_uchar4 *tex = NULL;
|
2016-05-09 00:22:01 +00:00
|
|
|
int id = atoi(name + strlen("__tex_image_byte4_"));
|
Unlimited number of textures for Cycles
This patch allows for an unlimited number of textures in Cycles where the hardware allows. It replaces a number static arrays with dynamic arrays and changes the way the flat_slot indices are calculated. Eventually, I'd like to get to a point where there are only flat slots left and textures off all kinds are stored in a single array.
Note that the arrays in DeviceScene are changed from containing device_vector<T> objects to device_vector<T>* pointers. Ideally, I'd like to store objects, but dynamic resizing of a std:vector in pre-C++11 calls the copy constructor, which for a good reason is not implemented for device_vector. Once we require C++11 for Cycles builds, we can implement a move constructor for device_vector and store objects again.
The limits for CUDA Fermi hardware still apply.
Reviewers: tod_baudais, InsigMathK, dingto, #cycles
Reviewed By: dingto, #cycles
Subscribers: dingto, smellslikedonkey
Differential Revision: https://developer.blender.org/D2650
2017-04-27 07:34:51 +00:00
|
|
|
int array_index = kernel_tex_index(id);
|
|
|
|
|
|
|
|
if(array_index >= 0) {
|
|
|
|
if(array_index >= kg->texture_byte4_images.size()) {
|
|
|
|
kg->texture_byte4_images.resize(array_index+1);
|
|
|
|
}
|
2016-05-09 07:16:41 +00:00
|
|
|
tex = &kg->texture_byte4_images[array_index];
|
2011-04-27 11:58:34 +00:00
|
|
|
}
|
|
|
|
|
|
|
|
if(tex) {
|
|
|
|
tex->data = (uchar4*)mem;
|
2014-05-04 20:57:33 +00:00
|
|
|
tex->dimensions_set(width, height, depth);
|
2014-03-07 22:16:09 +00:00
|
|
|
tex->interpolation = interpolation;
|
2015-07-28 11:51:10 +00:00
|
|
|
tex->extension = extension;
|
2011-04-27 11:58:34 +00:00
|
|
|
}
|
|
|
|
}
|
2016-05-12 12:51:42 +00:00
|
|
|
else if(strstr(name, "__tex_image_byte")) {
|
|
|
|
texture_image_uchar *tex = NULL;
|
|
|
|
int id = atoi(name + strlen("__tex_image_byte_"));
|
Unlimited number of textures for Cycles
This patch allows for an unlimited number of textures in Cycles where the hardware allows. It replaces a number static arrays with dynamic arrays and changes the way the flat_slot indices are calculated. Eventually, I'd like to get to a point where there are only flat slots left and textures off all kinds are stored in a single array.
Note that the arrays in DeviceScene are changed from containing device_vector<T> objects to device_vector<T>* pointers. Ideally, I'd like to store objects, but dynamic resizing of a std:vector in pre-C++11 calls the copy constructor, which for a good reason is not implemented for device_vector. Once we require C++11 for Cycles builds, we can implement a move constructor for device_vector and store objects again.
The limits for CUDA Fermi hardware still apply.
Reviewers: tod_baudais, InsigMathK, dingto, #cycles
Reviewed By: dingto, #cycles
Subscribers: dingto, smellslikedonkey
Differential Revision: https://developer.blender.org/D2650
2017-04-27 07:34:51 +00:00
|
|
|
int array_index = kernel_tex_index(id);
|
|
|
|
|
|
|
|
if(array_index >= 0) {
|
|
|
|
if(array_index >= kg->texture_byte_images.size()) {
|
|
|
|
kg->texture_byte_images.resize(array_index+1);
|
|
|
|
}
|
2016-05-12 12:51:42 +00:00
|
|
|
tex = &kg->texture_byte_images[array_index];
|
|
|
|
}
|
|
|
|
|
|
|
|
if(tex) {
|
|
|
|
tex->data = (uchar*)mem;
|
|
|
|
tex->dimensions_set(width, height, depth);
|
|
|
|
tex->interpolation = interpolation;
|
|
|
|
tex->extension = extension;
|
|
|
|
}
|
|
|
|
}
|
2016-06-19 15:31:16 +00:00
|
|
|
else if(strstr(name, "__tex_image_half4")) {
|
|
|
|
texture_image_half4 *tex = NULL;
|
|
|
|
int id = atoi(name + strlen("__tex_image_half4_"));
|
Unlimited number of textures for Cycles
This patch allows for an unlimited number of textures in Cycles where the hardware allows. It replaces a number static arrays with dynamic arrays and changes the way the flat_slot indices are calculated. Eventually, I'd like to get to a point where there are only flat slots left and textures off all kinds are stored in a single array.
Note that the arrays in DeviceScene are changed from containing device_vector<T> objects to device_vector<T>* pointers. Ideally, I'd like to store objects, but dynamic resizing of a std:vector in pre-C++11 calls the copy constructor, which for a good reason is not implemented for device_vector. Once we require C++11 for Cycles builds, we can implement a move constructor for device_vector and store objects again.
The limits for CUDA Fermi hardware still apply.
Reviewers: tod_baudais, InsigMathK, dingto, #cycles
Reviewed By: dingto, #cycles
Subscribers: dingto, smellslikedonkey
Differential Revision: https://developer.blender.org/D2650
2017-04-27 07:34:51 +00:00
|
|
|
int array_index = kernel_tex_index(id);
|
|
|
|
|
|
|
|
if(array_index >= 0) {
|
|
|
|
if(array_index >= kg->texture_half4_images.size()) {
|
|
|
|
kg->texture_half4_images.resize(array_index+1);
|
|
|
|
}
|
2016-06-19 15:31:16 +00:00
|
|
|
tex = &kg->texture_half4_images[array_index];
|
|
|
|
}
|
|
|
|
|
|
|
|
if(tex) {
|
|
|
|
tex->data = (half4*)mem;
|
|
|
|
tex->dimensions_set(width, height, depth);
|
|
|
|
tex->interpolation = interpolation;
|
|
|
|
tex->extension = extension;
|
|
|
|
}
|
|
|
|
}
|
|
|
|
else if(strstr(name, "__tex_image_half")) {
|
|
|
|
texture_image_half *tex = NULL;
|
|
|
|
int id = atoi(name + strlen("__tex_image_half_"));
|
Unlimited number of textures for Cycles
This patch allows for an unlimited number of textures in Cycles where the hardware allows. It replaces a number static arrays with dynamic arrays and changes the way the flat_slot indices are calculated. Eventually, I'd like to get to a point where there are only flat slots left and textures off all kinds are stored in a single array.
Note that the arrays in DeviceScene are changed from containing device_vector<T> objects to device_vector<T>* pointers. Ideally, I'd like to store objects, but dynamic resizing of a std:vector in pre-C++11 calls the copy constructor, which for a good reason is not implemented for device_vector. Once we require C++11 for Cycles builds, we can implement a move constructor for device_vector and store objects again.
The limits for CUDA Fermi hardware still apply.
Reviewers: tod_baudais, InsigMathK, dingto, #cycles
Reviewed By: dingto, #cycles
Subscribers: dingto, smellslikedonkey
Differential Revision: https://developer.blender.org/D2650
2017-04-27 07:34:51 +00:00
|
|
|
int array_index = kernel_tex_index(id);
|
|
|
|
|
|
|
|
if(array_index >= 0) {
|
|
|
|
if(array_index >= kg->texture_half_images.size()) {
|
|
|
|
kg->texture_half_images.resize(array_index+1);
|
|
|
|
}
|
2016-06-19 15:31:16 +00:00
|
|
|
tex = &kg->texture_half_images[array_index];
|
|
|
|
}
|
|
|
|
|
|
|
|
if(tex) {
|
|
|
|
tex->data = (half*)mem;
|
|
|
|
tex->dimensions_set(width, height, depth);
|
|
|
|
tex->interpolation = interpolation;
|
|
|
|
tex->extension = extension;
|
|
|
|
}
|
|
|
|
}
|
2011-04-27 11:58:34 +00:00
|
|
|
else
|
|
|
|
assert(0);
|
|
|
|
}
|
|
|
|
|
|
|
|
CCL_NAMESPACE_END
|