f5f7024040
New ("fullframe") CPU compositor backend is being used now, and all the code
related to "tiled" CPU compositor is just never used anymore. The new backend
is faster, uses less memory, better matches GPU compositor, etc.
TL;DR: 20 thousand lines of code gone.
This commit:
- Removes various bits and pieces related to "tiled" compositor (execution
groups, one-pixel-at-a-time node processing, read/write buffer operations
related to node execution groups).
- "GPU" (OpenCL) execution device, that was only used by several nodes of
the tiled compositor.
- With that, remove CLEW external library too, since nothing within Blender
uses OpenCL directly anymore.
Pull Request: https://projects.blender.org/blender/blender/pulls/118819
131 lines
3.6 KiB
C++
131 lines
3.6 KiB
C++
/* SPDX-FileCopyrightText: 2011 Blender Authors
|
|
*
|
|
* SPDX-License-Identifier: GPL-2.0-or-later */
|
|
|
|
#include "COM_CalculateMeanOperation.h"
|
|
|
|
#include "COM_ExecutionSystem.h"
|
|
|
|
#include "IMB_colormanagement.hh"
|
|
|
|
namespace blender::compositor {
|
|
|
|
CalculateMeanOperation::CalculateMeanOperation()
|
|
{
|
|
this->add_input_socket(DataType::Color, ResizeMode::Align);
|
|
this->add_output_socket(DataType::Value);
|
|
is_calculated_ = false;
|
|
setting_ = 1;
|
|
flags_.is_constant_operation = true;
|
|
|
|
needs_canvas_to_get_constant_ = true;
|
|
}
|
|
|
|
void CalculateMeanOperation::init_execution()
|
|
{
|
|
is_calculated_ = false;
|
|
}
|
|
|
|
void CalculateMeanOperation::set_setting(int setting)
|
|
{
|
|
setting_ = setting;
|
|
switch (setting) {
|
|
case 1: {
|
|
setting_func_ = IMB_colormanagement_get_luminance;
|
|
break;
|
|
}
|
|
case 2: {
|
|
setting_func_ = [](const float *elem) { return elem[0]; };
|
|
break;
|
|
}
|
|
case 3: {
|
|
setting_func_ = [](const float *elem) { return elem[1]; };
|
|
break;
|
|
}
|
|
case 4: {
|
|
setting_func_ = [](const float *elem) { return elem[2]; };
|
|
break;
|
|
}
|
|
case 5: {
|
|
setting_func_ = [](const float *elem) {
|
|
float yuv[3];
|
|
rgb_to_yuv(elem[0], elem[1], elem[2], &yuv[0], &yuv[1], &yuv[2], BLI_YUV_ITU_BT709);
|
|
return yuv[0];
|
|
};
|
|
break;
|
|
}
|
|
}
|
|
}
|
|
|
|
void CalculateMeanOperation::get_area_of_interest(int input_idx,
|
|
const rcti & /*output_area*/,
|
|
rcti &r_input_area)
|
|
{
|
|
BLI_assert(input_idx == 0);
|
|
r_input_area = get_input_operation(input_idx)->get_canvas();
|
|
}
|
|
|
|
void CalculateMeanOperation::determine_canvas(const rcti &preferred_area, rcti &r_area)
|
|
{
|
|
ConstantOperation::determine_canvas(preferred_area, r_area);
|
|
r_area = preferred_area;
|
|
}
|
|
|
|
const float *CalculateMeanOperation::get_constant_elem()
|
|
{
|
|
/* Node de-duplication uses the constant value as part of a hash for constant operations.
|
|
* The constant is not known in advance here, but need to return something. The value does
|
|
* not really matter, because if two CalculateMean operations are connected to different
|
|
* inputs it will be handled via hash of the input subtree. */
|
|
static float f = 0;
|
|
return &f;
|
|
}
|
|
|
|
void CalculateMeanOperation::update_memory_buffer(MemoryBuffer *output,
|
|
const rcti &area,
|
|
Span<MemoryBuffer *> inputs)
|
|
{
|
|
if (!is_calculated_) {
|
|
MemoryBuffer *input = inputs[0];
|
|
constant_value_ = calculate_value(input);
|
|
is_calculated_ = true;
|
|
}
|
|
|
|
output->fill(area, &constant_value_);
|
|
}
|
|
|
|
float CalculateMeanOperation::calculate_value(const MemoryBuffer *input) const
|
|
{
|
|
return calculate_mean(input);
|
|
}
|
|
|
|
float CalculateMeanOperation::calculate_mean(const MemoryBuffer *input) const
|
|
{
|
|
PixelsSum total = {0};
|
|
exec_system_->execute_work<PixelsSum>(
|
|
input->get_rect(),
|
|
[=](const rcti &split) { return calc_area_sum(input, split); },
|
|
total,
|
|
[](PixelsSum &join, const PixelsSum &chunk) {
|
|
join.sum += chunk.sum;
|
|
join.num_pixels += chunk.num_pixels;
|
|
});
|
|
return total.num_pixels == 0 ? 0.0f : total.sum / total.num_pixels;
|
|
}
|
|
|
|
using PixelsSum = CalculateMeanOperation::PixelsSum;
|
|
PixelsSum CalculateMeanOperation::calc_area_sum(const MemoryBuffer *input, const rcti &area) const
|
|
{
|
|
PixelsSum result = {0};
|
|
for (const float *elem : input->get_buffer_area(area)) {
|
|
if (elem[3] <= 0.0f) {
|
|
continue;
|
|
}
|
|
result.sum += setting_func_(elem);
|
|
result.num_pixels++;
|
|
}
|
|
return result;
|
|
}
|
|
|
|
} // namespace blender::compositor
|