archive/blender-archive
Archived
1
1
Fork 0
Code Pull Requests Packages Activity

Cycles: Remove sum_all_radiance kernel

Browse Source 
This was only needed for the previous implementation of parallel samples. As
we don't have that any more it can be removed.

Real reason for removal tho is this: `per_sample_output_buffers` was being
calculated too small and artifacts resulted. The tile buffer is already
the correct size and calculating the size for `per_sample_output_buffers`
is a bit difficult with the current layout of the code. As
`per_sample_output_buffers` was only needed for `sum_all_radiance`,
removing that kernel and writing output to the tile buffer directly
fixes the artifacts.
This commit is contained in:
Mai Lavelle
2017-03-01 01:47:08 -05:00
parent 4cf501b835
commit cd7d5669d1
13 changed files with 46 additions and 137 deletions
Download Patch File Download Diff File Expand all files Collapse all files

30
intern/cycles/kernel/split/kernel_data_init.h
View File

@@ -24,6 +24,21 @@ CCL_NAMESPACE_BEGIN
* The number of elements in the queues is initialized to 0;
*/
/* distributes an amount of work across all threads
* note: work done inside the loop may not show up to all threads till after the current kernel has completed
*/
#define parallel_for(kg, iter_name, work_size) \
for(size_t _size = (work_size), \
_global_size = ccl_global_size(0) * ccl_global_size(1), \
_n = _size / _global_size, \
_thread = ccl_global_id(0) + ccl_global_id(1) * ccl_global_size(0), \
iter_name = (_n > 0) ? (_thread * _n) : (_thread) \
; \
(iter_name < (_thread+1) * _n) || (iter_name == _n * _global_size + _thread && _thread < _size % _global_size) \
; \
iter_name = (iter_name != (_thread+1) * _n - 1) ? (iter_name + 1) : (_n * _global_size + _thread) \
)
#ifndef __KERNEL_CPU__
ccl_device void kernel_data_init(
#else
@@ -110,6 +125,21 @@ void KERNEL_FUNCTION_FULL_NAME(data_init)(
*/
*use_queues_flag = 0;
}
/* zero the tiles pixels if this is the first sample */
if(start_sample == 0) {
parallel_for(kg, i, sw * sh * kernel_data.film.pass_stride) {
int pixel = i / kernel_data.film.pass_stride;
int pass = i % kernel_data.film.pass_stride;
int x = sx + pixel % sw;
int y = sy + pixel / sw;
int index = (offset + x + y*stride) * kernel_data.film.pass_stride + pass;
*(buffer + index) = 0.0f;
}
}
}
CCL_NAMESPACE_END