forked from bartvdbraak/blender
Cycles: Cleanup and optimization comment update
This commit is contained in:
parent
eff2fe9a72
commit
09ac6cae09
@ -27,7 +27,12 @@ CCL_NAMESPACE_BEGIN
|
|||||||
|
|
||||||
ccl_device_inline int find_attribute_curve_motion(KernelGlobals *kg, int object, uint id, AttributeElement *elem)
|
ccl_device_inline int find_attribute_curve_motion(KernelGlobals *kg, int object, uint id, AttributeElement *elem)
|
||||||
{
|
{
|
||||||
/* todo: find a better (faster) solution for this, maybe store offset per object */
|
/* todo: find a better (faster) solution for this, maybe store offset per object.
|
||||||
|
*
|
||||||
|
* NOTE: currently it's not a bottleneck because in test scenes the loop below runs
|
||||||
|
* zero iterations and rendering is really slow with motion curves. For until other
|
||||||
|
* areas are speed up it's probably not so crucial to optimize this out.
|
||||||
|
*/
|
||||||
uint attr_offset = object*kernel_data.bvh.attributes_map_stride + ATTR_PRIM_CURVE;
|
uint attr_offset = object*kernel_data.bvh.attributes_map_stride + ATTR_PRIM_CURVE;
|
||||||
uint4 attr_map = kernel_tex_fetch(__attributes_map, attr_offset);
|
uint4 attr_map = kernel_tex_fetch(__attributes_map, attr_offset);
|
||||||
|
|
||||||
|
Loading…
Reference in New Issue
Block a user