Blender Git Loki
Git Commits -> Revision 57beb92
Revision 57beb92 by Geraldine Chua (soc-2018-cycles-volumes) August 12, 2018, 09:14 (GMT) |
Minor bug-fixes and optimizations. * Change most instances of division and modulo with TILE_SIZE to bit shifting. Much more efficient since TILE_SIZE should be a power of 2. * Added intialization for some Mesh members. Previously had motion blur randomly toggle on and off otherwise. * Fixed issue where voxel to tile correspondence is different between external VDBs and internal sparse grids. The fix requires generating a new VDB grid which may be too memory intensive. Better method would be translating grids in place. * Fixed misc OpenVDB to grid conversion issues. |
Commit Details:
Full Hash: 57beb92ccec7d4952636231fc54b106aacb17d55
Parent Commit: aa20e7b
Lines Changed: +70, -58
8 Modified Paths:
/intern/cycles/blender/blender_object.cpp (+0, -4) (Diff)
/intern/cycles/kernel/kernels/cpu/kernel_cpu_image.h (+10, -10) (Diff)
/intern/cycles/kernel/kernels/cuda/kernel_cuda_image.h (+7, -8) (Diff)
/intern/cycles/kernel/kernels/opencl/kernel_opencl_image.h (+8, -10) (Diff)
/intern/cycles/render/mesh.cpp (+4, -0) (Diff)
/intern/cycles/render/openvdb.cpp (+29, -18) (Diff)
/intern/cycles/util/util_sparse_grid.h (+8, -8) (Diff)
/intern/cycles/util/util_texture.h (+4, -0) (Diff)
/intern/cycles/kernel/kernels/cpu/kernel_cpu_image.h (+10, -10) (Diff)
/intern/cycles/kernel/kernels/cuda/kernel_cuda_image.h (+7, -8) (Diff)
/intern/cycles/kernel/kernels/opencl/kernel_opencl_image.h (+8, -10) (Diff)
/intern/cycles/render/mesh.cpp (+4, -0) (Diff)
/intern/cycles/render/openvdb.cpp (+29, -18) (Diff)
/intern/cycles/util/util_sparse_grid.h (+8, -8) (Diff)
/intern/cycles/util/util_texture.h (+4, -0) (Diff)