Blender Git Loki

Fix crash when deleting active layer

dynamic frac fixes (better activation, ability to merge cluster count and group)

Merge branch 'master' into soc-2016-layer_manager

Conflicts:
source/blender/blenloader/intern/versioning_270.c
source/blender/editors/interface/resources.c

Merge branch 'master' into temp_localview_split

Conflicts:
source/blender/makesrna/intern/rna_space.c

Fix T49631: radial control operators not using DPI properly.

Merge branch 'master' into HMD_viewport

Proper default value for custom HMD IPD

Use modelview matrix from OpenHMD, get custom IPD to work

Also renamed interocular distance to IPD in UI (slightly different things), removed unused matrix getters and made sure enabling/disabling custom IPD behaves nicely. Note that both projection and modelview matrices from OpenHMD get the Blender equivalents applied on top, so they act as relative transformations. Rotating, zooming and panning in HMD view works properly too now.

Fix T49640: Cycles constant folding incorrect for texture coordinates.

Fix build error with WITH_CYCLES_NATIVE_ONLY and recent AVX2 changes.

crash fixes for dynamic split, also attempt to get rotations correct with limit impact

Fix T49636: material draw mode crash with displacement and missing group input node.

FFmpeg: Fix off by one error in number of detected frames in matroska container

Seems to be rounding error. Hopefully new code handles the error fixed back in
SVN revision 28901 and still have proper frame number for Hjalti.

What could possibly go wrong here..

Cycles: Use const reference for register variables in non-OpenCL code

This is something tested by @LazyDodo and suggested by Maxym to make
MSVC happier.

Cycles: Use more SSE intrinsics for float3 type

This gives about 5% speedup on AVX2 kernels (other kernels still
have SSE disabled for math operations) and this solves the slowdown
of koro scene mention in the previous commit.

The title says it all actually. This commit also contains
changes to pass float3 as const reference in affected functions.

This should make MSVC happier without breaking OpenCL because it's
only done in areas which are ifdef-ed for non-OpenCL.

Another patch based on inspiration from Maxym Dmytrychenko, thanks!

Cycles: Implement AVX2 version of triangle_intersect

This commit basically vectorizes existing code using AVX2 instructions
(without modifying algorithm itself). This gives quite nice speedups:

BMW: -8%
Classroom: -5%
Cat: -5%
Koro: +1%
Barcelona: -8%

That's on Linux machine, reported performance improvement on Windows
goes up to 20%.

Not currently sure why Koro is somewhat slower because it mainly uses
curve intersection tests, could be a time noise? Or osmething with the
cache utilization perhaps? In any case speedup in other scenes makes
me thinking that current state is acceptable for initial implementation.

This is again inspired by Maxym Dmytrychenko.

Cycles: Add new avxf vectorized data type

Based on existing ssef data type and to my knowledge it's also what happens in
Embree nowadays.

Inspired by Maxym Dmytrychenko and required for the upcoming triangle
intersection commit.

Hopefully the copyright message is correct.

Always use projection matrix from OpenHMD

Was previously only used while in camera view.
Right now, you can't zoom during HMD session, should be fixable though. Custom interocular distance and optional HMD rotation don't work now either, need to check if we want to keep them.

Cycles: Enable SSE options of math module for AVX2 kernels

Currently this does not give measurable difference, but is required
ground work for some upcoming further optimization of AVX2 kernels.

Cycles: Get rid of ifdef-ed noinline policy