Blender Git Commits

API: Fix Links

Self-explanatory. to find broken links run `sphinx-build -b linkcheck sphinx-in sphinx-out`

Reviewers: mont29

Tags: #bf_blender, #python, #infrastructure:_websites

Differential Revision: https://developer.blender.org/D2297

Partial fix for T49836: Camera DOF properties not updating via graph editor

Do this for the new dependency graph: was missing handle of OB_UPDATE_TIME in tag update.

Hopefully it's all correct still.

Old dependency graph needs work, but i'm tempting to call it unsupported and move on
to 2.8 branch.

Cycles: Enable SSE math optimization for AVX kernels

This gives about 5% speedup for AVX processors.

Benefit of such optimization on other microarchitectures is still
under investigation.

Cycles: Use new SSE version of offset calculation for all QBVH flavors

Gives up to ~1% speedup again.

While it seems to be small, still nice since the code now is actually more
clean that it used to be before.

Cycles: Move QBVH near/far offset calculation to an utility function

Just preparing for new optimization to be used in all traversal implementation.

Should be no measurable difference.

Cycles: BVH-related SSE optimization

Several ideas here:

- Optimize calculation of near_{x,y,z} in a way that does not require
3 if() statements per update, which avoids negative effect of wrong
branch prediction.

- Optimization of direction clamping for BVH.

- Optimization of point/direction transform.

Brings ~1.5% speedup again depending on a scene (unfortunately, this
speedup can't be sum across all previous commits because speedup of
each of the changes varies from scene to scene, but it still seems to
be nice solid speedup of few percent on Linux and bigger speedup was
reported on Windows).

Once again ,thanks Maxym for inspiration!

Still TODO: We have multiple places where we need to calculate near
x,y,z indices in BVH, for now it's only done for main BVH traversal.
Will try to move this calculation to an utility function and see if
that can be easily re-used across all the BVH flavors.

Cycles: Avoid branching in SSE version of intersection pre-calculation

Similar to the previous commit, avoid negative effect of bad branch prediction.

Gives measurable performance up to ~2% in tests here.

Once again, thanks to Maxym Dmytrychenko!

Cycles: Implement SSE-optimized path of util_max_axis()

The idea here is to avoid if statements which could cause wrong
branch prediction.

Gives a bit of measurable speedup up to ~1%. Still nice :)

Inspired by Maxym Dmytrychenko, thanks!

Viewport: fix depth not being cleared in ortho mode

Merge remote-tracking branch 'origin/master' into blender2.8

OpenGL: mesh batch cache (WIP)

Gawain batches are built on demand while drawing, then kept in this per-DerivedMesh cache.

A mesh's batches try to share vertex buffers as much as possible.

Not sure if this file is the best home for this code, but functions in this file are the only users of the cache. So maybe.

Big part of T49165

make View3D depth debug code optional

Quick hack to optionally disable depth buffer debugging at compile time.

Will this be useful for users or just during development?

OpenGL: 3D View background & depth buffer details

Surveying buffer usage & clears for new viewport. Not yet perfect, but closer. Committing from Mac so I can test this on Windows.

Using new matrix API (T49450) for gradient background.

CMake: Followup to previous commit, try to ensure -ldl is always last

Seems CMake will rearrange and copy libraries which are passed to the linker
when some of the libraries is listed twice (for example, -lz from png libraries
and -l for blender itself). This was causing libopenimageio to be added somewhere
at the end of linking flags without -ldl followed after which was causing linking
issues.

Fix possible compilation error with OIIO enabled

OIIO library has plugin API which uses dlopen()/dlclose() so need to
link OIO libraries against dl library.

OpenGL: new built-in shaders for mesh edges

OpenGL: clean up glActiveTexture usage

Removed some of my earlier glActiveTexture calls. After reviewing the
code I now trust that GL_TEXTURE0 is active by default. Fewer GL calls,
same results.

Fixed some misuse of glActiveTexture & glUniformi, mostly my fault.
Caught by --debug-gpu on Windows. Don't know why this appeared to be
working previously!

Plus some easy cleanup nearby.

Viewport: original gradient background

Last but not least, fix the original 3d view background based on
@merwin's comment.

Viewport: move the background to the "modern" viewport

This also leaves room for the gradient background Mike Erwin wants to
implement.

immediate mode: background

There is a problem here, which is that we can't use immediate mode here
until we rely on `GPU_material_bind`.

(cc: @merwin)