Blender Git Commits

Cycles: Don't use std::set directly, use util_set instead

minor cleanup, comments

Make UI follow laplacian deform

Merge branch 'master' into asset-experiments

Quick attempt to resume thumbnails generation on cache.

Note this is rather stupid anyway, we need to rewrite this from scratch,
using a job for that makes no sense now, we should stick to a lighter
threading schema with a queue, or something like that!

Cycles: Don't limit OpenCL devices by the env variable

Silly whitespace fix to test git config

Merge branch 'master' into wiggly-widgets

Conflicts:
source/blender/editors/include/ED_transform.h
source/blender/gpu/intern/gpu_buffers.c

Merge branch 'alembic' into gooseberry

Cycles: Compilation error fixes

Also made checks about __SPLIT_KERNEL__ srvive cases
when it's not defined.

Fix for crashes when changing draw types that affect dupli objects.

The draw code was restoring the derivedFinal mesh unconditionally after
drawing duplis. However, the drawing code can in some cases replace the
derivedFinal pointer during drawing (bad!). This then leads to invalid
pointer access.

Fix is to only touch the derivedFinal when it's actually a dupli cache
mesh.

Cycles: Make code style closer to Cycles, avoid some unneeded changes

Merge branch 'alembic' into gooseberry

`cachelibrary.modifier_bake operator` has been removed in favor of
`cachelibrary.bake`.

Preliminary fix for bad level calls to WM job functions from BKE.

The job invocation should happen only in editors/. Also it makes sense
to modify the bake operator such that it supports both the basic cache
library and an optional modifier output.

The design for this area is in progress anyway, probably the cache
library archive paths will be changed so that there is always one input
and optionally an output path, with all the modifiers working in
sequence.

Removed unused code for ill-defined "cache objects".

The cache libraries now work based on dupli objects instead of
recursive group content, so this concept is no longer needed.

Simplified functions for generating absolute file paths from the base
cache library and modifiers.

Sanity check: motion state data in curves is optional.

Write out the complete dupli group in cache modifiers.

Theoretically we could write a stitcher that writes only the parts of
the data affected by a modifier and simply copies the rest, but this
is much more involved and not necessary at this point.

Merge branch 'master' into cycles_kernel_split

Conflicts:
intern/cycles/device/device_opencl.cpp
intern/cycles/kernel/kernel_shader.h
intern/cycles/kernel/svm/svm_closure.h
intern/cycles/kernel/svm/svm_wavelength.h
intern/cycles/kernel/svm/svm_wireframe.h

Cycles: OpenCL kernel-splitting work

This patch contains work to split the OpenCL mega-kernel into separate kernels
to obtain better GPU utilization and therefore performance.

A description of the optimizations included in this patch is located at

https://docs.google.com/document/d/1LuXW-CV-sVJkQaEGZlMJ86jZ8FmoPfecaMdR-oiWbUY/edit?usp=sharing