Blender Git Loki

Cycles: Code cleanup: use scoped timer to measure parts of SVM compiler

Cycles: Wrap SVM compiler state variables into a single struct

This way it's easier to pass stuff around and also much easier to add more
state variables.

So far should be no functional changes for artists.

Cycles: Optionally output luminance from the shader evaluation kernel

This makes it possible to move some parts of evaluation from host to the device
and hopefully reduce memory usage by avoid having full RGBA buffer on the host.

Reviewers: juicyfruit, lukasstockner97, brecht

Reviewed By: lukasstockner97, brecht

Differential Revision: https://developer.blender.org/D1702

Cycles: Don't install CPU-related kernel files

Cycles: Deduplicte CPU kernel declaration and definition code

Main goal is to make kernel signatures editing easier and less prone to the
errors caused by missing function signature update or so.

This will also make it easier to add new CPU architectures.

Reviewers: juicyfruit, dingto, lukasstockner97, brecht

Reviewed By: dingto, lukasstockner97, brecht

Differential Revision: https://developer.blender.org/D1703

BGE: Fix invalid operator< for microsoft compiler.

It fixes the strict weak ordering assertion failure, see : https://support.microsoft.com/en-us/kb/949171.
sybren and youle are the author of this commit.

GLog: Solve some compilation warnings

Those are actually sent to a pull-request, see

https://github.com/google/glog/pull/81

Libmv: Update to the latest upstream version

Main reason is to bring up new Glog which should have MSVC 2015
support.

Fix Moto overflow warnings

This fault (level 2, C4056) was introduced by changing Moto from double to float.

Skip some more workarounds when using MSVC 2015

Make cycles compile with MSVC 2015

Cycles: Add timing information to the SVM compiler stats

Cycles: Optimize nodes deduplication routines

The idea is to have separate sets per node name in order to speed up the
comparison process. This will use a bit more memory and slow down simple
shaders, but this extra memory is not so much huge and time penalty is
not really measurable (at least from initial tests).

This saves orders of magnitude seconds when de-duplicating 17K nodes and
overall process now takes 0.01sec on my laptop,

Cycles: Remove assert which is no longer valid

Was a left-over since 5b33115.

More from T47045: Add i18n translations to render status from compo.

Also fix inconsistency for freestyle translation - we use IFACE_ everywhere
(TIP_ may be more suited, but let's be consistent first!).

Building w/o OpenMP now works again

rename define since its no longer OpenMP specific.

Fix invalid memory access in multiview code during render, reported on irc.

OS X: remove code for OS X <= 10.5, which is already not supported since 2 years.

Fix memleaks in our gtests.

BLI_math_statistics: switch from OMP to BLI_task.

This time, with have over 300% speedup!

But no, this is not due to switch to BLI_task (which 'only' gives usal 15% speedup),
but to enhancement of the algorithm, flatten loop over covariance matrix items now allows
to compute (usually) all items in parallel, instead of having at most 3 or 4 working threads
(with unbalanced load even)...