49199963bf
Fix incorrect CUDA remaining time estimate after previous commit.
2017-10-04 23:25:51 +02:00
6da6f8d33f
Cycles: CUDA faster rendering of small tiles, using multiple samples like OpenCL.
...
The work size is still very conservative, and this doesn't help for progressive
refine. For that we will need to render multiple tiles at the same time. But this
should already help for denoising renders that require too much memory with big
tiles, and just generally soften the performance dropoff with small tiles.
Differential Revision: https://developer.blender.org/D2856
2017-10-04 21:58:47 +02:00
77f300e2a9
Fix use of uninitialized memory in Cycles normal baking.
2017-10-04 21:11:14 +02:00
5bb677e592
Code refactor: zero render buffers outside of kernel.
...
This was originally done with the first sample in the kernel for better
performance, but it doesn't work anymore with atomics. Any benefit was
very minor anyway, too small to measure it seems.
2017-10-04 21:11:14 +02:00
12f4538205
Code refactor: use split variance calculation for mega kernels too.
...
There is no significant difference in denoised benchmark scenes and
denoising ctests, so might as well make it all consistent.
2017-10-04 21:11:14 +02:00
e3e16cecc4
Code refactor: remove rng_state buffer and compute hash on the fly.
...
A little faster on some benchmark scenes, a little slower on others, seems
about performance neutral on average and saves a little memory.
2017-10-04 21:11:14 +02:00
5b7d6ea54b
Code refactor: add WorkTile struct for passing work to kernel.
...
This makes sharing some code between mega/split in following commits a bit
easier, and also paves the way for rendering multiple tiles later.
2017-10-04 21:11:14 +02:00
660e8e59e7
Fix T52645, T52645: AMD OpenCL compiler crash with recent drivers.
...
Work around the bug by reshuffling code.
2017-10-04 21:00:46 +02:00
Bastien Montagne
754f5d39c5
installdeps: Fix building numpy with newer py system...
...
Many thanks to Jens Verwiebe for investigating and finding this fix.
2017-10-04 20:16:39 +02:00
Ray Molenkamp
57d7e5b6ee
Fix T42489 and T52936: Loading blend with minimized window results in crash or empty screen on windows.
...
Reviewed By: @brecht , @sergey
Differential Revision: http://developer.blender.org/D2866
2017-10-04 11:44:22 -06:00
Campbell Barton
296757804c
Fix error copying smoke modifier uv layer
2017-10-05 02:44:02 +11:00
Bastien Montagne
9609828000
Tssst, we use USA english, not Oxford one... :P
2017-10-04 12:42:33 +02:00
Bastien Montagne
a22af1a84b
Fix i18n messages extraction script, and a few more UI messages...
2017-10-04 12:39:06 +02:00
Sergey Sharybin
b22b2c783d
Fix T52982: Join operator with context override crashes Blender 2.79
2017-10-04 15:00:05 +05:00
Campbell Barton
d1dfed206b
Fix T52923: Circle diameter is in fact radius
2017-10-04 18:01:53 +11:00
Campbell Barton
08f728a3e9
Cleanup: redundant casts
2017-10-04 16:52:55 +11:00
Campbell Barton
d3f0fa1277
Docs: remove outdated PyAPI guide
2017-10-04 15:57:43 +11:00
Campbell Barton
92272f56e6
PyAPI: fast keyword parsing for __import__
...
No functional changes.
2017-10-04 15:54:41 +11:00
Campbell Barton
728cffa9ce
PyAPI: fast keyword parsing for bpy.props
...
No functional changes.
2017-10-04 15:13:52 +11:00
Bastien Montagne
ad22d29dac
Some UI/messages fixes.
2017-10-03 21:32:46 +02:00
Campbell Barton
239b38cd64
CMake: avoid naming collision
2017-10-04 01:35:24 +11:00
Campbell Barton
cbe07d980c
Python: bump minimum version to 3.6
2017-10-03 23:50:34 +11:00
Campbell Barton
7ae02c2ddb
Cleanup: remove unused vpaint projection handle
2017-10-03 23:43:30 +11:00
Campbell Barton
a4721c0504
Weight Paint: blend smear tool
2017-10-03 23:35:30 +11:00
Campbell Barton
6a7dc3348a
Vertex Paint: fix feedback loop w/ smear tool
...
Was reading and writing from same array,
this also makes smear thread-safe.
2017-10-03 23:07:13 +11:00
Bastien Montagne
314d93838e
Fix bplayer (c)
2017-10-03 12:31:51 +02:00
Sergey Sharybin
61d5c5a64f
Fix T52981: 2D Curve shapes do not render untill extruded
...
Regression since 9298c53.
2017-10-03 15:29:39 +05:00
Alexander Gavrilov
7c2d5eaacb
Change weigth paint accumulate mechanics to fix problems with mirroring.
...
Restoring weights is problematic when the stroke overlaps its mirror.
It's better to simply compute the new weight based on the saved data
rather than restoring things, and check that the change is monotonic.
This way is also closer to how things worked before the merge.
2017-10-03 12:41:37 +03:00
Campbell Barton
975ed05edb
Vertex Paint: smear got wrong last-location
2017-10-03 20:19:59 +11:00
Campbell Barton
7640d5e92a
Error in last commit, weight must be set for non-accumulate
2017-10-03 19:24:02 +11:00
Campbell Barton
f4b1f9b925
Vertex Paint: remove accumulate (soc-2017-vertex_paint)
...
This secondary accumulation option accumulated brush falloff.
The same option in image painting accumulates color
as vertex paiht 'Spray' does.
Giving this option different behavior for vertex paint seems strange.
Also this is basically increasing falloff over time.
Remove the new code, expose existing 'Spray' as 'Accumulate'
to match other paint modes.
2017-10-03 19:09:28 +11:00
Campbell Barton
515e69999e
Weight Paint: don't set weight w/ average brush
...
Also use 'const' for brush.
2017-10-03 18:43:18 +11:00
Campbell Barton
2b3ff61d80
Correct int division in last commit
2017-10-03 17:36:55 +11:00
Campbell Barton
6092ded127
Weight Paint: add some color blending modes
...
While these mostly make sense for color,
there is no harm in supporting.
2017-10-03 17:24:29 +11:00
Sergey Sharybin
9f8bb5813b
Buildbot: Re-enable cuda binaries, those should work
2017-10-02 20:24:58 +05:00
Alexander Gavrilov
88dd45e1e1
Disable multithreaded weight paint when mirroring is enabled.
...
Mirroring writes to the mirror vertex within the direct vertex
draw function, which violates assumptions required for processing
vertices in parallel.
2017-10-02 17:33:51 +03:00
Campbell Barton
64afba0244
Remove brush versioning (again!)
...
Causes crash, see c133927bb
Merging soc-2017-vertex_paint brought this back by accident.
2017-10-03 00:56:45 +11:00
Sergey Sharybin
64c1a48c50
Fix T52927: Compositor wrong scale when scale size input is connected to complex node
...
The issue here is that we can not read scale from socket when determining
dependent area of interest. This area will depend on current pixel. Now fall
back to more stupid but reliable thing: if scale size input is connected to some
nodes, we use the whole frame as area of interest.
2017-10-02 15:55:52 +05:00
Campbell Barton
b2a10fa615
Vertex Paint: projection options
...
This makes vertex paint match image painting more closely.
- Add falloff shape option sphere/circle
where sphere uses a 3D radius around the cursor and
circle uses a 2D radius (projected), like previous releases.
- Add normal angle option so you can control the falloff.
- Add Cull option, to paint onto faces pointing away.
Disabling normals, culling and using circle falloff
allows you to paint through the mesh.
2017-10-02 22:07:58 +11:00
Campbell Barton
9d34ae6048
Sculpt: calculate plane instead of point,normal
...
No functional changes.
2017-10-02 22:07:57 +11:00
Campbell Barton
1cfee60419
Cleanup: add 'sphere' to sculpt test function name
2017-10-02 22:07:57 +11:00
Campbell Barton
2a744a9f2e
Cleanup: indentation, whitespace
2017-10-02 22:07:57 +11:00
Campbell Barton
deb16defd5
Math Lib: distance to AABB
...
Original code by @mano-wii, modified for general use.
2017-10-02 22:07:39 +11:00
Sergey Sharybin
ff0938870f
Compositor: Fix strict compilation warnings when debug is enabled
2017-10-02 15:18:10 +05:00
Sergey Sharybin
f06a60c25d
Compositor: Fix compilation error with debug enabled
2017-10-02 15:12:04 +05:00
Sergey Sharybin
04704297d2
Fix T52920: Saving Tiff Files type Blender crashes
...
Was only happening for 3 and 1 channel sources.
2017-10-02 13:44:13 +05:00
Campbell Barton
864a53e276
Fix gradient tool crash with recent changes
...
Having gesture automatic free memory isn't practical if this
has it's own allocations.
Add option not to free userdata.
2017-10-02 01:34:51 +11:00
Campbell Barton
9a6aba2a94
Vertex Paint: use squared distance test
...
Avoids sqrt for verts that are later ignored
2017-10-02 00:44:04 +11:00
Campbell Barton
f1e7891588
Vertex Paint: respect 'normal' option
...
Recent changes ignored this option.
2017-10-01 23:53:24 +11:00
f55735e533
CMake: support CUDA 9 toolkit, and automatically disable sm_2x binaries.
...
Fermi cards (GTX 4xx and 5xx) are no longer supported with this version, so
we can keep supporting both CUDA 8 and 9 for a while.
2017-10-01 14:14:53 +02:00