griefith/test2
1
0
Fork 0
Code Issues Pull Requests Actions 1 Packages Projects Releases Wiki Activity
51,918 Commits 1 Branch 0 Tags
d03e7687613ff288e027198ac58df59cd9f1677e
Commit Graph

3854 Commits

Author SHA1 Message Date
Campbell Barton
63ccb26303 Code Cleanup: spelling 2014-01-17 17:35:03 +11:00
Brecht Van Lommel
e9227c76d4 Fix T37706: avoid cycles crash when using a stack that exceeds SVM stack limits. 
This should be pretty rare, the shader in question had many parallel node links
because of copying the nodes many times, which is inefficient to run anyway.
 2014-01-16 22:36:30 +01:00
Jens Verwiebe
1b4ced86b9 Workaround for avx assembler not compiling with vanilla gcc ( with openMP case ) 2014-01-16 19:44:25 +01:00
Thomas Dinges
da523185fb Fix compilation of Cycles AVX kernel with cmake. 2014-01-16 18:32:54 +01:00
Thomas Dinges
de28a4d4b2 Cycles: Add an AVX kernel for CPU rendering. 
* AVX is available on Intel Sandy Bridge and newer and AMD Bulldozer and newer.
* We don't use dedicated AVX intrinsics yet, but gcc auto vectorization gives a 3% performance improvement for Caminandes. Tested on an i5-3570, Linux x64.
* No change for Windows yet, MSVC 2008 does not support AVX.

Reviewed by: brecht
Differential Revision: https://developer.blender.org/D216
 2014-01-16 17:04:11 +01:00
Brecht Van Lommel
d9e52ac98b Code cleanup: move half float functions to separate header file. 2014-01-15 15:29:22 +01:00
Brecht Van Lommel
8af782ad22 Code cleanup: some reshuffling of SIMD defines moving more code to util_optimization.h. 2014-01-15 15:11:50 +01:00
Sergey Sharybin
5cd321203e Fix compilation error with stricg GCC flags 2014-01-15 16:21:53 +06:00
Bastien Montagne
6ba02681de Fix a bunch of UI string issues... 2014-01-15 10:40:54 +01:00
Thomas Dinges
9e3ddd70d4 Cycles: Disable SSE41 kernel on 32bit, we don't use intrinsics here anyway. Also disable it for Visual Studio < 2012, broken blendv instruction. 2014-01-14 23:51:59 +01:00
Martijn Berger
0f3fed2970 OS X linker does not like empty compilation unit by itself in a library. Scons creates one library (.a) per kernel. This fixes that 2014-01-14 22:48:31 +01:00
Thomas Dinges
5d88f7c7db Cycles: Build SSE41 kernel per default, remove build option. This hopefully also fixes some compile errors on various systems. 2014-01-14 22:04:32 +01:00
Martijn Berger
993b946681 DingTo forgot to make sure kernel_sse41 is compiled in even when empty 2014-01-14 21:49:48 +01:00
Thomas Dinges
9351ac0d85 Cycles: Skip the compilation of the dedicated SSE2 kernel on x86-64, we can assume SSE2 here, so just re-use the regular one. Saves 500kb in the blender binary. 
Reviewed by: brecht
Differential Revision: https://developer.blender.org/D199
 2014-01-14 20:39:54 +01:00
Brecht Van Lommel
d980c3eccb Further fix for T37817: non-ascii paths fix in Cycles broke OSL rendering. 
Not quite sure yet what is going on here, but this works for me.
 2014-01-14 20:01:26 +01:00
Sv. Lockal
1c49eb0072 Cycles, Code cleanup: simplify code for color linear interpolation and float math 
Reviewed By: brecht

Differential Revision: https://developer.blender.org/D215
 2014-01-14 22:55:02 +04:00
Brecht Van Lommel
a02a97753b Fix T38192: OSL got incorrectly enabled in preview render when GPU was selected 
for rendering of the main scene.
 2014-01-13 23:29:35 +01:00
Thomas Dinges
e9984653a8 Cycles: Fix Wave texture difference between OSL and SVM, OSL wasn't using the "Scale" properly for distortion. 2014-01-13 22:01:39 +01:00
Thomas Dinges
6b61f7f755 Code cleanup / Cycles: Don't pass scale to texture functions, do the multiplication in the function call already. 2014-01-13 21:17:55 +01:00
Bastien Montagne
75ab57efed Fix T38190: Linux tablet: Issue with XInput/GHOST? 
With edits by Campbell, thanks!

Looks like in some cases (driver dependent?), `XDeviceMotionEvent` get generated with only part of expected data
(e.g. only x coordinate, only pressure, etc.), data which did not change since last event being NULL.
We know which data to actually handle with `XDeviceMotionEvent.first_axis` and `XDeviceMotionEvent.axes_count` values.

Reviewed by: campbellbarton

Differential Revision: https://developer.blender.org/D208
 2014-01-13 17:51:17 +01:00
Sv. Lockal
9cf6946d31 Fix cycles texture crash on win x86-64 + msvc 11 
Use union for __m128 aliasing; while gcc supports no-strict-aliasing attribute, unions are the most common way to deal with __m128 in msvc.
 2014-01-13 18:31:02 +04:00
Marcus von Appen
f0fb60f8c9 Fix cycles standalone compile on some systems, CMAKE_DL_LIBS should be enough. 
Reviewed By: brecht

Differential Revision: https://developer.blender.org/D205
 2014-01-13 10:36:02 +01:00
Antony Riakiotakis
717bf85545 Fix some harmless warnings that mostly appeared on MinGW64 2014-01-13 04:28:51 +02:00
Sv. Lockal
d6c022d6d7 Fix compilation for OpenCL (and small stype fixes) 2014-01-12 18:18:43 +04:00
Sv. Lockal
47c5898fa1 Cycles: SSE for Voronoi textures (targeted for Haswell CPUs) 
Gives up to 15% speedup scenes with voronoi-based textures (up to 25% with volumes) on Haswell. The performance change for other CPUs is much smaller: 1-2%.

Reviewed By: brecht

Differential Revision: https://developer.blender.org/D203
 2014-01-12 18:14:00 +04:00
Sv. Lockal
da3fdf0b4b Code Cleanup: in Cycles SSE replace macros with templates, skip unused code with preprocessor, simplify casts 2014-01-11 22:20:03 +04:00
Sv. Lockal
b886c26d1f Cycles: mix hair minimum width code with SSE intersection code 
Gives 6.5% speedup for hair.blend from testsuite.

Reviewed By: brecht

Differential Revision: https://developer.blender.org/D197
 2014-01-11 20:47:30 +04:00
Campbell Barton
50650e28ea Events: support for buttons 6 & 7 (some trackballs have these) X11 only 
Patch by Marcus von Appen

Note: this patch makes ISMOUSE accept INBETWEEN_MOUSEMOVE as a mouse
event where before it didnt.
 2014-01-12 00:40:23 +11:00
Campbell Barton
1bfa64895e Ghost: update multitest for changes in BLF 2014-01-11 23:53:23 +11:00
Brecht Van Lommel
4d72a5e34a Fix T38129: cycles viewport render display with very bright colors turning black. 
This happened when exceeding the maximum value representable by half floats.
 2014-01-11 00:50:53 +01:00
Brecht Van Lommel
241fccaf6a Fix T37817: cycles CUDA detection problem on Windows with non-ascii paths. 2014-01-11 00:47:58 +01:00
Brecht Van Lommel
79e040ba95 Fix T38143: pressing F12 key in the text editor on OS X would type an invalid character. 2014-01-10 19:41:01 +01:00
Brecht Van Lommel
8c0f9365c0 Fix T38134: missing cycles update when removing world volume scatter shader. 2014-01-09 01:26:43 +01:00
Brecht Van Lommel
2d073bbf21 Fix T38134: cycles viewport render not updating when adding a world volume shader. 2014-01-08 23:12:54 +01:00
Sv. Lockal
20b046d763 Cycles: workaround for noise performance regression in CUDA 5.5 
Use manual ternary operation widening in grad(). Without it nvcc 5.5 produces multiple branch splits with very big branches (because of inlining). This solves 19% performance regression for BMW1M-MikePan.blend.

Also remove one redundant instruction in perlin SSE (when h == 12 or h == 14, then h is always >= 4).

Reviewed By: brecht

Differential Revision: https://developer.blender.org/D190
 2014-01-08 22:25:55 +04:00
Thomas Dinges
8094ac9194 Cycles: Fix World Homogeneous volume setting, had no influence. 
Thanks to "storm" for noticing this!
 2014-01-08 11:00:06 +01:00
Thomas Dinges
073be286b4 Cycles: Update Integrator Presets and Total Samples info for latest changes. 2014-01-07 15:55:32 +01:00
Brecht Van Lommel
7b0a46b1ff Fix CUDA/OpenCL compile errors in scattering commit. 2014-01-07 15:48:04 +01:00
Brecht Van Lommel
01df756bd1 Cycles Volume Render: scattering support. 
This is done by adding a Volume Scatter node. In many cases you will want to
add together a Volume Absorption and Volume Scatter node with the same color
and density to get the expected results.

This should work with branched path tracing, mixing closures, overlapping
volumes, etc. However there's still various optimizations needed for sampling.
The main missing thing from the volume branch is the equiangular sampling for
homogeneous volumes.

The heterogeneous scattering code was arranged such that we can use a single
stratified random number for distance sampling, which gives less noise than
pseudo random numbers for each step. For volumes where the color is textured
there still seems to be something off, needs to be investigated.
 2014-01-07 15:03:41 +01:00
Campbell Barton
fee66f7bc8 Code cleanup: defines for statfs were getting out of hand for BSD's. 
add __DragonFly__ and internal defines to avoid copy-pasting checks.

also remove __CYGWIN32__ check, since cygwin is no longer supported.
 2014-01-07 13:39:00 +11:00
Martijn Berger
62c8bae872 Added scons option to actually build the sse41 kernel. CMake already has it 2014-01-06 21:38:56 +01:00
Thomas Dinges
5022d9f81b Cycles: Don't include SIMD util file for OpenCL/CUDA, this fixes OpenCL compilation. 2014-01-06 21:28:18 +01:00
Jens Verwiebe
a0b424aa4c Take back last header copy, due it is for native only, must be a runtime solution, todo: do by definitions 2014-01-06 20:43:54 +01:00
Jens Verwiebe
48d8faeb79 Cmake: fix kernelcompile after introduction of util_simd.h 2014-01-06 20:26:02 +01:00
Sv. Lockal
acc90b40bf Cycles: Minor optimization (~1%) for texture access on CPU 2014-01-06 22:05:31 +04:00
Sv. Lockal
4817da0df4 Cleanup: use blend() in perlin noise (gives 12 less instructions on SSE4.1) 2014-01-06 21:24:28 +04:00
Sv. Lockal
96903508bc Cycles: SSE optimization for sRGB conversion (gives 7% speedup on CPU for pavillon_barcelone scene) 
Thanks brecht/dingto/juicyfruit et al. for testing and reviewing this patch in T38034.
 2014-01-06 20:03:30 +04:00
Campbell Barton
64fc94e93f Code Cleanup: osl style 2014-01-06 13:58:33 +11:00
Thomas Dinges
ce6dce3b13 Code cleanup / Cycles: else/if for SSE41 kernel functions. 2014-01-06 03:22:14 +01:00
Campbell Barton
c3bc2fd941 CMake: cleanup and add include 2014-01-04 13:17:07 +11:00