[Bf-blender-cvs] SVN commit: /data/svn/bf-blender [61245] trunk/lib/darwin-9.x.universal/osl : OSX/OSL: update to v 1.4
jens verwiebe
info at jensverwiebe.de
Tue Dec 3 17:02:12 CET 2013
Revision: 61245
http://projects.blender.org/scm/viewvc.php?view=rev&root=bf-blender&revision=61245
Author: jensverwiebe
Date: 2013-12-03 16:02:11 +0000 (Tue, 03 Dec 2013)
Log Message:
-----------
OSX/OSL: update to v 1.4
Modified Paths:
--------------
trunk/lib/darwin-9.x.universal/osl/CHANGES
trunk/lib/darwin-9.x.universal/osl/bin/oslc
trunk/lib/darwin-9.x.universal/osl/bin/oslinfo
trunk/lib/darwin-9.x.universal/osl/doc/osl-languagespec.pdf
trunk/lib/darwin-9.x.universal/osl/include/OSL/dual_vec.h
trunk/lib/darwin-9.x.universal/osl/include/OSL/oslclosure.h
trunk/lib/darwin-9.x.universal/osl/include/OSL/oslexec.h
trunk/lib/darwin-9.x.universal/osl/include/OSL/oslversion.h
trunk/lib/darwin-9.x.universal/osl/lib/liboslcomp.a
trunk/lib/darwin-9.x.universal/osl/lib/liboslexec.a
trunk/lib/darwin-9.x.universal/osl/lib/liboslquery.a
trunk/lib/darwin-9.x.universal/osl/shaders/emitter.oso
trunk/lib/darwin-9.x.universal/osl/shaders/glass.oso
trunk/lib/darwin-9.x.universal/osl/shaders/image.oso
trunk/lib/darwin-9.x.universal/osl/shaders/matte.oso
trunk/lib/darwin-9.x.universal/osl/shaders/metal.oso
trunk/lib/darwin-9.x.universal/osl/shaders/stdosl.h
trunk/lib/darwin-9.x.universal/osl/shaders/ubersurface.oso
Modified: trunk/lib/darwin-9.x.universal/osl/CHANGES
===================================================================
--- trunk/lib/darwin-9.x.universal/osl/CHANGES 2013-12-03 15:59:25 UTC (rev 61244)
+++ trunk/lib/darwin-9.x.universal/osl/CHANGES 2013-12-03 16:02:11 UTC (rev 61245)
@@ -1,3 +1,161 @@
+Release 1.4.0 - 25 November 2013 (compared to 1.3)
+--------------------------------------------------
+
+Language, standard libary, and compiler changes (for shader writers):
+
+* Simplex noise: new noise varieties "simplex" and "usimplex" are
+ faster than Perlin noise (and MUCH faster for the 4-D variety),
+ but has a somewhat different appearance. Judge for yourself, use
+ whichever noise suits your purpose.
+* oslc now catches mismatches between the format string and the types
+ of arguments to format(), printf(), fprintf(), warning(), and error().
+* oslc now gives a compile error when trying to use a function as
+ if it were an ordinary variable. (1.3.2)
+* Improvement to oslc function type checking -- allow more nuanced
+ matching of polymorphism based on return types. First we check for
+ identical return types, then we test allowing non-identical spatial
+ triples to match, then finally we allow any return type to match.
+* New ShadingSystem attribute "debug_uninit", which when set to nonzero
+ tries to detect use of uninitialized variables at runtime. (Off by
+ default because it's very expensive, just use if you are debugging.)
+* Implemented the OSL aastep() function, which was always described in
+ the docs but never implemented.
+* The texture lookup functions all take new optional parameters,
+ "missingtexture", (color), "missingalpha", (float), which when supplied
+ will substitute that color and alpha for missing or broken textures,
+ rather than treating it as an error condition.
+* New gettextureinfo() queries: "datawindow", "displaywindow" (both take
+ an int[4] as destination argument).
+
+ShadingSystem API changes and new options (for renderer writers):
+
+* New ShadingSystem attribute "renderer_outputs" enumerates the renderer
+ outputs (AOVs) explicitly, so that elision of unused outputs can be done
+ with full understanding of which params will or will not be sued.
+ Beware: if you enable "opt_elide_unconnected_outputs", you MUST pass
+ "renderer_outputs" or your AOVs may be eliminated!
+
+Performance improvements:
+
+* "Middleman" optimization notices when you have 3 layers A,B,C with
+ connections A.Aout -> B.Bin and B.Bout -> C.Cin, an Bout is simply
+ copied unconditionally from Bin. In this case, Aout is connected
+ directly to Cin, potentially leading to further optimizations, including
+ layer B getting culled entirely.
+* Layer-to-layer global propagation: Notice cases where an upstream
+ output is guaranteed to be aassigned a global variable (such as N or u)
+ and then alias the downstream parameter to that global (eliminating
+ the connection and possibly culling the upstream layer entirely).
+* Improve constant propagation from layer to layer (it used to do so only
+ if the output of the earlier layer was never used in the upstream layer;
+ but actually it's ok for it to be read, as long as it's not rewritten).
+* Runtime constant-folding of cellnoise() calls.
+* Local variables that are arrays initialized with float, int, or string
+ constants have been greatly optimized. If that's the only time anything
+ is written into the array, it will be turned into a true static constant
+ with no per-shader-invocation initialization or allocation costs.
+ Thus, it is no longer expensive to use a large array of constants as a
+ table... it may look like a per-shade cost to set up the array, but by
+ the time the optimizer gets done with it, it is not.
+* Improved runtime constant-folding of max/min (now works for the integer
+ case, as well as previously working for the float-based types).
+* Runtime constant-folding of cos, sin, acos, asin, normalize, inversesqrt,
+ exp, exp2, expm1, log, log2, log10, logb, erf, erfc, radians, degrees.
+* Optimization that speeds up certain closure manipulations and can
+ cut overhead especially for relatively simple shaders.
+* New peephole optimizations for two common patterns:
+ (1) add A B C ; sub D A C => [add A B C;] assign D B
+ and the same if the order of add & sub are switched.
+ (2) OP A B... ; assign A C => OP C B... as long as A is not subsequently
+ used and OP fully overwrites A and writes to no other variables.
+* Constant folding of Dx, Dy, Dz, area, and filterwidth, when passed a
+ constant (the deriv will be 0 in that case).
+* Substantial multithread speedups by eliminating testing of some shared
+ variables.
+* Constant folding of add & sub for the mixed float/triple cases.
+* Speed up closure construction by combining the closure primitive with
+ its (nearly inevitable) multiplication by weight, into a single internal
+ operation.
+* The OSLQuery library has been sped up by 20x when querying the parameters
+ of large shaders.
+
+Bug fixes:
+
+* Fix 'debugnan' tendency to give false positives for certain assignments
+ to arrays and multi-component variables (colors, points, etc.).
+* The integer versions of min(), max(), and clamp() were not working
+ properly. (1.3.1)
+* fmod(triple,triple,float) was not working properly in Windows. (1.3.2)
+* Avoid incorrect heap addressing when retrieving a symbol's address. (1.3.2)
+* Fix for a bug where an early 'return' from a shader could lead to
+ output parameters not being properly copied to the inputs of their
+ downstream connections. (1.3.2)
+* Fix bug where filterwidth() didn't properly mark itself as needing
+ derivatives of its argument. (1.3.2)
+* Bug fix: Was not properly initializing the default values of output
+ params if they were written but never subsequently read.
+* Better error handling for missing pointclouds.
+* Fixed improper initialization of arrays of closures.
+* Fixed constant folding bug for the split() function.
+* Fixed constant folding bug for pow(x,2.0).
+
+Under the hood:
+
+* Eliminate use of custom DEBUG preprocessor symbol in favor of the more
+ standard NDEBUG (beware: its sense is reversed).
+* All references to and vestiges of TBB have been removed.
+
+Build & test system improvements and developer goodies:
+
+* testshade: --options lets you set arbitrary options to the ShadingSystem
+ for test runs.
+* Added CMake option 'ENABLERTTI' that enables use of RTTI for sites that
+ must link against an RTTI-enabled LLVM (the default is to assume LLVM
+ is built without RTTI, so we don't use it either). (1.3.1)
+* Fix some ambiguity about which variety of shared_ptr we are using --
+ this could produce problems in apps that include the OSL headers but also
+ use different shared_ptr varieties.
+* Fix a variety of build issues and warnings under Windows.
+* Work around CMake bug where on some platforms, CMake doesn't define the
+ 'NDEBUG' when making the 'RelWithDebInfo' target (which is sometimes
+ used for profiling). (1.3.1)
+* Add Make/CMake option 'USE_EXTERNAL_PUGIXML' which if set to nonzero
+ will find and use a system-installed PugiXML rather than assuming that
+ it's included in libOpenImageIO. When used, it will also use the
+ environment variable PUGIXML_HOME (if set) as a hint for where to find
+ it, in case it's not in a usual system library directory. (1.3.1)
+* Nicer debug printing (for developers) when an op doesn't have an
+ associated source line. (1.3.2)
+* Reorder include directories for building LLVM bitcode so that the
+ openimageio, ilmbase and boost directories are ahead of any system
+ directories given by llvm-config --cxxflags. This avoids the LLVM
+ bitcode getting compiled against different library versions than the
+ rest of the code if multiple versions are installed.
+* Have debug_groupname and debug_layername apply to LLVM debugging info
+ as well.
+* Support for LLVM 3.3.
+* Fix compiler warnings for clang 3.3 and libc++.
+* CMake files now more carefully quote assembled file paths to more
+ gracefully handle paths with spaces in them.
+* CMakeLists.txt has been moved to the top level (not in src/) to conform
+ to the usual CMake conventions.
+
+Documentation:
+
+* Eliminate references to OSL random() function; it was never
+ implemented and nobody seemed to miss it or have a compelling use
+ case. (1.3.1)
+
+
+
+
+Release 1.3.3 - 11 July 2013 (compared to 1.3.2)
+------------------------------------------------
+* Fix bug in the implementation of filterwidth() when passed a constant
+ (for which no derivative could be taken).
+* Changes to support LLVM 3.3.
+
+
Release 1.3.2 - 19 Jun 2013 (compared to 1.3.1)
-----------------------------------------------
* fmod(triple,triple,float) was not working properly in Windows.
@@ -32,7 +190,6 @@
it, in case it's not in a usual system library directory.
-
Release 1.3.0 - 14 Feb 2013 (compared to 1.2)
----------------------------------------------
Modified: trunk/lib/darwin-9.x.universal/osl/bin/oslc
===================================================================
(Binary files differ)
Modified: trunk/lib/darwin-9.x.universal/osl/bin/oslinfo
===================================================================
(Binary files differ)
Modified: trunk/lib/darwin-9.x.universal/osl/doc/osl-languagespec.pdf
===================================================================
(Binary files differ)
Modified: trunk/lib/darwin-9.x.universal/osl/include/OSL/dual_vec.h
===================================================================
--- trunk/lib/darwin-9.x.universal/osl/include/OSL/dual_vec.h 2013-12-03 15:59:25 UTC (rev 61244)
+++ trunk/lib/darwin-9.x.universal/osl/include/OSL/dual_vec.h 2013-12-03 16:02:11 UTC (rev 61245)
@@ -353,18 +353,7 @@
inline Dual2<float>
distance (const Dual2<Vec3> &a, const Dual2<Vec3> &b)
{
- Dual2<float> ax (a.val().x, a.dx().x, a.dy().x);
- Dual2<float> ay (a.val().y, a.dx().y, a.dy().y);
- Dual2<float> az (a.val().z, a.dx().z, a.dy().z);
- Dual2<float> bx (b.val().x, b.dx().x, b.dy().x);
- Dual2<float> by (b.val().y, b.dx().y, b.dy().y);
- Dual2<float> bz (b.val().z, b.dx().z, b.dy().z);
-
- Dual2<float> dx = bx - ax;
- Dual2<float> dy = by - ay;
- Dual2<float> dz = bz - az;
-
- return sqrt(dx*dx + dy*dy + dz*dz);
+ return length (a - b);
}
@@ Diff output truncated at 10240 characters. @@
More information about the Bf-blender-cvs
mailing list