sandey/blender
1
0
Fork 0
forked from bartvdbraak/blender
Code Pull Requests Activity
8bf9bf37ec
blender/intern/cycles/CMakeLists.txt

118 lines
3.7 KiB
CMake
Raw Normal View History

Cycles render engine, initial commit. This is the engine itself, blender modifications and build instructions will follow later. Cycles uses code from some great open source projects, many thanks them: * BVH building and traversal code from NVidia's "Understanding the Efficiency of Ray Traversal on GPUs": http://code.google.com/p/understanding-the-efficiency-of-ray-traversal-on-gpus/ * Open Shading Language for a large part of the shading system: http://code.google.com/p/openshadinglanguage/ * Blender for procedural textures and a few other nodes. * Approximate Catmull Clark subdivision from NVidia Mesh tools: http://code.google.com/p/nvidia-mesh-tools/ * Sobol direction vectors from: http://web.maths.unsw.edu.au/~fkuo/sobol/ * Film response functions from: http://www.cs.columbia.edu/CAVE/software/softlib/dorf.php
2011-04-27 11:58:34 +00:00
Cycles: Add a cmake config to easily compile Cycles Standalone. On Linux/Mac OS X, simply type "make cycles" inside the Blender source directory, to get a standalone build of the engine. Reviewed by: Brecht Differential Revision: https://developer.blender.org/D228
2014-01-20 19:41:54 +00:00
# Standalone or with Blender
if(NOT WITH_BLENDER AND WITH_CYCLES_STANDALONE)
set(CYCLES_INSTALL_PATH "")
else()
set(WITH_CYCLES_BLENDER ON)
set(CYCLES_INSTALL_PATH "scripts/addons/cycles")
endif()
Cycles: * add some (disabled) test code for using OpenImageIO in imbuf * link cycles, openimageio and boost into blender instead of a shared library * some cmakefile changes to simplify the code and follow conventions better * this may solve running cycles problems on windows XP, or give a different and hopefully more useful error message
2011-08-16 16:15:34 +00:00
# External Libraries
include(cmake/external_libs.cmake)
# Build Flags
Cycles * Compile all of cycles with -ffast-math again * Add scons compilation of cuda binaries, tested on mac/linux. * Add UI option for supported/experimental features, to make it more clear what is supported, opencl/subdivision is experimental. * Remove cycles xml exporter, was just for testing.
2011-12-01 16:33:21 +00:00
if(WIN32 AND MSVC)
Cycles: attempt to fix internal compile error with some visual studio builds
2013-06-18 13:19:16 +00:00
# there is no /arch:SSE3, but intrinsics are available anyway
Cleanup: * Silence /arch:SSE2 warning on msvc x64.
2013-08-31 03:18:55 +00:00
if(CMAKE_CL_64)
set(CYCLES_SSE2_KERNEL_FLAGS "/fp:fast -D_CRT_SECURE_NO_WARNINGS /Gs-")
set(CYCLES_SSE3_KERNEL_FLAGS "/fp:fast -D_CRT_SECURE_NO_WARNINGS /Gs-")
Cycles: test code for sse 4.1 kernel and alignment for some vector types. This is mostly work towards enabling the __KERNEL_SSE__ option to start using SIMD operations for vector math operations. This 4.1 kernel performes about 8% faster with that option but overall is still slower than without the option. WITH_CYCLES_OPTIMIZED_KERNEL_SSE41 is the cmake flag for testing this kernel. Alignment of int3, int4, float3, float4 to 16 bytes seems to give a slight 1-2% speedup on tested systems with the current kernel already, so is enabled now.
2013-11-22 13:16:47 +00:00
set(CYCLES_SSE41_KERNEL_FLAGS "/fp:fast -D_CRT_SECURE_NO_WARNINGS /Gs-")
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 16:04:11 +00:00
set(CYCLES_AVX_KERNEL_FLAGS "/fp:fast -D_CRT_SECURE_NO_WARNINGS /Gs-") #/arch:AVX for VC2012 and above
Cleanup: * Silence /arch:SSE2 warning on msvc x64.
2013-08-31 03:18:55 +00:00
else()
set(CYCLES_SSE2_KERNEL_FLAGS "/arch:SSE2 /fp:fast -D_CRT_SECURE_NO_WARNINGS /Gs-")
set(CYCLES_SSE3_KERNEL_FLAGS "/arch:SSE2 /fp:fast -D_CRT_SECURE_NO_WARNINGS /Gs-")
Cycles: test code for sse 4.1 kernel and alignment for some vector types. This is mostly work towards enabling the __KERNEL_SSE__ option to start using SIMD operations for vector math operations. This 4.1 kernel performes about 8% faster with that option but overall is still slower than without the option. WITH_CYCLES_OPTIMIZED_KERNEL_SSE41 is the cmake flag for testing this kernel. Alignment of int3, int4, float3, float4 to 16 bytes seems to give a slight 1-2% speedup on tested systems with the current kernel already, so is enabled now.
2013-11-22 13:16:47 +00:00
set(CYCLES_SSE41_KERNEL_FLAGS "/arch:SSE2 /fp:fast -D_CRT_SECURE_NO_WARNINGS /Gs-")
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 16:04:11 +00:00
set(CYCLES_AVX_KERNEL_FLAGS "/arch:SSE2 /fp:fast -D_CRT_SECURE_NO_WARNINGS /Gs-") #/arch:AVX for VC2012 and above
Cleanup: * Silence /arch:SSE2 warning on msvc x64.
2013-08-31 03:18:55 +00:00
endif()
Cycles: optimized SSE BVH traversal now also works with SSE2 CPUs, so all the way back to Pentium 4, using a slightly less efficient instruction. Also ensure /Ox is used for Visual Studio for RelWithDebInfo builds.
2013-06-19 17:54:26 +00:00
set(CMAKE_CXX_FLAGS "${CMAKE_CXX_FLAGS} /fp:fast -D_CRT_SECURE_NO_WARNINGS /Gs-")
Fix for Debug build on MSVC2008 and MSVC2012 after Cycles Kernel optimization: Compiler optimization was accidentally set to /Ox for debug build too. Changed this to be /Od in Debug and /Ox in Release mode.
2013-06-19 17:17:51 +00:00
set(CMAKE_CXX_FLAGS_RELEASE "${CMAKE_CXX_FLAGS_RELEASE} /Ox")
Fix wrong cycles cmake + msvc build flags with relwithdebinfo and minsizerel. Patch by Karsten Schwenk.
2013-08-05 12:02:43 +00:00
set(CMAKE_CXX_FLAGS_RELWITHDEBINFO "${CMAKE_CXX_FLAGS_RELWITHDEBINFO} /Ox")
set(CMAKE_CXX_FLAGS_MINSIZEREL "${CMAKE_CXX_FLAGS_MINSIZEREL} /Ox")
Cycles * Compile all of cycles with -ffast-math again * Add scons compilation of cuda binaries, tested on mac/linux. * Add UI option for supported/experimental features, to make it more clear what is supported, opencl/subdivision is experimental. * Remove cycles xml exporter, was just for testing.
2011-12-01 16:33:21 +00:00
elseif(CMAKE_COMPILER_IS_GNUCC)
Fix cycles intersection issue with overlapping faces on windows 32 bit and CPU without SSE3 support, due to 80 bit precision float register being used for one bounding box but not the one next to it.
2013-02-04 16:12:37 +00:00
set(CYCLES_SSE2_KERNEL_FLAGS "-ffast-math -msse -msse2 -mfpmath=sse")
Cycles / BVH: * GCC needs -mssse3 compile flag too (4.7.2) to compile.
2013-06-18 10:05:47 +00:00
set(CYCLES_SSE3_KERNEL_FLAGS "-ffast-math -msse -msse2 -msse3 -mssse3 -mfpmath=sse")
Cycles: test code for sse 4.1 kernel and alignment for some vector types. This is mostly work towards enabling the __KERNEL_SSE__ option to start using SIMD operations for vector math operations. This 4.1 kernel performes about 8% faster with that option but overall is still slower than without the option. WITH_CYCLES_OPTIMIZED_KERNEL_SSE41 is the cmake flag for testing this kernel. Alignment of int3, int4, float3, float4 to 16 bytes seems to give a slight 1-2% speedup on tested systems with the current kernel already, so is enabled now.
2013-11-22 13:16:47 +00:00
set(CYCLES_SSE41_KERNEL_FLAGS "-ffast-math -msse -msse2 -msse3 -mssse3 -msse4.1 -mfpmath=sse")
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 16:04:11 +00:00
set(CYCLES_AVX_KERNEL_FLAGS "-ffast-math -msse -msse2 -msse3 -mssse3 -msse4.1 -mavx -mfpmath=sse")
Cycles * Compile all of cycles with -ffast-math again * Add scons compilation of cuda binaries, tested on mac/linux. * Add UI option for supported/experimental features, to make it more clear what is supported, opencl/subdivision is experimental. * Remove cycles xml exporter, was just for testing.
2011-12-01 16:33:21 +00:00
set(CMAKE_CXX_FLAGS "${CMAKE_CXX_FLAGS} -ffast-math")
Arrempt to fix Cycles compilation with Clang Issue is caused by missing sse flags for Clang compilers, this flags only was set for GNU C compilers. Added if branch for Clang now, which contains the same flags apart from -mfpmath=sse, This is because Clang was claiming it's unused argument. Probably OSX would need some further checks since it's also using Clang. I've got no idea why it could have worked for OSX before..
2013-07-07 18:29:57 +00:00
elseif(CMAKE_CXX_COMPILER_ID MATCHES "Clang")
set(CYCLES_SSE2_KERNEL_FLAGS "-ffast-math -msse -msse2")
set(CYCLES_SSE3_KERNEL_FLAGS "-ffast-math -msse -msse2 -msse3 -mssse3")
Cycles: test code for sse 4.1 kernel and alignment for some vector types. This is mostly work towards enabling the __KERNEL_SSE__ option to start using SIMD operations for vector math operations. This 4.1 kernel performes about 8% faster with that option but overall is still slower than without the option. WITH_CYCLES_OPTIMIZED_KERNEL_SSE41 is the cmake flag for testing this kernel. Alignment of int3, int4, float3, float4 to 16 bytes seems to give a slight 1-2% speedup on tested systems with the current kernel already, so is enabled now.
2013-11-22 13:16:47 +00:00
set(CYCLES_SSE41_KERNEL_FLAGS "-ffast-math -msse -msse2 -msse3 -mssse3 -msse4.1")
Fix compilation of Cycles AVX kernel with cmake.
2014-01-16 17:32:39 +00:00
set(CYCLES_AVX_KERNEL_FLAGS "-ffast-math -msse -msse2 -msse3 -mssse3 -msse4.1 -mavx")
Arrempt to fix Cycles compilation with Clang Issue is caused by missing sse flags for Clang compilers, this flags only was set for GNU C compilers. Added if branch for Clang now, which contains the same flags apart from -mfpmath=sse, This is because Clang was claiming it's unused argument. Probably OSX would need some further checks since it's also using Clang. I've got no idea why it could have worked for OSX before..
2013-07-07 18:29:57 +00:00
set(CMAKE_CXX_FLAGS "${CMAKE_CXX_FLAGS} -ffast-math")
edits to cycles cmake files so cmake_consistency_check.py can parse them.
2011-11-08 20:27:37 +00:00
endif()
Cycles: hook up the CMake build system. New build instructions for Ubuntu Linux in the wiki: http://wiki.blender.org/index.php/Dev:2.5/Source/Cycles
2011-04-28 13:47:27 +00:00
RTTI needs to be disabled in cycles for OSL.
2012-09-03 11:38:20 +00:00
# for OSL
-fno-rtti is a gcc option which gives a warning when using MSVC
2012-11-26 13:48:48 +00:00
if(WIN32 AND MSVC)
set(RTTI_DISABLE_FLAGS "/GR- -DBOOST_NO_RTTI -DBOOST_NO_TYPEID")
Fix compile error with clang. Patch by sambler via irc
2012-11-29 12:22:07 +00:00
elseif(CMAKE_COMPILER_IS_GNUCC OR (CMAKE_C_COMPILER_ID MATCHES "Clang"))
-fno-rtti is a gcc option which gives a warning when using MSVC
2012-11-26 13:48:48 +00:00
set(RTTI_DISABLE_FLAGS "-fno-rtti -DBOOST_NO_RTTI -DBOOST_NO_TYPEID")
endif()
Cycles: * Fix excessive fireflies in Velvet BSDF (patch by David). * Disable some unused SSE code * Remove RTTI disabling flags for now, this is giving some compile issues and was only needed of OSL which we're not using yet.
2011-11-10 14:32:16 +00:00
Cycles: * add some (disabled) test code for using OpenImageIO in imbuf * link cycles, openimageio and boost into blender instead of a shared library * some cmakefile changes to simplify the code and follow conventions better * this may solve running cycles problems on windows XP, or give a different and hopefully more useful error message
2011-08-16 16:15:34 +00:00
# Definitions and Includes
add_definitions(${BOOST_DEFINITIONS} ${OPENIMAGEIO_DEFINITIONS})
add_definitions(-DCCL_NAMESPACE_BEGIN=namespace\ ccl\ {)
add_definitions(-DCCL_NAMESPACE_END=})
if(WITH_CYCLES_NETWORK)
add UNLIKELY / LIKELY util defines to help optimize iterators that are called a lot, only defined for gcc at the moment. also some cmake formatting edits.
2011-12-29 01:46:58 +00:00
add_definitions(-DWITH_NETWORK)
Cycles: hook up the CMake build system. New build instructions for Ubuntu Linux in the wiki: http://wiki.blender.org/index.php/Dev:2.5/Source/Cycles
2011-04-28 13:47:27 +00:00
endif()
Cycles Standalone: GUI flag was not passed correctly to the code, own mistake when I added the GUI build option.
2014-01-20 20:12:28 +00:00
if(WITH_CYCLES_STANDALONE_GUI)
add_definitions(-DWITH_CYCLES_STANDALONE_GUI)
endif()
Cycles: experimental OSL ptex reading code. This code can't actually be enabled for building and is incomplete, but it's here because we know we want to support this at some point and there's not much reason to have it in a separate branch if a simple #ifdef can disable it.
2013-11-28 00:38:23 +00:00
if(WITH_CYCLES_PTEX)
add_definitions(-DWITH_PTEX)
endif()
Cycles: experimental OpenSubdiv code. This code can't actually be enabled for building and is incomplete, but it's here because we know we want to support this at some point and there's not much reason to have it in a separate branch if a simple #ifdef can disable it.
2013-11-28 01:10:02 +00:00
if(WITH_CYCLES_OPENSUBDIV)
add_definitions(-DWITH_OPENSUBDIV)
endif()
Cycles: * add some (disabled) test code for using OpenImageIO in imbuf * link cycles, openimageio and boost into blender instead of a shared library * some cmakefile changes to simplify the code and follow conventions better * this may solve running cycles problems on windows XP, or give a different and hopefully more useful error message
2011-08-16 16:15:34 +00:00
if(WITH_CYCLES_OSL)
add UNLIKELY / LIKELY util defines to help optimize iterators that are called a lot, only defined for gcc at the moment. also some cmake formatting edits.
2011-12-29 01:46:58 +00:00
add_definitions(-DWITH_OSL)
Fix cycles windows link errors when building with OSL master branch.
2013-01-30 21:58:47 +00:00
add_definitions(-DOSL_STATIC_LIBRARY)
Cycles OSL: some build system tweaks to avoid global includes and definitions, which ensures there is no conflict with other libraries, and avoids full rebuild when toggling OSL on/off.
2012-10-20 12:17:45 +00:00
include_directories(${OSL_INCLUDES})
Cycles: * add some (disabled) test code for using OpenImageIO in imbuf * link cycles, openimageio and boost into blender instead of a shared library * some cmakefile changes to simplify the code and follow conventions better * this may solve running cycles problems on windows XP, or give a different and hopefully more useful error message
2011-08-16 16:15:34 +00:00
endif()
Cycles: * Fix crash in light path node * Fix struct alignment issue for cuda * Fix issue with instances taking up too much memory * Fix issue with ray visibility working incorrect on some objects * Enable OpenCL always and remove option, it has no dependencies so may as well * Refuse to load kernel if OpenCL version < 1.1, recent drivers are needed * Better error handling for OpenCL device * 3D views with rendered draw mode will now revert to wireframe on file load
2011-09-02 14:55:06 +00:00
add_definitions(-DWITH_OPENCL)
Cycles: * OpenCL now only uses GPU/Accelerator devices, it's only confusing if CPU device is used, easy to enable in the code for debugging. * OpenCL kernel binaries are now cached for faster startup after the first time compiling. * CUDA kernels can now be compiled and cached at runtime if the CUDA toolkit is installed. This means that even if the build does not have CUDA enabled, it's still possible to use it as long as you install the toolkit.
2011-09-09 12:04:39 +00:00
add_definitions(-DWITH_CUDA)
Cycles: * Fix missing update when editing objects with emission materials. * Fix preview pass rendering set to 1 not showing full resolution. * Fix CUDA runtime compiling failing due to missing cache directory. * Use settings from first render layer for visibility and material override. And a bunch of incomplete and still disabled code mostly related to closure sampling.
2011-09-12 13:13:56 +00:00
add_definitions(-DWITH_MULTI)
Cycles render engine, initial commit. This is the engine itself, blender modifications and build instructions will follow later. Cycles uses code from some great open source projects, many thanks them: * BVH building and traversal code from NVidia's "Understanding the Efficiency of Ray Traversal on GPUs": http://code.google.com/p/understanding-the-efficiency-of-ray-traversal-on-gpus/ * Open Shading Language for a large part of the shading system: http://code.google.com/p/openshadinglanguage/ * Blender for procedural textures and a few other nodes. * Approximate Catmull Clark subdivision from NVidia Mesh tools: http://code.google.com/p/nvidia-mesh-tools/ * Sobol direction vectors from: http://web.maths.unsw.edu.au/~fkuo/sobol/ * Film response functions from: http://www.cs.columbia.edu/CAVE/software/softlib/dorf.php
2011-04-27 11:58:34 +00:00
Cycles: * add some (disabled) test code for using OpenImageIO in imbuf * link cycles, openimageio and boost into blender instead of a shared library * some cmakefile changes to simplify the code and follow conventions better * this may solve running cycles problems on windows XP, or give a different and hopefully more useful error message
2011-08-16 16:15:34 +00:00
include_directories(
code cleanup: cycles now uses system includes for boost/oiio.. etc, so we dont get warnings from system headers.
2012-09-20 09:04:43 +00:00
SYSTEM
Cycles: * add some (disabled) test code for using OpenImageIO in imbuf * link cycles, openimageio and boost into blender instead of a shared library * some cmakefile changes to simplify the code and follow conventions better * this may solve running cycles problems on windows XP, or give a different and hopefully more useful error message
2011-08-16 16:15:34 +00:00
${BOOST_INCLUDE_DIR}
added FindOpenImageIO cmake module.
2011-09-18 04:55:22 +00:00
${OPENIMAGEIO_INCLUDE_DIRS}
OSL/cmake: the non-intrusive parts, so other os-maintainers can take code over
2012-09-24 19:06:28 +00:00
${OPENIMAGEIO_INCLUDE_DIRS}/OpenImageIO
Revert recent OpenEXR cmake/linux fix and earlier OS X fix to use GCC for AVX. Both need a better solution and break working builds.
2014-01-27 18:11:36 +00:00
${OPENEXR_INCLUDE_DIR}
Attempt to fix T38373: cmake linux build failure with OpenEXR.
2014-01-27 14:25:22 +00:00
${OPENEXR_INCLUDE_DIRS})
Cycles render engine, initial commit. This is the engine itself, blender modifications and build instructions will follow later. Cycles uses code from some great open source projects, many thanks them: * BVH building and traversal code from NVidia's "Understanding the Efficiency of Ray Traversal on GPUs": http://code.google.com/p/understanding-the-efficiency-of-ray-traversal-on-gpus/ * Open Shading Language for a large part of the shading system: http://code.google.com/p/openshadinglanguage/ * Blender for procedural textures and a few other nodes. * Approximate Catmull Clark subdivision from NVidia Mesh tools: http://code.google.com/p/nvidia-mesh-tools/ * Sobol direction vectors from: http://web.maths.unsw.edu.au/~fkuo/sobol/ * Film response functions from: http://www.cs.columbia.edu/CAVE/software/softlib/dorf.php
2011-04-27 11:58:34 +00:00
# Subdirectories
Cycles: * add some (disabled) test code for using OpenImageIO in imbuf * link cycles, openimageio and boost into blender instead of a shared library * some cmakefile changes to simplify the code and follow conventions better * this may solve running cycles problems on windows XP, or give a different and hopefully more useful error message
2011-08-16 16:15:34 +00:00
if(WITH_CYCLES_BLENDER)
UI: small tweak to tooltips for enum menus, it wasn't very clear which description was for the property and which for the item.
2011-11-18 21:06:36 +00:00
add_subdirectory(blender)
code cleanup: warnings
2012-06-01 14:59:06 +00:00
endif()
Cycles: * add some (disabled) test code for using OpenImageIO in imbuf * link cycles, openimageio and boost into blender instead of a shared library * some cmakefile changes to simplify the code and follow conventions better * this may solve running cycles problems on windows XP, or give a different and hopefully more useful error message
2011-08-16 16:15:34 +00:00
Cycles / Standalone: * Rename test to standalone. Note: New CMAKE flag is WITH_CYCLES_STANDALONE.
2013-08-27 02:37:48 +00:00
if(WITH_CYCLES_STANDALONE OR WITH_CYCLES_NETWORK)
Cycles / CMake: * Removed PARTIO building code, partio code was removed already. * Include "app" dir only when building with CYCLES_TEST enabled.
2012-11-09 18:31:04 +00:00
add_subdirectory(app)
endif()
Cycles: * add some (disabled) test code for using OpenImageIO in imbuf * link cycles, openimageio and boost into blender instead of a shared library * some cmakefile changes to simplify the code and follow conventions better * this may solve running cycles problems on windows XP, or give a different and hopefully more useful error message
2011-08-16 16:15:34 +00:00
add_subdirectory(bvh)
add_subdirectory(device)
add_subdirectory(doc)
add_subdirectory(kernel)
add_subdirectory(render)
add_subdirectory(subd)
add_subdirectory(util)
Cycles: * Add back option to bundle CUDA kernel binaries with builds. * Disable runtime CUDA kernel compilation on Windows, couldn't get this working, since it seems to depend on visual studio being installed, even though for this particular case it shouldn't be needed. CMake only at the moment. * Runtime compilation on linux/mac should now work if nvcc is not installed in the default location, but available in PATH.
2011-11-10 12:52:17 +00:00
Cycles: Add a cmake config to easily compile Cycles Standalone. On Linux/Mac OS X, simply type "make cycles" inside the Blender source directory, to get a standalone build of the engine. Reviewed by: Brecht Differential Revision: https://developer.blender.org/D228
2014-01-20 19:41:54 +00:00
if(NOT WITH_BLENDER AND WITH_CYCLES_STANDALONE)
delayed_do_install(${CMAKE_BINARY_DIR}/bin)
endif()
Copy Permalink