OpenSubdiv/opensubdiv/osd/CMakeLists.txt

426 lines
11 KiB
CMake
Raw Normal View History

#
# Copyright 2013 Pixar
#
# Licensed under the Apache License, Version 2.0 (the "Apache License")
# with the following modification; you may not use this file except in
# compliance with the Apache License and the following modification to it:
# Section 6. Trademarks. is deleted and replaced with:
#
# 6. Trademarks. This License does not grant permission to use the trade
# names, trademarks, service marks, or product names of the Licensor
# and its affiliates, except as required to comply with Section 4(c) of
# the License and to reproduce the content of the NOTICE file.
#
# You may obtain a copy of the Apache License at
#
# http://www.apache.org/licenses/LICENSE-2.0
#
# Unless required by applicable law or agreed to in writing, software
# distributed under the Apache License with the above modification is
# distributed on an "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY
# KIND, either express or implied. See the Apache License for the specific
# language governing permissions and limitations under the Apache License.
#
2012-06-09 21:22:57 +00:00
#-------------------------------------------------------------------------------
#-------------------------------------------------------------------------------
# source & headers
set(CPU_SOURCE_FILES
Refurbish osd layer API. In OpenSubdiv 2.x, we encapsulated subdivision tables into compute context in osd layer since those tables are order-dependent and have to be applied in a certain manner. In 3.0, we adopted stencil table based refinement. It's more simple and such an encapsulation is no longer needed. Also 2.0 API has several ownership issues of GPU kernel caching, and forces unnecessary instantiation of controllers even though the cpu kernels typically don't need instances unlike GPU ones. This change completely revisit osd client facing APIs. All contexts and controllers were replaced with device-specific tables and evaluators. While we can still use consistent API across various device backends, unnecessary complexities have been removed. For example, cpu evaluator is just a set of static functions and also there's no need to replicate FarStencilTables to ComputeContext. Also the new API delegates the ownership of compiled GPU kernels to clients, for the better management of resources especially in multiple GPU environment. In addition to integrating ComputeController and EvalStencilController into a single function Evaluator::EvalStencils(), EvalLimit API is also added into Evaluator. This is working but still in progress, and we'll make a followup change for the complete implementation. -some naming convention changes: GLSLTransformFeedback to GLXFBEvaluator GLSLCompute to GLComputeEvaluator -move LimitLocation struct into examples/glEvalLimit. We're still discussing patch evaluation interface. Basically we'd like to tease all ptex-specific parametrization out of far/osd layer. TODO: -implments EvalPatches() in the right way -derivative evaluation API is still interim. -VertexBufferDescriptor needs a better API to advance its location -synchronization mechanism is not ideal (too global). -OsdMesh class is hacky. need to fix it.
2015-05-09 00:31:26 +00:00
cpuEvaluator.cpp
cpuKernel.cpp
cpuPatchTable.cpp
cpuVertexBuffer.cpp
)
set(GPU_SOURCE_FILES )
set(INC_FILES )
set(PRIVATE_HEADER_FILES
cpuKernel.h
)
set(PUBLIC_HEADER_FILES
bufferDescriptor.h
Refurbish osd layer API. In OpenSubdiv 2.x, we encapsulated subdivision tables into compute context in osd layer since those tables are order-dependent and have to be applied in a certain manner. In 3.0, we adopted stencil table based refinement. It's more simple and such an encapsulation is no longer needed. Also 2.0 API has several ownership issues of GPU kernel caching, and forces unnecessary instantiation of controllers even though the cpu kernels typically don't need instances unlike GPU ones. This change completely revisit osd client facing APIs. All contexts and controllers were replaced with device-specific tables and evaluators. While we can still use consistent API across various device backends, unnecessary complexities have been removed. For example, cpu evaluator is just a set of static functions and also there's no need to replicate FarStencilTables to ComputeContext. Also the new API delegates the ownership of compiled GPU kernels to clients, for the better management of resources especially in multiple GPU environment. In addition to integrating ComputeController and EvalStencilController into a single function Evaluator::EvalStencils(), EvalLimit API is also added into Evaluator. This is working but still in progress, and we'll make a followup change for the complete implementation. -some naming convention changes: GLSLTransformFeedback to GLXFBEvaluator GLSLCompute to GLComputeEvaluator -move LimitLocation struct into examples/glEvalLimit. We're still discussing patch evaluation interface. Basically we'd like to tease all ptex-specific parametrization out of far/osd layer. TODO: -implments EvalPatches() in the right way -derivative evaluation API is still interim. -VertexBufferDescriptor needs a better API to advance its location -synchronization mechanism is not ideal (too global). -OsdMesh class is hacky. need to fix it.
2015-05-09 00:31:26 +00:00
cpuEvaluator.h
cpuPatchTable.h
cpuVertexBuffer.h
mesh.h
nonCopyable.h
opengl.h
types.h
)
list(APPEND KERNEL_FILES
patchBasisCommonTypes.h
patchBasisCommon.h
patchBasisCommonEval.h
)
set(DOXY_HEADER_FILES ${PUBLIC_HEADER_FILES})
#-------------------------------------------------------------------------------
set(OPENMP_PUBLIC_HEADERS
Refurbish osd layer API. In OpenSubdiv 2.x, we encapsulated subdivision tables into compute context in osd layer since those tables are order-dependent and have to be applied in a certain manner. In 3.0, we adopted stencil table based refinement. It's more simple and such an encapsulation is no longer needed. Also 2.0 API has several ownership issues of GPU kernel caching, and forces unnecessary instantiation of controllers even though the cpu kernels typically don't need instances unlike GPU ones. This change completely revisit osd client facing APIs. All contexts and controllers were replaced with device-specific tables and evaluators. While we can still use consistent API across various device backends, unnecessary complexities have been removed. For example, cpu evaluator is just a set of static functions and also there's no need to replicate FarStencilTables to ComputeContext. Also the new API delegates the ownership of compiled GPU kernels to clients, for the better management of resources especially in multiple GPU environment. In addition to integrating ComputeController and EvalStencilController into a single function Evaluator::EvalStencils(), EvalLimit API is also added into Evaluator. This is working but still in progress, and we'll make a followup change for the complete implementation. -some naming convention changes: GLSLTransformFeedback to GLXFBEvaluator GLSLCompute to GLComputeEvaluator -move LimitLocation struct into examples/glEvalLimit. We're still discussing patch evaluation interface. Basically we'd like to tease all ptex-specific parametrization out of far/osd layer. TODO: -implments EvalPatches() in the right way -derivative evaluation API is still interim. -VertexBufferDescriptor needs a better API to advance its location -synchronization mechanism is not ideal (too global). -OsdMesh class is hacky. need to fix it.
2015-05-09 00:31:26 +00:00
ompEvaluator.h
ompKernel.h
)
if(OPENMP_FOUND )
list(APPEND CPU_SOURCE_FILES
Refurbish osd layer API. In OpenSubdiv 2.x, we encapsulated subdivision tables into compute context in osd layer since those tables are order-dependent and have to be applied in a certain manner. In 3.0, we adopted stencil table based refinement. It's more simple and such an encapsulation is no longer needed. Also 2.0 API has several ownership issues of GPU kernel caching, and forces unnecessary instantiation of controllers even though the cpu kernels typically don't need instances unlike GPU ones. This change completely revisit osd client facing APIs. All contexts and controllers were replaced with device-specific tables and evaluators. While we can still use consistent API across various device backends, unnecessary complexities have been removed. For example, cpu evaluator is just a set of static functions and also there's no need to replicate FarStencilTables to ComputeContext. Also the new API delegates the ownership of compiled GPU kernels to clients, for the better management of resources especially in multiple GPU environment. In addition to integrating ComputeController and EvalStencilController into a single function Evaluator::EvalStencils(), EvalLimit API is also added into Evaluator. This is working but still in progress, and we'll make a followup change for the complete implementation. -some naming convention changes: GLSLTransformFeedback to GLXFBEvaluator GLSLCompute to GLComputeEvaluator -move LimitLocation struct into examples/glEvalLimit. We're still discussing patch evaluation interface. Basically we'd like to tease all ptex-specific parametrization out of far/osd layer. TODO: -implments EvalPatches() in the right way -derivative evaluation API is still interim. -VertexBufferDescriptor needs a better API to advance its location -synchronization mechanism is not ideal (too global). -OsdMesh class is hacky. need to fix it.
2015-05-09 00:31:26 +00:00
ompEvaluator.cpp
ompKernel.cpp
)
list(APPEND PUBLIC_HEADER_FILES ${OPENMP_PUBLIC_HEADERS})
if (CMAKE_COMPILER_IS_GNUCXX)
list(APPEND PLATFORM_CPU_LIBRARIES gomp)
endif()
endif()
list(APPEND DOXY_HEADER_FILES ${OPENMP_PUBLIC_HEADERS})
#-------------------------------------------------------------------------------
set(TBB_PUBLIC_HEADERS
Refurbish osd layer API. In OpenSubdiv 2.x, we encapsulated subdivision tables into compute context in osd layer since those tables are order-dependent and have to be applied in a certain manner. In 3.0, we adopted stencil table based refinement. It's more simple and such an encapsulation is no longer needed. Also 2.0 API has several ownership issues of GPU kernel caching, and forces unnecessary instantiation of controllers even though the cpu kernels typically don't need instances unlike GPU ones. This change completely revisit osd client facing APIs. All contexts and controllers were replaced with device-specific tables and evaluators. While we can still use consistent API across various device backends, unnecessary complexities have been removed. For example, cpu evaluator is just a set of static functions and also there's no need to replicate FarStencilTables to ComputeContext. Also the new API delegates the ownership of compiled GPU kernels to clients, for the better management of resources especially in multiple GPU environment. In addition to integrating ComputeController and EvalStencilController into a single function Evaluator::EvalStencils(), EvalLimit API is also added into Evaluator. This is working but still in progress, and we'll make a followup change for the complete implementation. -some naming convention changes: GLSLTransformFeedback to GLXFBEvaluator GLSLCompute to GLComputeEvaluator -move LimitLocation struct into examples/glEvalLimit. We're still discussing patch evaluation interface. Basically we'd like to tease all ptex-specific parametrization out of far/osd layer. TODO: -implments EvalPatches() in the right way -derivative evaluation API is still interim. -VertexBufferDescriptor needs a better API to advance its location -synchronization mechanism is not ideal (too global). -OsdMesh class is hacky. need to fix it.
2015-05-09 00:31:26 +00:00
tbbEvaluator.h
tbbKernel.h
)
if( TBB_FOUND )
include_directories("${TBB_INCLUDE_DIR}")
list(APPEND CPU_SOURCE_FILES
Refurbish osd layer API. In OpenSubdiv 2.x, we encapsulated subdivision tables into compute context in osd layer since those tables are order-dependent and have to be applied in a certain manner. In 3.0, we adopted stencil table based refinement. It's more simple and such an encapsulation is no longer needed. Also 2.0 API has several ownership issues of GPU kernel caching, and forces unnecessary instantiation of controllers even though the cpu kernels typically don't need instances unlike GPU ones. This change completely revisit osd client facing APIs. All contexts and controllers were replaced with device-specific tables and evaluators. While we can still use consistent API across various device backends, unnecessary complexities have been removed. For example, cpu evaluator is just a set of static functions and also there's no need to replicate FarStencilTables to ComputeContext. Also the new API delegates the ownership of compiled GPU kernels to clients, for the better management of resources especially in multiple GPU environment. In addition to integrating ComputeController and EvalStencilController into a single function Evaluator::EvalStencils(), EvalLimit API is also added into Evaluator. This is working but still in progress, and we'll make a followup change for the complete implementation. -some naming convention changes: GLSLTransformFeedback to GLXFBEvaluator GLSLCompute to GLComputeEvaluator -move LimitLocation struct into examples/glEvalLimit. We're still discussing patch evaluation interface. Basically we'd like to tease all ptex-specific parametrization out of far/osd layer. TODO: -implments EvalPatches() in the right way -derivative evaluation API is still interim. -VertexBufferDescriptor needs a better API to advance its location -synchronization mechanism is not ideal (too global). -OsdMesh class is hacky. need to fix it.
2015-05-09 00:31:26 +00:00
tbbEvaluator.cpp
tbbKernel.cpp
)
list(APPEND PUBLIC_HEADER_FILES ${TBB_PUBLIC_HEADERS})
if (CMAKE_COMPILER_IS_GNUCXX)
list(APPEND PLATFORM_CPU_LIBRARIES gomp)
endif()
list(APPEND PLATFORM_CPU_LIBRARIES
${TBB_LIBRARIES}
)
endif()
list(APPEND DOXY_HEADER_FILES ${TBB_PUBLIC_HEADERS})
#-------------------------------------------------------------------------------
# GL code & dependencies
set(GL_PUBLIC_HEADERS
cpuGLVertexBuffer.h
glLegacyGregoryPatchTable.h
glPatchTable.h
glVertexBuffer.h
glMesh.h
glslPatchShaderSource.h
)
2016-07-26 19:41:18 +00:00
if( (NOT NO_OPENGL) AND (OPENGL_FOUND OR OPENGLES_FOUND) )
list(APPEND GPU_SOURCE_FILES
cpuGLVertexBuffer.cpp
glLegacyGregoryPatchTable.cpp
glPatchTable.cpp
glVertexBuffer.cpp
glslPatchShaderSource.cpp
)
list(APPEND PUBLIC_HEADER_FILES ${GL_PUBLIC_HEADERS})
if ( OPENGL_FOUND )
list(APPEND KERNEL_FILES
glslPatchCommon.glsl
glslPatchCommonTess.glsl
glslPatchBoxSplineTriangle.glsl
glslPatchBSpline.glsl
glslPatchGregory.glsl
glslPatchGregoryBasis.glsl
glslPatchGregoryTriangle.glsl
glslPatchLegacy.glsl
)
endif()
endif()
list(APPEND DOXY_HEADER_FILES ${GL_PUBLIC_HEADERS})
#-------------------------------------------------------------------------------
# OpenGL 4.2 dependencies
# note : (GLSL transform feedback kernels require GL 4.2)
set(GL_4_2_PUBLIC_HEADERS
Refurbish osd layer API. In OpenSubdiv 2.x, we encapsulated subdivision tables into compute context in osd layer since those tables are order-dependent and have to be applied in a certain manner. In 3.0, we adopted stencil table based refinement. It's more simple and such an encapsulation is no longer needed. Also 2.0 API has several ownership issues of GPU kernel caching, and forces unnecessary instantiation of controllers even though the cpu kernels typically don't need instances unlike GPU ones. This change completely revisit osd client facing APIs. All contexts and controllers were replaced with device-specific tables and evaluators. While we can still use consistent API across various device backends, unnecessary complexities have been removed. For example, cpu evaluator is just a set of static functions and also there's no need to replicate FarStencilTables to ComputeContext. Also the new API delegates the ownership of compiled GPU kernels to clients, for the better management of resources especially in multiple GPU environment. In addition to integrating ComputeController and EvalStencilController into a single function Evaluator::EvalStencils(), EvalLimit API is also added into Evaluator. This is working but still in progress, and we'll make a followup change for the complete implementation. -some naming convention changes: GLSLTransformFeedback to GLXFBEvaluator GLSLCompute to GLComputeEvaluator -move LimitLocation struct into examples/glEvalLimit. We're still discussing patch evaluation interface. Basically we'd like to tease all ptex-specific parametrization out of far/osd layer. TODO: -implments EvalPatches() in the right way -derivative evaluation API is still interim. -VertexBufferDescriptor needs a better API to advance its location -synchronization mechanism is not ideal (too global). -OsdMesh class is hacky. need to fix it.
2015-05-09 00:31:26 +00:00
glXFBEvaluator.h
)
if( OPENGL_4_2_FOUND )
list(APPEND GPU_SOURCE_FILES
Refurbish osd layer API. In OpenSubdiv 2.x, we encapsulated subdivision tables into compute context in osd layer since those tables are order-dependent and have to be applied in a certain manner. In 3.0, we adopted stencil table based refinement. It's more simple and such an encapsulation is no longer needed. Also 2.0 API has several ownership issues of GPU kernel caching, and forces unnecessary instantiation of controllers even though the cpu kernels typically don't need instances unlike GPU ones. This change completely revisit osd client facing APIs. All contexts and controllers were replaced with device-specific tables and evaluators. While we can still use consistent API across various device backends, unnecessary complexities have been removed. For example, cpu evaluator is just a set of static functions and also there's no need to replicate FarStencilTables to ComputeContext. Also the new API delegates the ownership of compiled GPU kernels to clients, for the better management of resources especially in multiple GPU environment. In addition to integrating ComputeController and EvalStencilController into a single function Evaluator::EvalStencils(), EvalLimit API is also added into Evaluator. This is working but still in progress, and we'll make a followup change for the complete implementation. -some naming convention changes: GLSLTransformFeedback to GLXFBEvaluator GLSLCompute to GLComputeEvaluator -move LimitLocation struct into examples/glEvalLimit. We're still discussing patch evaluation interface. Basically we'd like to tease all ptex-specific parametrization out of far/osd layer. TODO: -implments EvalPatches() in the right way -derivative evaluation API is still interim. -VertexBufferDescriptor needs a better API to advance its location -synchronization mechanism is not ideal (too global). -OsdMesh class is hacky. need to fix it.
2015-05-09 00:31:26 +00:00
glXFBEvaluator.cpp
)
list(APPEND PUBLIC_HEADER_FILES ${GL_4_2_PUBLIC_HEADERS})
list(APPEND KERNEL_FILES
Refurbish osd layer API. In OpenSubdiv 2.x, we encapsulated subdivision tables into compute context in osd layer since those tables are order-dependent and have to be applied in a certain manner. In 3.0, we adopted stencil table based refinement. It's more simple and such an encapsulation is no longer needed. Also 2.0 API has several ownership issues of GPU kernel caching, and forces unnecessary instantiation of controllers even though the cpu kernels typically don't need instances unlike GPU ones. This change completely revisit osd client facing APIs. All contexts and controllers were replaced with device-specific tables and evaluators. While we can still use consistent API across various device backends, unnecessary complexities have been removed. For example, cpu evaluator is just a set of static functions and also there's no need to replicate FarStencilTables to ComputeContext. Also the new API delegates the ownership of compiled GPU kernels to clients, for the better management of resources especially in multiple GPU environment. In addition to integrating ComputeController and EvalStencilController into a single function Evaluator::EvalStencils(), EvalLimit API is also added into Evaluator. This is working but still in progress, and we'll make a followup change for the complete implementation. -some naming convention changes: GLSLTransformFeedback to GLXFBEvaluator GLSLCompute to GLComputeEvaluator -move LimitLocation struct into examples/glEvalLimit. We're still discussing patch evaluation interface. Basically we'd like to tease all ptex-specific parametrization out of far/osd layer. TODO: -implments EvalPatches() in the right way -derivative evaluation API is still interim. -VertexBufferDescriptor needs a better API to advance its location -synchronization mechanism is not ideal (too global). -OsdMesh class is hacky. need to fix it.
2015-05-09 00:31:26 +00:00
glslXFBKernel.glsl
)
list(APPEND PLATFORM_GPU_LIBRARIES
${OPENGL_LOADER_LIBRARIES}
)
endif()
list(APPEND DOXY_HEADER_FILES ${GL_4_2_PUBLIC_HEADERS})
#-------------------------------------------------------------------------------
# OpenGL 4.3 dependencies
# note : (GLSL compute shader kernels require GL 4.3)
set(GL_4_3_PUBLIC_HEADERS
Refurbish osd layer API. In OpenSubdiv 2.x, we encapsulated subdivision tables into compute context in osd layer since those tables are order-dependent and have to be applied in a certain manner. In 3.0, we adopted stencil table based refinement. It's more simple and such an encapsulation is no longer needed. Also 2.0 API has several ownership issues of GPU kernel caching, and forces unnecessary instantiation of controllers even though the cpu kernels typically don't need instances unlike GPU ones. This change completely revisit osd client facing APIs. All contexts and controllers were replaced with device-specific tables and evaluators. While we can still use consistent API across various device backends, unnecessary complexities have been removed. For example, cpu evaluator is just a set of static functions and also there's no need to replicate FarStencilTables to ComputeContext. Also the new API delegates the ownership of compiled GPU kernels to clients, for the better management of resources especially in multiple GPU environment. In addition to integrating ComputeController and EvalStencilController into a single function Evaluator::EvalStencils(), EvalLimit API is also added into Evaluator. This is working but still in progress, and we'll make a followup change for the complete implementation. -some naming convention changes: GLSLTransformFeedback to GLXFBEvaluator GLSLCompute to GLComputeEvaluator -move LimitLocation struct into examples/glEvalLimit. We're still discussing patch evaluation interface. Basically we'd like to tease all ptex-specific parametrization out of far/osd layer. TODO: -implments EvalPatches() in the right way -derivative evaluation API is still interim. -VertexBufferDescriptor needs a better API to advance its location -synchronization mechanism is not ideal (too global). -OsdMesh class is hacky. need to fix it.
2015-05-09 00:31:26 +00:00
glComputeEvaluator.h
)
if( OPENGL_4_3_FOUND )
list(APPEND GPU_SOURCE_FILES
Refurbish osd layer API. In OpenSubdiv 2.x, we encapsulated subdivision tables into compute context in osd layer since those tables are order-dependent and have to be applied in a certain manner. In 3.0, we adopted stencil table based refinement. It's more simple and such an encapsulation is no longer needed. Also 2.0 API has several ownership issues of GPU kernel caching, and forces unnecessary instantiation of controllers even though the cpu kernels typically don't need instances unlike GPU ones. This change completely revisit osd client facing APIs. All contexts and controllers were replaced with device-specific tables and evaluators. While we can still use consistent API across various device backends, unnecessary complexities have been removed. For example, cpu evaluator is just a set of static functions and also there's no need to replicate FarStencilTables to ComputeContext. Also the new API delegates the ownership of compiled GPU kernels to clients, for the better management of resources especially in multiple GPU environment. In addition to integrating ComputeController and EvalStencilController into a single function Evaluator::EvalStencils(), EvalLimit API is also added into Evaluator. This is working but still in progress, and we'll make a followup change for the complete implementation. -some naming convention changes: GLSLTransformFeedback to GLXFBEvaluator GLSLCompute to GLComputeEvaluator -move LimitLocation struct into examples/glEvalLimit. We're still discussing patch evaluation interface. Basically we'd like to tease all ptex-specific parametrization out of far/osd layer. TODO: -implments EvalPatches() in the right way -derivative evaluation API is still interim. -VertexBufferDescriptor needs a better API to advance its location -synchronization mechanism is not ideal (too global). -OsdMesh class is hacky. need to fix it.
2015-05-09 00:31:26 +00:00
glComputeEvaluator.cpp
)
list(APPEND PUBLIC_HEADER_FILES ${GL_4_3_PUBLIC_HEADERS})
list(APPEND KERNEL_FILES
glslComputeKernel.glsl
)
list(APPEND PLATFORM_GPU_LIBRARIES
${OPENGL_LOADER_LIBRARIES}
)
endif()
list(APPEND DOXY_HEADER_FILES ${GL_4_3_PUBLIC_HEADERS})
#-------------------------------------------------------------------------------
# DX11 code & dependencies
set(DXSDK_PUBLIC_HEADERS
cpuD3D11VertexBuffer.h
Refurbish osd layer API. In OpenSubdiv 2.x, we encapsulated subdivision tables into compute context in osd layer since those tables are order-dependent and have to be applied in a certain manner. In 3.0, we adopted stencil table based refinement. It's more simple and such an encapsulation is no longer needed. Also 2.0 API has several ownership issues of GPU kernel caching, and forces unnecessary instantiation of controllers even though the cpu kernels typically don't need instances unlike GPU ones. This change completely revisit osd client facing APIs. All contexts and controllers were replaced with device-specific tables and evaluators. While we can still use consistent API across various device backends, unnecessary complexities have been removed. For example, cpu evaluator is just a set of static functions and also there's no need to replicate FarStencilTables to ComputeContext. Also the new API delegates the ownership of compiled GPU kernels to clients, for the better management of resources especially in multiple GPU environment. In addition to integrating ComputeController and EvalStencilController into a single function Evaluator::EvalStencils(), EvalLimit API is also added into Evaluator. This is working but still in progress, and we'll make a followup change for the complete implementation. -some naming convention changes: GLSLTransformFeedback to GLXFBEvaluator GLSLCompute to GLComputeEvaluator -move LimitLocation struct into examples/glEvalLimit. We're still discussing patch evaluation interface. Basically we'd like to tease all ptex-specific parametrization out of far/osd layer. TODO: -implments EvalPatches() in the right way -derivative evaluation API is still interim. -VertexBufferDescriptor needs a better API to advance its location -synchronization mechanism is not ideal (too global). -OsdMesh class is hacky. need to fix it.
2015-05-09 00:31:26 +00:00
d3d11ComputeEvaluator.h
d3d11LegacyGregoryPatchTable.h
d3d11PatchTable.h
d3d11VertexBuffer.h
d3d11Mesh.h
hlslPatchShaderSource.h
)
if( DXSDK_FOUND )
list(APPEND GPU_SOURCE_FILES
cpuD3D11VertexBuffer.cpp
Refurbish osd layer API. In OpenSubdiv 2.x, we encapsulated subdivision tables into compute context in osd layer since those tables are order-dependent and have to be applied in a certain manner. In 3.0, we adopted stencil table based refinement. It's more simple and such an encapsulation is no longer needed. Also 2.0 API has several ownership issues of GPU kernel caching, and forces unnecessary instantiation of controllers even though the cpu kernels typically don't need instances unlike GPU ones. This change completely revisit osd client facing APIs. All contexts and controllers were replaced with device-specific tables and evaluators. While we can still use consistent API across various device backends, unnecessary complexities have been removed. For example, cpu evaluator is just a set of static functions and also there's no need to replicate FarStencilTables to ComputeContext. Also the new API delegates the ownership of compiled GPU kernels to clients, for the better management of resources especially in multiple GPU environment. In addition to integrating ComputeController and EvalStencilController into a single function Evaluator::EvalStencils(), EvalLimit API is also added into Evaluator. This is working but still in progress, and we'll make a followup change for the complete implementation. -some naming convention changes: GLSLTransformFeedback to GLXFBEvaluator GLSLCompute to GLComputeEvaluator -move LimitLocation struct into examples/glEvalLimit. We're still discussing patch evaluation interface. Basically we'd like to tease all ptex-specific parametrization out of far/osd layer. TODO: -implments EvalPatches() in the right way -derivative evaluation API is still interim. -VertexBufferDescriptor needs a better API to advance its location -synchronization mechanism is not ideal (too global). -OsdMesh class is hacky. need to fix it.
2015-05-09 00:31:26 +00:00
d3d11ComputeEvaluator.cpp
d3d11LegacyGregoryPatchTable.cpp
d3d11PatchTable.cpp
d3d11VertexBuffer.cpp
hlslPatchShaderSource.cpp
)
list(APPEND PUBLIC_HEADER_FILES ${DXSDK_PUBLIC_HEADERS})
list(APPEND KERNEL_FILES
hlslComputeKernel.hlsl
hlslPatchCommon.hlsl
hlslPatchCommonTess.hlsl
hlslPatchBoxSplineTriangle.hlsl
hlslPatchBSpline.hlsl
hlslPatchGregory.hlsl
hlslPatchGregoryBasis.hlsl
hlslPatchGregoryTriangle.hlsl
hlslPatchLegacy.hlsl
)
list(APPEND PLATFORM_GPU_LIBRARIES
${DXSDK_LIBRARIES}
)
endif()
list(APPEND DOXY_HEADER_FILES ${DXSDK_PUBLIC_HEADERS})
2016-07-26 19:41:18 +00:00
#-------------------------------------------------------------------------------
# Metal code & dependencies
set(METAL_PUBLIC_HEADERS
mtlVertexBuffer.h
mtlComputeEvaluator.h
mtlLegacyGregoryPatchTable.h
mtlPatchTable.h
mtlMesh.h
mtlPatchShaderSource.h
mtlCommon.h
)
if( METAL_FOUND )
set(METAL_SOURCE_FILES
mtlVertexBuffer.mm
mtlComputeEvaluator.mm
mtlLegacyGregoryPatchTable.mm
mtlPatchTable.mm
mtlVertexBuffer.mm
mtlPatchShaderSource.mm
)
set_source_files_properties(
${METAL_SOURCE_FILES}
PROPERTIES
COMPILE_FLAGS
"-fobjc-arc")
list(APPEND GPU_SOURCE_FILES ${METAL_SOURCE_FILES})
list(APPEND PUBLIC_HEADER_FILES ${METAL_PUBLIC_HEADERS})
list(APPEND KERNEL_FILES
mtlComputeKernel.metal
mtlPatchCommon.metal
mtlPatchCommonTess.metal
mtlPatchBoxSplineTriangle.metal
2016-07-26 19:41:18 +00:00
mtlPatchBSpline.metal
mtlPatchGregory.metal
mtlPatchGregoryBasis.metal
mtlPatchGregoryTriangle.metal
mtlPatchLegacy.metal
2016-07-26 19:41:18 +00:00
)
list(APPEND PLATFORM_GPU_LIBRARIES
${METAL_LIBRARIES}
)
endif()
list(APPEND DOXY_HEADER_FILES ${METAL_PUBLIC_HEADERS})
#-------------------------------------------------------------------------------
# OpenCL code & dependencies
set(OPENCL_PUBLIC_HEADERS
Refurbish osd layer API. In OpenSubdiv 2.x, we encapsulated subdivision tables into compute context in osd layer since those tables are order-dependent and have to be applied in a certain manner. In 3.0, we adopted stencil table based refinement. It's more simple and such an encapsulation is no longer needed. Also 2.0 API has several ownership issues of GPU kernel caching, and forces unnecessary instantiation of controllers even though the cpu kernels typically don't need instances unlike GPU ones. This change completely revisit osd client facing APIs. All contexts and controllers were replaced with device-specific tables and evaluators. While we can still use consistent API across various device backends, unnecessary complexities have been removed. For example, cpu evaluator is just a set of static functions and also there's no need to replicate FarStencilTables to ComputeContext. Also the new API delegates the ownership of compiled GPU kernels to clients, for the better management of resources especially in multiple GPU environment. In addition to integrating ComputeController and EvalStencilController into a single function Evaluator::EvalStencils(), EvalLimit API is also added into Evaluator. This is working but still in progress, and we'll make a followup change for the complete implementation. -some naming convention changes: GLSLTransformFeedback to GLXFBEvaluator GLSLCompute to GLComputeEvaluator -move LimitLocation struct into examples/glEvalLimit. We're still discussing patch evaluation interface. Basically we'd like to tease all ptex-specific parametrization out of far/osd layer. TODO: -implments EvalPatches() in the right way -derivative evaluation API is still interim. -VertexBufferDescriptor needs a better API to advance its location -synchronization mechanism is not ideal (too global). -OsdMesh class is hacky. need to fix it.
2015-05-09 00:31:26 +00:00
clEvaluator.h
clPatchTable.h
clVertexBuffer.h
opencl.h
)
if ( OPENCL_FOUND )
list(APPEND GPU_SOURCE_FILES
Refurbish osd layer API. In OpenSubdiv 2.x, we encapsulated subdivision tables into compute context in osd layer since those tables are order-dependent and have to be applied in a certain manner. In 3.0, we adopted stencil table based refinement. It's more simple and such an encapsulation is no longer needed. Also 2.0 API has several ownership issues of GPU kernel caching, and forces unnecessary instantiation of controllers even though the cpu kernels typically don't need instances unlike GPU ones. This change completely revisit osd client facing APIs. All contexts and controllers were replaced with device-specific tables and evaluators. While we can still use consistent API across various device backends, unnecessary complexities have been removed. For example, cpu evaluator is just a set of static functions and also there's no need to replicate FarStencilTables to ComputeContext. Also the new API delegates the ownership of compiled GPU kernels to clients, for the better management of resources especially in multiple GPU environment. In addition to integrating ComputeController and EvalStencilController into a single function Evaluator::EvalStencils(), EvalLimit API is also added into Evaluator. This is working but still in progress, and we'll make a followup change for the complete implementation. -some naming convention changes: GLSLTransformFeedback to GLXFBEvaluator GLSLCompute to GLComputeEvaluator -move LimitLocation struct into examples/glEvalLimit. We're still discussing patch evaluation interface. Basically we'd like to tease all ptex-specific parametrization out of far/osd layer. TODO: -implments EvalPatches() in the right way -derivative evaluation API is still interim. -VertexBufferDescriptor needs a better API to advance its location -synchronization mechanism is not ideal (too global). -OsdMesh class is hacky. need to fix it.
2015-05-09 00:31:26 +00:00
clEvaluator.cpp
clPatchTable.cpp
clVertexBuffer.cpp
)
list(APPEND PUBLIC_HEADER_FILES ${OPENCL_PUBLIC_HEADERS})
list(APPEND KERNEL_FILES
clKernel.cl
)
list(APPEND PLATFORM_GPU_LIBRARIES
${OPENCL_LIBRARIES}
)
include_directories( "${OPENCL_INCLUDE_DIRS}" )
if ( OPENGL_FOUND )
list(APPEND GPU_SOURCE_FILES clGLVertexBuffer.cpp)
list(APPEND PUBLIC_HEADER_FILES clGLVertexBuffer.h)
endif()
if ( OPENCL_D3D11_INTEROP_FOUND )
list(APPEND GPU_SOURCE_FILES
clD3D11VertexBuffer.cpp
)
list(APPEND PUBLIC_HEADER_FILES
clD3D11VertexBuffer.h
)
endif()
endif()
list(APPEND DOXY_HEADER_FILES ${OPENCL_PUBLIC_HEADERS})
2012-06-09 21:22:57 +00:00
#-------------------------------------------------------------------------------
# CUDA code & dependencies
set(CUDA_PUBLIC_HEADERS
Refurbish osd layer API. In OpenSubdiv 2.x, we encapsulated subdivision tables into compute context in osd layer since those tables are order-dependent and have to be applied in a certain manner. In 3.0, we adopted stencil table based refinement. It's more simple and such an encapsulation is no longer needed. Also 2.0 API has several ownership issues of GPU kernel caching, and forces unnecessary instantiation of controllers even though the cpu kernels typically don't need instances unlike GPU ones. This change completely revisit osd client facing APIs. All contexts and controllers were replaced with device-specific tables and evaluators. While we can still use consistent API across various device backends, unnecessary complexities have been removed. For example, cpu evaluator is just a set of static functions and also there's no need to replicate FarStencilTables to ComputeContext. Also the new API delegates the ownership of compiled GPU kernels to clients, for the better management of resources especially in multiple GPU environment. In addition to integrating ComputeController and EvalStencilController into a single function Evaluator::EvalStencils(), EvalLimit API is also added into Evaluator. This is working but still in progress, and we'll make a followup change for the complete implementation. -some naming convention changes: GLSLTransformFeedback to GLXFBEvaluator GLSLCompute to GLComputeEvaluator -move LimitLocation struct into examples/glEvalLimit. We're still discussing patch evaluation interface. Basically we'd like to tease all ptex-specific parametrization out of far/osd layer. TODO: -implments EvalPatches() in the right way -derivative evaluation API is still interim. -VertexBufferDescriptor needs a better API to advance its location -synchronization mechanism is not ideal (too global). -OsdMesh class is hacky. need to fix it.
2015-05-09 00:31:26 +00:00
cudaEvaluator.h
cudaPatchTable.h
cudaVertexBuffer.h
)
2012-06-09 21:22:57 +00:00
if( CUDA_FOUND )
list(APPEND GPU_SOURCE_FILES
Refurbish osd layer API. In OpenSubdiv 2.x, we encapsulated subdivision tables into compute context in osd layer since those tables are order-dependent and have to be applied in a certain manner. In 3.0, we adopted stencil table based refinement. It's more simple and such an encapsulation is no longer needed. Also 2.0 API has several ownership issues of GPU kernel caching, and forces unnecessary instantiation of controllers even though the cpu kernels typically don't need instances unlike GPU ones. This change completely revisit osd client facing APIs. All contexts and controllers were replaced with device-specific tables and evaluators. While we can still use consistent API across various device backends, unnecessary complexities have been removed. For example, cpu evaluator is just a set of static functions and also there's no need to replicate FarStencilTables to ComputeContext. Also the new API delegates the ownership of compiled GPU kernels to clients, for the better management of resources especially in multiple GPU environment. In addition to integrating ComputeController and EvalStencilController into a single function Evaluator::EvalStencils(), EvalLimit API is also added into Evaluator. This is working but still in progress, and we'll make a followup change for the complete implementation. -some naming convention changes: GLSLTransformFeedback to GLXFBEvaluator GLSLCompute to GLComputeEvaluator -move LimitLocation struct into examples/glEvalLimit. We're still discussing patch evaluation interface. Basically we'd like to tease all ptex-specific parametrization out of far/osd layer. TODO: -implments EvalPatches() in the right way -derivative evaluation API is still interim. -VertexBufferDescriptor needs a better API to advance its location -synchronization mechanism is not ideal (too global). -OsdMesh class is hacky. need to fix it.
2015-05-09 00:31:26 +00:00
cudaEvaluator.cpp
cudaPatchTable.cpp
cudaVertexBuffer.cpp
2012-06-09 21:22:57 +00:00
)
list(APPEND PUBLIC_HEADER_FILES ${CUDA_PUBLIC_HEADERS})
list(APPEND CUDA_KERNEL_FILES
osd/cudaKernel.cu
2012-06-09 21:22:57 +00:00
)
set (CUDA_KERNEL_FILES ${CUDA_KERNEL_FILES} PARENT_SCOPE)
if ( OPENGL_FOUND )
list(APPEND GPU_SOURCE_FILES
cudaGLVertexBuffer.cpp
)
list(APPEND PUBLIC_HEADER_FILES
cudaGLVertexBuffer.h
)
endif()
if( DXSDK_FOUND )
list(APPEND GPU_SOURCE_FILES
cudaD3D11VertexBuffer.cpp
)
list(APPEND PUBLIC_HEADER_FILES
cudaD3D11VertexBuffer.h
)
endif()
2012-06-09 21:22:57 +00:00
endif()
list(APPEND DOXY_HEADER_FILES ${CUDA_PUBLIC_HEADERS})
2012-06-09 21:22:57 +00:00
#-------------------------------------------------------------------------------
osd_stringify("${KERNEL_FILES}" INC_FILES)
include_directories("${CMAKE_CURRENT_BINARY_DIR}")
2012-06-09 21:22:57 +00:00
#-------------------------------------------------------------------------------
source_group("Kernels" FILES ${KERNEL_FILES})
source_group("Inc" FILES ${INC_FILES})
# Compile objs first for both the CPU and GPU libs -----
add_library(osd_cpu_obj
OBJECT
${CPU_SOURCE_FILES}
${PRIVATE_HEADER_FILES}
${PUBLIC_HEADER_FILES}
)
set_target_properties(osd_cpu_obj
PROPERTIES
FOLDER "opensubdiv"
)
if( GPU_SOURCE_FILES )
add_library(osd_gpu_obj
OBJECT
${GPU_SOURCE_FILES}
${PRIVATE_HEADER_FILES}
${PUBLIC_HEADER_FILES}
${INC_FILES}
)
set_target_properties(osd_gpu_obj
PROPERTIES
FOLDER "opensubdiv"
)
endif()
osd_add_doxy_headers( "${DOXY_HEADER_FILES}" )
install(
FILES
${PUBLIC_HEADER_FILES}
DESTINATION
"${CMAKE_INCDIR_BASE}/osd"
PERMISSIONS
OWNER_READ
GROUP_READ
WORLD_READ )
if (ANDROID)
install(
FILES
Android.mk
DESTINATION
"${LIBRARY_OUTPUT_PATH_ROOT}"
PERMISSIONS
OWNER_READ
GROUP_READ
WORLD_READ )
endif()
2012-06-09 21:22:57 +00:00
#-------------------------------------------------------------------------------