AuroraMiddleware/OpenSubdiv
1
0
Fork 0
mirror of https://github.com/PixarAnimationStudios/OpenSubdiv synced 2024-12-01 15:50:07 +00:00
Code Issues Releases Wiki Activity
c4e760e722
OpenSubdiv/opensubdiv/osd/d3d11ComputeController.cpp

316 lines
11 KiB
C++
Raw Normal View History

Added DirectCompute backend for Osd Compute
2012-12-13 18:22:30 +00:00
//
Amending Apache license language & file headers. New text: 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.
2013-09-26 19:04:57 +00:00
// Copyright 2013 Pixar
Added DirectCompute backend for Osd Compute
2012-12-13 18:22:30 +00:00
//
Amending Apache license language & file headers. New text: 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.
2013-09-26 19:04:57 +00:00
// 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:
Added DirectCompute backend for Osd Compute
2012-12-13 18:22:30 +00:00
//
Amending Apache license language & file headers. New text: 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.
2013-09-26 19:04:57 +00:00
// 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.
Added DirectCompute backend for Osd Compute
2012-12-13 18:22:30 +00:00
//
Amending Apache license language & file headers. New text: 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.
2013-09-26 19:04:57 +00:00
// You may obtain a copy of the Apache License at
Added DirectCompute backend for Osd Compute
2012-12-13 18:22:30 +00:00
//
Amending Apache license language & file headers. New text: 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.
2013-09-26 19:04:57 +00:00
// http://www.apache.org/licenses/LICENSE-2.0
Updating EULA
2013-07-18 21:19:50 +00:00
//
Amending Apache license language & file headers. New text: 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.
2013-09-26 19:04:57 +00:00
// 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.
Added DirectCompute backend for Osd Compute
2012-12-13 18:22:30 +00:00
//
#include "../osd/d3d11ComputeController.h"
#include "../osd/d3d11ComputeContext.h"
#include "../osd/d3d11KernelBundle.h"
#include <D3D11.h>
#include <algorithm>
#include <cassert>
namespace OpenSubdiv {
namespace OPENSUBDIV_VERSION {
#define SAFE_RELEASE(p) { if(p) { (p)->Release(); (p)=NULL; } }
OsdD3D11ComputeController::OsdD3D11ComputeController(
ID3D11DeviceContext *deviceContext)
Interleaved buffer support in OsdCompute. Removed OsdVertexDescriptor and replaced with OsdVertexBufferDescriptor. All kernels take offset/length/stride to apply subdivision partially in each vertex elements. Also the offset can be used for client-based VBO aggregation, without modifying index buffers. This is useful for topology sharing, in conjunction with glDrawElementsBaseVertex etc. However, gregory patch shader fetches vertex buffer via texture buffer, which index should also be offsetted too. Although gl_BaseVertexARB extension should be able to do that job, it's a relatively new extension. So we use OsdBaseVertex() call to mitigate the compatibility issue as clients can provide it in their way at least for the time being.
2014-05-09 00:20:54 +00:00
: _deviceContext(deviceContext), _query(0) {
Added DirectCompute backend for Osd Compute
2012-12-13 18:22:30 +00:00
}
OsdD3D11ComputeController::~OsdD3D11ComputeController() {
for (std::vector<OsdD3D11ComputeKernelBundle*>::iterator it =
_kernelRegistry.begin();
it != _kernelRegistry.end(); ++it) {
delete *it;
}
SAFE_RELEASE(_query);
}
void
OsdD3D11ComputeController::Synchronize() {
if (! _query) {
ID3D11Device *device = NULL;
_deviceContext->GetDevice(&device);
assert(device);
D3D11_QUERY_DESC desc;
desc.Query = D3D11_QUERY_EVENT;
desc.MiscFlags = 0;
device->CreateQuery(&desc, &_query);
}
_deviceContext->Flush();
_deviceContext->End(_query);
while (S_OK != _deviceContext->GetData(_query, NULL, 0, 0));
}
OsdD3D11ComputeKernelBundle *
Interleaved buffer support in OsdCompute. Removed OsdVertexDescriptor and replaced with OsdVertexBufferDescriptor. All kernels take offset/length/stride to apply subdivision partially in each vertex elements. Also the offset can be used for client-based VBO aggregation, without modifying index buffers. This is useful for topology sharing, in conjunction with glDrawElementsBaseVertex etc. However, gregory patch shader fetches vertex buffer via texture buffer, which index should also be offsetted too. Although gl_BaseVertexARB extension should be able to do that job, it's a relatively new extension. So we use OsdBaseVertex() call to mitigate the compatibility issue as clients can provide it in their way at least for the time being.
2014-05-09 00:20:54 +00:00
OsdD3D11ComputeController::getKernels(OsdVertexBufferDescriptor const &vertexDesc,
OsdVertexBufferDescriptor const &varyingDesc) {
Added DirectCompute backend for Osd Compute
2012-12-13 18:22:30 +00:00
std::vector<OsdD3D11ComputeKernelBundle*>::iterator it =
std::find_if(_kernelRegistry.begin(), _kernelRegistry.end(),
Interleaved buffer support in OsdCompute. Removed OsdVertexDescriptor and replaced with OsdVertexBufferDescriptor. All kernels take offset/length/stride to apply subdivision partially in each vertex elements. Also the offset can be used for client-based VBO aggregation, without modifying index buffers. This is useful for topology sharing, in conjunction with glDrawElementsBaseVertex etc. However, gregory patch shader fetches vertex buffer via texture buffer, which index should also be offsetted too. Although gl_BaseVertexARB extension should be able to do that job, it's a relatively new extension. So we use OsdBaseVertex() call to mitigate the compatibility issue as clients can provide it in their way at least for the time being.
2014-05-09 00:20:54 +00:00
OsdD3D11ComputeKernelBundle::Match(
vertexDesc, varyingDesc));
Added DirectCompute backend for Osd Compute
2012-12-13 18:22:30 +00:00
if (it != _kernelRegistry.end()) {
return *it;
} else {
OsdD3D11ComputeKernelBundle *kernelBundle =
new OsdD3D11ComputeKernelBundle(_deviceContext);
_kernelRegistry.push_back(kernelBundle);
Interleaved buffer support in OsdCompute. Removed OsdVertexDescriptor and replaced with OsdVertexBufferDescriptor. All kernels take offset/length/stride to apply subdivision partially in each vertex elements. Also the offset can be used for client-based VBO aggregation, without modifying index buffers. This is useful for topology sharing, in conjunction with glDrawElementsBaseVertex etc. However, gregory patch shader fetches vertex buffer via texture buffer, which index should also be offsetted too. Although gl_BaseVertexARB extension should be able to do that job, it's a relatively new extension. So we use OsdBaseVertex() call to mitigate the compatibility issue as clients can provide it in their way at least for the time being.
2014-05-09 00:20:54 +00:00
kernelBundle->Compile(vertexDesc, varyingDesc);
Added DirectCompute backend for Osd Compute
2012-12-13 18:22:30 +00:00
return kernelBundle;
}
}
Reorganize ComputeContext and ComputeController. Moved transient states (current vertex buffer etc) to controller. ComputeContext becomes constant so that it's well suited for coarse-grain parallelism on cpu. The prims sharing same topology (ComputeContext) can be refined simultaneously by having mutiple compute controllers. Client facing API doesn't change.
2014-05-06 15:53:36 +00:00
void
OsdD3D11ComputeController::bindShaderResources()
{
// Unbind the vertexBuffer from the input assembler
ID3D11Buffer *NULLBuffer = 0;
UINT voffset = 0;
UINT vstride = 0;
_deviceContext->IASetVertexBuffers(0, 1, &NULLBuffer, &voffset, &vstride);
// Unbind the vertexBuffer from the vertex shader (gregory patch vertex srv)
ID3D11ShaderResourceView *NULLSRV = 0;
_deviceContext->VSSetShaderResources(0, 1, &NULLSRV);
Interleaved buffer support in OsdCompute. Removed OsdVertexDescriptor and replaced with OsdVertexBufferDescriptor. All kernels take offset/length/stride to apply subdivision partially in each vertex elements. Also the offset can be used for client-based VBO aggregation, without modifying index buffers. This is useful for topology sharing, in conjunction with glDrawElementsBaseVertex etc. However, gregory patch shader fetches vertex buffer via texture buffer, which index should also be offsetted too. Although gl_BaseVertexARB extension should be able to do that job, it's a relatively new extension. So we use OsdBaseVertex() call to mitigate the compatibility issue as clients can provide it in their way at least for the time being.
2014-05-09 00:20:54 +00:00
if (_currentBindState.vertexBuffer)
_deviceContext->CSSetUnorderedAccessViews(0, 1, &_currentBindState.vertexBuffer, 0); // u0
Reorganize ComputeContext and ComputeController. Moved transient states (current vertex buffer etc) to controller. ComputeContext becomes constant so that it's well suited for coarse-grain parallelism on cpu. The prims sharing same topology (ComputeContext) can be refined simultaneously by having mutiple compute controllers. Client facing API doesn't change.
2014-05-06 15:53:36 +00:00
Interleaved buffer support in OsdCompute. Removed OsdVertexDescriptor and replaced with OsdVertexBufferDescriptor. All kernels take offset/length/stride to apply subdivision partially in each vertex elements. Also the offset can be used for client-based VBO aggregation, without modifying index buffers. This is useful for topology sharing, in conjunction with glDrawElementsBaseVertex etc. However, gregory patch shader fetches vertex buffer via texture buffer, which index should also be offsetted too. Although gl_BaseVertexARB extension should be able to do that job, it's a relatively new extension. So we use OsdBaseVertex() call to mitigate the compatibility issue as clients can provide it in their way at least for the time being.
2014-05-09 00:20:54 +00:00
if (_currentBindState.varyingBuffer)
_deviceContext->CSSetUnorderedAccessViews(1, 1, &_currentBindState.varyingBuffer, 0); // u1
Reorganize ComputeContext and ComputeController. Moved transient states (current vertex buffer etc) to controller. ComputeContext becomes constant so that it's well suited for coarse-grain parallelism on cpu. The prims sharing same topology (ComputeContext) can be refined simultaneously by having mutiple compute controllers. Client facing API doesn't change.
2014-05-06 15:53:36 +00:00
}
void
OsdD3D11ComputeController::unbindShaderResources()
{
ID3D11UnorderedAccessView *UAViews[] = { 0, 0 };
_deviceContext->CSSetUnorderedAccessViews(0, 2, UAViews, 0); // u0-u2
}
Batching stuffs: generalized kernel batches, table/dispatcher refactoring, multiMeshFactory, drawContext, etc. 2 client APIs are changed. - VertexBuffer::UpdateData() takes start vertex offset - ComputeController::Refine() takes FarKernelBatchVector Also, ComputeContext no longer holds farmesh. Client can free farmesh after OsdComputeContext is created. (but still need FarKernelBatchVector to apply subdivision kernels)
2013-03-08 01:50:15 +00:00
void
OsdD3D11ComputeController::ApplyBilinearFaceVerticesKernel(
Reorganize ComputeContext and ComputeController. Moved transient states (current vertex buffer etc) to controller. ComputeContext becomes constant so that it's well suited for coarse-grain parallelism on cpu. The prims sharing same topology (ComputeContext) can be refined simultaneously by having mutiple compute controllers. Client facing API doesn't change.
2014-05-06 15:53:36 +00:00
FarKernelBatch const &batch, OsdD3D11ComputeContext const *context) const {
Batching stuffs: generalized kernel batches, table/dispatcher refactoring, multiMeshFactory, drawContext, etc. 2 client APIs are changed. - VertexBuffer::UpdateData() takes start vertex offset - ComputeController::Refine() takes FarKernelBatchVector Also, ComputeContext no longer holds farmesh. Client can free farmesh after OsdComputeContext is created. (but still need FarKernelBatchVector to apply subdivision kernels)
2013-03-08 01:50:15 +00:00
assert(context);
Interleaved buffer support in OsdCompute. Removed OsdVertexDescriptor and replaced with OsdVertexBufferDescriptor. All kernels take offset/length/stride to apply subdivision partially in each vertex elements. Also the offset can be used for client-based VBO aggregation, without modifying index buffers. This is useful for topology sharing, in conjunction with glDrawElementsBaseVertex etc. However, gregory patch shader fetches vertex buffer via texture buffer, which index should also be offsetted too. Although gl_BaseVertexARB extension should be able to do that job, it's a relatively new extension. So we use OsdBaseVertex() call to mitigate the compatibility issue as clients can provide it in their way at least for the time being.
2014-05-09 00:20:54 +00:00
_currentBindState.kernelBundle->ApplyBilinearFaceVerticesKernel(
batch.GetVertexOffset(), batch.GetTableOffset(),
batch.GetStart(), batch.GetEnd(),
_currentBindState.vertexDesc.offset, _currentBindState.varyingDesc.offset);
Batching stuffs: generalized kernel batches, table/dispatcher refactoring, multiMeshFactory, drawContext, etc. 2 client APIs are changed. - VertexBuffer::UpdateData() takes start vertex offset - ComputeController::Refine() takes FarKernelBatchVector Also, ComputeContext no longer holds farmesh. Client can free farmesh after OsdComputeContext is created. (but still need FarKernelBatchVector to apply subdivision kernels)
2013-03-08 01:50:15 +00:00
}
void
OsdD3D11ComputeController::ApplyBilinearEdgeVerticesKernel(
Reorganize ComputeContext and ComputeController. Moved transient states (current vertex buffer etc) to controller. ComputeContext becomes constant so that it's well suited for coarse-grain parallelism on cpu. The prims sharing same topology (ComputeContext) can be refined simultaneously by having mutiple compute controllers. Client facing API doesn't change.
2014-05-06 15:53:36 +00:00
FarKernelBatch const &batch, OsdD3D11ComputeContext const *context) const {
Batching stuffs: generalized kernel batches, table/dispatcher refactoring, multiMeshFactory, drawContext, etc. 2 client APIs are changed. - VertexBuffer::UpdateData() takes start vertex offset - ComputeController::Refine() takes FarKernelBatchVector Also, ComputeContext no longer holds farmesh. Client can free farmesh after OsdComputeContext is created. (but still need FarKernelBatchVector to apply subdivision kernels)
2013-03-08 01:50:15 +00:00
assert(context);
Interleaved buffer support in OsdCompute. Removed OsdVertexDescriptor and replaced with OsdVertexBufferDescriptor. All kernels take offset/length/stride to apply subdivision partially in each vertex elements. Also the offset can be used for client-based VBO aggregation, without modifying index buffers. This is useful for topology sharing, in conjunction with glDrawElementsBaseVertex etc. However, gregory patch shader fetches vertex buffer via texture buffer, which index should also be offsetted too. Although gl_BaseVertexARB extension should be able to do that job, it's a relatively new extension. So we use OsdBaseVertex() call to mitigate the compatibility issue as clients can provide it in their way at least for the time being.
2014-05-09 00:20:54 +00:00
_currentBindState.kernelBundle->ApplyBilinearEdgeVerticesKernel(
batch.GetVertexOffset(), batch.GetTableOffset(),
batch.GetStart(), batch.GetEnd(),
_currentBindState.vertexDesc.offset, _currentBindState.varyingDesc.offset);
Batching stuffs: generalized kernel batches, table/dispatcher refactoring, multiMeshFactory, drawContext, etc. 2 client APIs are changed. - VertexBuffer::UpdateData() takes start vertex offset - ComputeController::Refine() takes FarKernelBatchVector Also, ComputeContext no longer holds farmesh. Client can free farmesh after OsdComputeContext is created. (but still need FarKernelBatchVector to apply subdivision kernels)
2013-03-08 01:50:15 +00:00
}
void
OsdD3D11ComputeController::ApplyBilinearVertexVerticesKernel(
Reorganize ComputeContext and ComputeController. Moved transient states (current vertex buffer etc) to controller. ComputeContext becomes constant so that it's well suited for coarse-grain parallelism on cpu. The prims sharing same topology (ComputeContext) can be refined simultaneously by having mutiple compute controllers. Client facing API doesn't change.
2014-05-06 15:53:36 +00:00
FarKernelBatch const &batch, OsdD3D11ComputeContext const *context) const {
Batching stuffs: generalized kernel batches, table/dispatcher refactoring, multiMeshFactory, drawContext, etc. 2 client APIs are changed. - VertexBuffer::UpdateData() takes start vertex offset - ComputeController::Refine() takes FarKernelBatchVector Also, ComputeContext no longer holds farmesh. Client can free farmesh after OsdComputeContext is created. (but still need FarKernelBatchVector to apply subdivision kernels)
2013-03-08 01:50:15 +00:00
assert(context);
Interleaved buffer support in OsdCompute. Removed OsdVertexDescriptor and replaced with OsdVertexBufferDescriptor. All kernels take offset/length/stride to apply subdivision partially in each vertex elements. Also the offset can be used for client-based VBO aggregation, without modifying index buffers. This is useful for topology sharing, in conjunction with glDrawElementsBaseVertex etc. However, gregory patch shader fetches vertex buffer via texture buffer, which index should also be offsetted too. Although gl_BaseVertexARB extension should be able to do that job, it's a relatively new extension. So we use OsdBaseVertex() call to mitigate the compatibility issue as clients can provide it in their way at least for the time being.
2014-05-09 00:20:54 +00:00
_currentBindState.kernelBundle->ApplyBilinearVertexVerticesKernel(
batch.GetVertexOffset(), batch.GetTableOffset(),
batch.GetStart(), batch.GetEnd(),
_currentBindState.vertexDesc.offset, _currentBindState.varyingDesc.offset);
Batching stuffs: generalized kernel batches, table/dispatcher refactoring, multiMeshFactory, drawContext, etc. 2 client APIs are changed. - VertexBuffer::UpdateData() takes start vertex offset - ComputeController::Refine() takes FarKernelBatchVector Also, ComputeContext no longer holds farmesh. Client can free farmesh after OsdComputeContext is created. (but still need FarKernelBatchVector to apply subdivision kernels)
2013-03-08 01:50:15 +00:00
}
void
OsdD3D11ComputeController::ApplyCatmarkFaceVerticesKernel(
Reorganize ComputeContext and ComputeController. Moved transient states (current vertex buffer etc) to controller. ComputeContext becomes constant so that it's well suited for coarse-grain parallelism on cpu. The prims sharing same topology (ComputeContext) can be refined simultaneously by having mutiple compute controllers. Client facing API doesn't change.
2014-05-06 15:53:36 +00:00
FarKernelBatch const &batch, OsdD3D11ComputeContext const *context) const {
Batching stuffs: generalized kernel batches, table/dispatcher refactoring, multiMeshFactory, drawContext, etc. 2 client APIs are changed. - VertexBuffer::UpdateData() takes start vertex offset - ComputeController::Refine() takes FarKernelBatchVector Also, ComputeContext no longer holds farmesh. Client can free farmesh after OsdComputeContext is created. (but still need FarKernelBatchVector to apply subdivision kernels)
2013-03-08 01:50:15 +00:00
assert(context);
Interleaved buffer support in OsdCompute. Removed OsdVertexDescriptor and replaced with OsdVertexBufferDescriptor. All kernels take offset/length/stride to apply subdivision partially in each vertex elements. Also the offset can be used for client-based VBO aggregation, without modifying index buffers. This is useful for topology sharing, in conjunction with glDrawElementsBaseVertex etc. However, gregory patch shader fetches vertex buffer via texture buffer, which index should also be offsetted too. Although gl_BaseVertexARB extension should be able to do that job, it's a relatively new extension. So we use OsdBaseVertex() call to mitigate the compatibility issue as clients can provide it in their way at least for the time being.
2014-05-09 00:20:54 +00:00
_currentBindState.kernelBundle->ApplyCatmarkFaceVerticesKernel(
batch.GetVertexOffset(), batch.GetTableOffset(),
batch.GetStart(), batch.GetEnd(),
_currentBindState.vertexDesc.offset, _currentBindState.varyingDesc.offset);
Batching stuffs: generalized kernel batches, table/dispatcher refactoring, multiMeshFactory, drawContext, etc. 2 client APIs are changed. - VertexBuffer::UpdateData() takes start vertex offset - ComputeController::Refine() takes FarKernelBatchVector Also, ComputeContext no longer holds farmesh. Client can free farmesh after OsdComputeContext is created. (but still need FarKernelBatchVector to apply subdivision kernels)
2013-03-08 01:50:15 +00:00
}
Implemented CATMARK_QUAD_FACE_VERTEX and CATMARK_TRI_QUAD_FACE_VERTEX kernels for DirectX 11, OpenMP, and TBB platforms. Fixed a bug in the GLSL Transform Feedback kernels. Minor changes for consistent formatting.
2014-05-30 06:02:19 +00:00
void
OsdD3D11ComputeController::ApplyCatmarkQuadFaceVerticesKernel(
FarKernelBatch const &batch, OsdD3D11ComputeContext const *context) const {
assert(context);
_currentBindState.kernelBundle->ApplyCatmarkQuadFaceVerticesKernel(
batch.GetVertexOffset(), batch.GetTableOffset(),
batch.GetStart(), batch.GetEnd(),
_currentBindState.vertexDesc.offset, _currentBindState.varyingDesc.offset);
}
Batching stuffs: generalized kernel batches, table/dispatcher refactoring, multiMeshFactory, drawContext, etc. 2 client APIs are changed. - VertexBuffer::UpdateData() takes start vertex offset - ComputeController::Refine() takes FarKernelBatchVector Also, ComputeContext no longer holds farmesh. Client can free farmesh after OsdComputeContext is created. (but still need FarKernelBatchVector to apply subdivision kernels)
2013-03-08 01:50:15 +00:00
Implemented CATMARK_QUAD_FACE_VERTEX and CATMARK_TRI_QUAD_FACE_VERTEX kernels for DirectX 11, OpenMP, and TBB platforms. Fixed a bug in the GLSL Transform Feedback kernels. Minor changes for consistent formatting.
2014-05-30 06:02:19 +00:00
void
OsdD3D11ComputeController::ApplyCatmarkTriQuadFaceVerticesKernel(
FarKernelBatch const &batch, OsdD3D11ComputeContext const *context) const {
assert(context);
_currentBindState.kernelBundle->ApplyCatmarkTriQuadFaceVerticesKernel(
batch.GetVertexOffset(), batch.GetTableOffset(),
batch.GetStart(), batch.GetEnd(),
_currentBindState.vertexDesc.offset, _currentBindState.varyingDesc.offset);
}
Batching stuffs: generalized kernel batches, table/dispatcher refactoring, multiMeshFactory, drawContext, etc. 2 client APIs are changed. - VertexBuffer::UpdateData() takes start vertex offset - ComputeController::Refine() takes FarKernelBatchVector Also, ComputeContext no longer holds farmesh. Client can free farmesh after OsdComputeContext is created. (but still need FarKernelBatchVector to apply subdivision kernels)
2013-03-08 01:50:15 +00:00
void
OsdD3D11ComputeController::ApplyCatmarkEdgeVerticesKernel(
Reorganize ComputeContext and ComputeController. Moved transient states (current vertex buffer etc) to controller. ComputeContext becomes constant so that it's well suited for coarse-grain parallelism on cpu. The prims sharing same topology (ComputeContext) can be refined simultaneously by having mutiple compute controllers. Client facing API doesn't change.
2014-05-06 15:53:36 +00:00
FarKernelBatch const &batch, OsdD3D11ComputeContext const *context) const {
Batching stuffs: generalized kernel batches, table/dispatcher refactoring, multiMeshFactory, drawContext, etc. 2 client APIs are changed. - VertexBuffer::UpdateData() takes start vertex offset - ComputeController::Refine() takes FarKernelBatchVector Also, ComputeContext no longer holds farmesh. Client can free farmesh after OsdComputeContext is created. (but still need FarKernelBatchVector to apply subdivision kernels)
2013-03-08 01:50:15 +00:00
assert(context);
Interleaved buffer support in OsdCompute. Removed OsdVertexDescriptor and replaced with OsdVertexBufferDescriptor. All kernels take offset/length/stride to apply subdivision partially in each vertex elements. Also the offset can be used for client-based VBO aggregation, without modifying index buffers. This is useful for topology sharing, in conjunction with glDrawElementsBaseVertex etc. However, gregory patch shader fetches vertex buffer via texture buffer, which index should also be offsetted too. Although gl_BaseVertexARB extension should be able to do that job, it's a relatively new extension. So we use OsdBaseVertex() call to mitigate the compatibility issue as clients can provide it in their way at least for the time being.
2014-05-09 00:20:54 +00:00
_currentBindState.kernelBundle->ApplyCatmarkEdgeVerticesKernel(
batch.GetVertexOffset(), batch.GetTableOffset(),
batch.GetStart(), batch.GetEnd(),
_currentBindState.vertexDesc.offset, _currentBindState.varyingDesc.offset);
Batching stuffs: generalized kernel batches, table/dispatcher refactoring, multiMeshFactory, drawContext, etc. 2 client APIs are changed. - VertexBuffer::UpdateData() takes start vertex offset - ComputeController::Refine() takes FarKernelBatchVector Also, ComputeContext no longer holds farmesh. Client can free farmesh after OsdComputeContext is created. (but still need FarKernelBatchVector to apply subdivision kernels)
2013-03-08 01:50:15 +00:00
}
void
OsdD3D11ComputeController::ApplyCatmarkVertexVerticesKernelB(
Reorganize ComputeContext and ComputeController. Moved transient states (current vertex buffer etc) to controller. ComputeContext becomes constant so that it's well suited for coarse-grain parallelism on cpu. The prims sharing same topology (ComputeContext) can be refined simultaneously by having mutiple compute controllers. Client facing API doesn't change.
2014-05-06 15:53:36 +00:00
FarKernelBatch const &batch, OsdD3D11ComputeContext const *context) const {
Batching stuffs: generalized kernel batches, table/dispatcher refactoring, multiMeshFactory, drawContext, etc. 2 client APIs are changed. - VertexBuffer::UpdateData() takes start vertex offset - ComputeController::Refine() takes FarKernelBatchVector Also, ComputeContext no longer holds farmesh. Client can free farmesh after OsdComputeContext is created. (but still need FarKernelBatchVector to apply subdivision kernels)
2013-03-08 01:50:15 +00:00
assert(context);
Interleaved buffer support in OsdCompute. Removed OsdVertexDescriptor and replaced with OsdVertexBufferDescriptor. All kernels take offset/length/stride to apply subdivision partially in each vertex elements. Also the offset can be used for client-based VBO aggregation, without modifying index buffers. This is useful for topology sharing, in conjunction with glDrawElementsBaseVertex etc. However, gregory patch shader fetches vertex buffer via texture buffer, which index should also be offsetted too. Although gl_BaseVertexARB extension should be able to do that job, it's a relatively new extension. So we use OsdBaseVertex() call to mitigate the compatibility issue as clients can provide it in their way at least for the time being.
2014-05-09 00:20:54 +00:00
_currentBindState.kernelBundle->ApplyCatmarkVertexVerticesKernelB(
batch.GetVertexOffset(), batch.GetTableOffset(),
batch.GetStart(), batch.GetEnd(),
_currentBindState.vertexDesc.offset, _currentBindState.varyingDesc.offset);
Batching stuffs: generalized kernel batches, table/dispatcher refactoring, multiMeshFactory, drawContext, etc. 2 client APIs are changed. - VertexBuffer::UpdateData() takes start vertex offset - ComputeController::Refine() takes FarKernelBatchVector Also, ComputeContext no longer holds farmesh. Client can free farmesh after OsdComputeContext is created. (but still need FarKernelBatchVector to apply subdivision kernels)
2013-03-08 01:50:15 +00:00
}
void
OsdD3D11ComputeController::ApplyCatmarkVertexVerticesKernelA1(
Reorganize ComputeContext and ComputeController. Moved transient states (current vertex buffer etc) to controller. ComputeContext becomes constant so that it's well suited for coarse-grain parallelism on cpu. The prims sharing same topology (ComputeContext) can be refined simultaneously by having mutiple compute controllers. Client facing API doesn't change.
2014-05-06 15:53:36 +00:00
FarKernelBatch const &batch, OsdD3D11ComputeContext const *context) const {
Batching stuffs: generalized kernel batches, table/dispatcher refactoring, multiMeshFactory, drawContext, etc. 2 client APIs are changed. - VertexBuffer::UpdateData() takes start vertex offset - ComputeController::Refine() takes FarKernelBatchVector Also, ComputeContext no longer holds farmesh. Client can free farmesh after OsdComputeContext is created. (but still need FarKernelBatchVector to apply subdivision kernels)
2013-03-08 01:50:15 +00:00
assert(context);
Interleaved buffer support in OsdCompute. Removed OsdVertexDescriptor and replaced with OsdVertexBufferDescriptor. All kernels take offset/length/stride to apply subdivision partially in each vertex elements. Also the offset can be used for client-based VBO aggregation, without modifying index buffers. This is useful for topology sharing, in conjunction with glDrawElementsBaseVertex etc. However, gregory patch shader fetches vertex buffer via texture buffer, which index should also be offsetted too. Although gl_BaseVertexARB extension should be able to do that job, it's a relatively new extension. So we use OsdBaseVertex() call to mitigate the compatibility issue as clients can provide it in their way at least for the time being.
2014-05-09 00:20:54 +00:00
_currentBindState.kernelBundle->ApplyCatmarkVertexVerticesKernelA(
batch.GetVertexOffset(), batch.GetTableOffset(),
batch.GetStart(), batch.GetEnd(), false,
_currentBindState.vertexDesc.offset, _currentBindState.varyingDesc.offset);
Batching stuffs: generalized kernel batches, table/dispatcher refactoring, multiMeshFactory, drawContext, etc. 2 client APIs are changed. - VertexBuffer::UpdateData() takes start vertex offset - ComputeController::Refine() takes FarKernelBatchVector Also, ComputeContext no longer holds farmesh. Client can free farmesh after OsdComputeContext is created. (but still need FarKernelBatchVector to apply subdivision kernels)
2013-03-08 01:50:15 +00:00
}
void
OsdD3D11ComputeController::ApplyCatmarkVertexVerticesKernelA2(
Reorganize ComputeContext and ComputeController. Moved transient states (current vertex buffer etc) to controller. ComputeContext becomes constant so that it's well suited for coarse-grain parallelism on cpu. The prims sharing same topology (ComputeContext) can be refined simultaneously by having mutiple compute controllers. Client facing API doesn't change.
2014-05-06 15:53:36 +00:00
FarKernelBatch const &batch, OsdD3D11ComputeContext const *context) const {
Batching stuffs: generalized kernel batches, table/dispatcher refactoring, multiMeshFactory, drawContext, etc. 2 client APIs are changed. - VertexBuffer::UpdateData() takes start vertex offset - ComputeController::Refine() takes FarKernelBatchVector Also, ComputeContext no longer holds farmesh. Client can free farmesh after OsdComputeContext is created. (but still need FarKernelBatchVector to apply subdivision kernels)
2013-03-08 01:50:15 +00:00
assert(context);
Interleaved buffer support in OsdCompute. Removed OsdVertexDescriptor and replaced with OsdVertexBufferDescriptor. All kernels take offset/length/stride to apply subdivision partially in each vertex elements. Also the offset can be used for client-based VBO aggregation, without modifying index buffers. This is useful for topology sharing, in conjunction with glDrawElementsBaseVertex etc. However, gregory patch shader fetches vertex buffer via texture buffer, which index should also be offsetted too. Although gl_BaseVertexARB extension should be able to do that job, it's a relatively new extension. So we use OsdBaseVertex() call to mitigate the compatibility issue as clients can provide it in their way at least for the time being.
2014-05-09 00:20:54 +00:00
_currentBindState.kernelBundle->ApplyCatmarkVertexVerticesKernelA(
batch.GetVertexOffset(), batch.GetTableOffset(),
batch.GetStart(), batch.GetEnd(), true,
_currentBindState.vertexDesc.offset, _currentBindState.varyingDesc.offset);
Batching stuffs: generalized kernel batches, table/dispatcher refactoring, multiMeshFactory, drawContext, etc. 2 client APIs are changed. - VertexBuffer::UpdateData() takes start vertex offset - ComputeController::Refine() takes FarKernelBatchVector Also, ComputeContext no longer holds farmesh. Client can free farmesh after OsdComputeContext is created. (but still need FarKernelBatchVector to apply subdivision kernels)
2013-03-08 01:50:15 +00:00
}
void
OsdD3D11ComputeController::ApplyLoopEdgeVerticesKernel(
Reorganize ComputeContext and ComputeController. Moved transient states (current vertex buffer etc) to controller. ComputeContext becomes constant so that it's well suited for coarse-grain parallelism on cpu. The prims sharing same topology (ComputeContext) can be refined simultaneously by having mutiple compute controllers. Client facing API doesn't change.
2014-05-06 15:53:36 +00:00
FarKernelBatch const &batch, OsdD3D11ComputeContext const *context) const {
Batching stuffs: generalized kernel batches, table/dispatcher refactoring, multiMeshFactory, drawContext, etc. 2 client APIs are changed. - VertexBuffer::UpdateData() takes start vertex offset - ComputeController::Refine() takes FarKernelBatchVector Also, ComputeContext no longer holds farmesh. Client can free farmesh after OsdComputeContext is created. (but still need FarKernelBatchVector to apply subdivision kernels)
2013-03-08 01:50:15 +00:00
assert(context);
Interleaved buffer support in OsdCompute. Removed OsdVertexDescriptor and replaced with OsdVertexBufferDescriptor. All kernels take offset/length/stride to apply subdivision partially in each vertex elements. Also the offset can be used for client-based VBO aggregation, without modifying index buffers. This is useful for topology sharing, in conjunction with glDrawElementsBaseVertex etc. However, gregory patch shader fetches vertex buffer via texture buffer, which index should also be offsetted too. Although gl_BaseVertexARB extension should be able to do that job, it's a relatively new extension. So we use OsdBaseVertex() call to mitigate the compatibility issue as clients can provide it in their way at least for the time being.
2014-05-09 00:20:54 +00:00
_currentBindState.kernelBundle->ApplyLoopEdgeVerticesKernel(
batch.GetVertexOffset(), batch.GetTableOffset(),
batch.GetStart(), batch.GetEnd(),
_currentBindState.vertexDesc.offset, _currentBindState.varyingDesc.offset);
Batching stuffs: generalized kernel batches, table/dispatcher refactoring, multiMeshFactory, drawContext, etc. 2 client APIs are changed. - VertexBuffer::UpdateData() takes start vertex offset - ComputeController::Refine() takes FarKernelBatchVector Also, ComputeContext no longer holds farmesh. Client can free farmesh after OsdComputeContext is created. (but still need FarKernelBatchVector to apply subdivision kernels)
2013-03-08 01:50:15 +00:00
}
void
OsdD3D11ComputeController::ApplyLoopVertexVerticesKernelB(
Reorganize ComputeContext and ComputeController. Moved transient states (current vertex buffer etc) to controller. ComputeContext becomes constant so that it's well suited for coarse-grain parallelism on cpu. The prims sharing same topology (ComputeContext) can be refined simultaneously by having mutiple compute controllers. Client facing API doesn't change.
2014-05-06 15:53:36 +00:00
FarKernelBatch const &batch, OsdD3D11ComputeContext const *context) const {
Batching stuffs: generalized kernel batches, table/dispatcher refactoring, multiMeshFactory, drawContext, etc. 2 client APIs are changed. - VertexBuffer::UpdateData() takes start vertex offset - ComputeController::Refine() takes FarKernelBatchVector Also, ComputeContext no longer holds farmesh. Client can free farmesh after OsdComputeContext is created. (but still need FarKernelBatchVector to apply subdivision kernels)
2013-03-08 01:50:15 +00:00
assert(context);
Interleaved buffer support in OsdCompute. Removed OsdVertexDescriptor and replaced with OsdVertexBufferDescriptor. All kernels take offset/length/stride to apply subdivision partially in each vertex elements. Also the offset can be used for client-based VBO aggregation, without modifying index buffers. This is useful for topology sharing, in conjunction with glDrawElementsBaseVertex etc. However, gregory patch shader fetches vertex buffer via texture buffer, which index should also be offsetted too. Although gl_BaseVertexARB extension should be able to do that job, it's a relatively new extension. So we use OsdBaseVertex() call to mitigate the compatibility issue as clients can provide it in their way at least for the time being.
2014-05-09 00:20:54 +00:00
_currentBindState.kernelBundle->ApplyLoopVertexVerticesKernelB(
batch.GetVertexOffset(), batch.GetTableOffset(),
batch.GetStart(), batch.GetEnd(),
_currentBindState.vertexDesc.offset, _currentBindState.varyingDesc.offset);
Batching stuffs: generalized kernel batches, table/dispatcher refactoring, multiMeshFactory, drawContext, etc. 2 client APIs are changed. - VertexBuffer::UpdateData() takes start vertex offset - ComputeController::Refine() takes FarKernelBatchVector Also, ComputeContext no longer holds farmesh. Client can free farmesh after OsdComputeContext is created. (but still need FarKernelBatchVector to apply subdivision kernels)
2013-03-08 01:50:15 +00:00
}
void
OsdD3D11ComputeController::ApplyLoopVertexVerticesKernelA1(
Reorganize ComputeContext and ComputeController. Moved transient states (current vertex buffer etc) to controller. ComputeContext becomes constant so that it's well suited for coarse-grain parallelism on cpu. The prims sharing same topology (ComputeContext) can be refined simultaneously by having mutiple compute controllers. Client facing API doesn't change.
2014-05-06 15:53:36 +00:00
FarKernelBatch const &batch, OsdD3D11ComputeContext const *context) const {
Batching stuffs: generalized kernel batches, table/dispatcher refactoring, multiMeshFactory, drawContext, etc. 2 client APIs are changed. - VertexBuffer::UpdateData() takes start vertex offset - ComputeController::Refine() takes FarKernelBatchVector Also, ComputeContext no longer holds farmesh. Client can free farmesh after OsdComputeContext is created. (but still need FarKernelBatchVector to apply subdivision kernels)
2013-03-08 01:50:15 +00:00
assert(context);
Interleaved buffer support in OsdCompute. Removed OsdVertexDescriptor and replaced with OsdVertexBufferDescriptor. All kernels take offset/length/stride to apply subdivision partially in each vertex elements. Also the offset can be used for client-based VBO aggregation, without modifying index buffers. This is useful for topology sharing, in conjunction with glDrawElementsBaseVertex etc. However, gregory patch shader fetches vertex buffer via texture buffer, which index should also be offsetted too. Although gl_BaseVertexARB extension should be able to do that job, it's a relatively new extension. So we use OsdBaseVertex() call to mitigate the compatibility issue as clients can provide it in their way at least for the time being.
2014-05-09 00:20:54 +00:00
_currentBindState.kernelBundle->ApplyLoopVertexVerticesKernelA(
batch.GetVertexOffset(), batch.GetTableOffset(),
batch.GetStart(), batch.GetEnd(), false,
_currentBindState.vertexDesc.offset, _currentBindState.varyingDesc.offset);
Batching stuffs: generalized kernel batches, table/dispatcher refactoring, multiMeshFactory, drawContext, etc. 2 client APIs are changed. - VertexBuffer::UpdateData() takes start vertex offset - ComputeController::Refine() takes FarKernelBatchVector Also, ComputeContext no longer holds farmesh. Client can free farmesh after OsdComputeContext is created. (but still need FarKernelBatchVector to apply subdivision kernels)
2013-03-08 01:50:15 +00:00
}
void
OsdD3D11ComputeController::ApplyLoopVertexVerticesKernelA2(
Reorganize ComputeContext and ComputeController. Moved transient states (current vertex buffer etc) to controller. ComputeContext becomes constant so that it's well suited for coarse-grain parallelism on cpu. The prims sharing same topology (ComputeContext) can be refined simultaneously by having mutiple compute controllers. Client facing API doesn't change.
2014-05-06 15:53:36 +00:00
FarKernelBatch const &batch, OsdD3D11ComputeContext const *context) const {
Batching stuffs: generalized kernel batches, table/dispatcher refactoring, multiMeshFactory, drawContext, etc. 2 client APIs are changed. - VertexBuffer::UpdateData() takes start vertex offset - ComputeController::Refine() takes FarKernelBatchVector Also, ComputeContext no longer holds farmesh. Client can free farmesh after OsdComputeContext is created. (but still need FarKernelBatchVector to apply subdivision kernels)
2013-03-08 01:50:15 +00:00
assert(context);
Interleaved buffer support in OsdCompute. Removed OsdVertexDescriptor and replaced with OsdVertexBufferDescriptor. All kernels take offset/length/stride to apply subdivision partially in each vertex elements. Also the offset can be used for client-based VBO aggregation, without modifying index buffers. This is useful for topology sharing, in conjunction with glDrawElementsBaseVertex etc. However, gregory patch shader fetches vertex buffer via texture buffer, which index should also be offsetted too. Although gl_BaseVertexARB extension should be able to do that job, it's a relatively new extension. So we use OsdBaseVertex() call to mitigate the compatibility issue as clients can provide it in their way at least for the time being.
2014-05-09 00:20:54 +00:00
_currentBindState.kernelBundle->ApplyLoopVertexVerticesKernelA(
batch.GetVertexOffset(), batch.GetTableOffset(),
batch.GetStart(), batch.GetEnd(), true,
_currentBindState.vertexDesc.offset, _currentBindState.varyingDesc.offset);
Batching stuffs: generalized kernel batches, table/dispatcher refactoring, multiMeshFactory, drawContext, etc. 2 client APIs are changed. - VertexBuffer::UpdateData() takes start vertex offset - ComputeController::Refine() takes FarKernelBatchVector Also, ComputeContext no longer holds farmesh. Client can free farmesh after OsdComputeContext is created. (but still need FarKernelBatchVector to apply subdivision kernels)
2013-03-08 01:50:15 +00:00
}
void
OsdD3D11ComputeController::ApplyVertexEdits(
Reorganize ComputeContext and ComputeController. Moved transient states (current vertex buffer etc) to controller. ComputeContext becomes constant so that it's well suited for coarse-grain parallelism on cpu. The prims sharing same topology (ComputeContext) can be refined simultaneously by having mutiple compute controllers. Client facing API doesn't change.
2014-05-06 15:53:36 +00:00
FarKernelBatch const &batch, OsdD3D11ComputeContext const *context) const {
Batching stuffs: generalized kernel batches, table/dispatcher refactoring, multiMeshFactory, drawContext, etc. 2 client APIs are changed. - VertexBuffer::UpdateData() takes start vertex offset - ComputeController::Refine() takes FarKernelBatchVector Also, ComputeContext no longer holds farmesh. Client can free farmesh after OsdComputeContext is created. (but still need FarKernelBatchVector to apply subdivision kernels)
2013-03-08 01:50:15 +00:00
assert(context);
Fixed DX11 build errors from KernelBatch refactoring
2013-05-07 06:55:45 +00:00
const OsdD3D11ComputeHEditTable * edit = context->GetEditTable(batch.GetTableIndex());
Batching stuffs: generalized kernel batches, table/dispatcher refactoring, multiMeshFactory, drawContext, etc. 2 client APIs are changed. - VertexBuffer::UpdateData() takes start vertex offset - ComputeController::Refine() takes FarKernelBatchVector Also, ComputeContext no longer holds farmesh. Client can free farmesh after OsdComputeContext is created. (but still need FarKernelBatchVector to apply subdivision kernels)
2013-03-08 01:50:15 +00:00
assert(edit);
Reorganize ComputeContext and ComputeController. Moved transient states (current vertex buffer etc) to controller. ComputeContext becomes constant so that it's well suited for coarse-grain parallelism on cpu. The prims sharing same topology (ComputeContext) can be refined simultaneously by having mutiple compute controllers. Client facing API doesn't change.
2014-05-06 15:53:36 +00:00
context->BindEditShaderStorageBuffers(batch.GetTableIndex(), _deviceContext);
Batching stuffs: generalized kernel batches, table/dispatcher refactoring, multiMeshFactory, drawContext, etc. 2 client APIs are changed. - VertexBuffer::UpdateData() takes start vertex offset - ComputeController::Refine() takes FarKernelBatchVector Also, ComputeContext no longer holds farmesh. Client can free farmesh after OsdComputeContext is created. (but still need FarKernelBatchVector to apply subdivision kernels)
2013-03-08 01:50:15 +00:00
int primvarOffset = edit->GetPrimvarOffset();
int primvarWidth = edit->GetPrimvarWidth();
Reorganize ComputeContext and ComputeController. Moved transient states (current vertex buffer etc) to controller. ComputeContext becomes constant so that it's well suited for coarse-grain parallelism on cpu. The prims sharing same topology (ComputeContext) can be refined simultaneously by having mutiple compute controllers. Client facing API doesn't change.
2014-05-06 15:53:36 +00:00
Batching stuffs: generalized kernel batches, table/dispatcher refactoring, multiMeshFactory, drawContext, etc. 2 client APIs are changed. - VertexBuffer::UpdateData() takes start vertex offset - ComputeController::Refine() takes FarKernelBatchVector Also, ComputeContext no longer holds farmesh. Client can free farmesh after OsdComputeContext is created. (but still need FarKernelBatchVector to apply subdivision kernels)
2013-03-08 01:50:15 +00:00
if (edit->GetOperation() == FarVertexEdit::Add) {
Interleaved buffer support in OsdCompute. Removed OsdVertexDescriptor and replaced with OsdVertexBufferDescriptor. All kernels take offset/length/stride to apply subdivision partially in each vertex elements. Also the offset can be used for client-based VBO aggregation, without modifying index buffers. This is useful for topology sharing, in conjunction with glDrawElementsBaseVertex etc. However, gregory patch shader fetches vertex buffer via texture buffer, which index should also be offsetted too. Although gl_BaseVertexARB extension should be able to do that job, it's a relatively new extension. So we use OsdBaseVertex() call to mitigate the compatibility issue as clients can provide it in their way at least for the time being.
2014-05-09 00:20:54 +00:00
_currentBindState.kernelBundle->ApplyEditAdd(primvarOffset, primvarWidth,
Reorganize ComputeContext and ComputeController. Moved transient states (current vertex buffer etc) to controller. ComputeContext becomes constant so that it's well suited for coarse-grain parallelism on cpu. The prims sharing same topology (ComputeContext) can be refined simultaneously by having mutiple compute controllers. Client facing API doesn't change.
2014-05-06 15:53:36 +00:00
batch.GetVertexOffset(),
batch.GetTableOffset(),
batch.GetStart(),
Interleaved buffer support in OsdCompute. Removed OsdVertexDescriptor and replaced with OsdVertexBufferDescriptor. All kernels take offset/length/stride to apply subdivision partially in each vertex elements. Also the offset can be used for client-based VBO aggregation, without modifying index buffers. This is useful for topology sharing, in conjunction with glDrawElementsBaseVertex etc. However, gregory patch shader fetches vertex buffer via texture buffer, which index should also be offsetted too. Although gl_BaseVertexARB extension should be able to do that job, it's a relatively new extension. So we use OsdBaseVertex() call to mitigate the compatibility issue as clients can provide it in their way at least for the time being.
2014-05-09 00:20:54 +00:00
batch.GetEnd(),
_currentBindState.vertexDesc.offset,
_currentBindState.varyingDesc.offset);
Batching stuffs: generalized kernel batches, table/dispatcher refactoring, multiMeshFactory, drawContext, etc. 2 client APIs are changed. - VertexBuffer::UpdateData() takes start vertex offset - ComputeController::Refine() takes FarKernelBatchVector Also, ComputeContext no longer holds farmesh. Client can free farmesh after OsdComputeContext is created. (but still need FarKernelBatchVector to apply subdivision kernels)
2013-03-08 01:50:15 +00:00
} else {
// XXX: edit SET is not implemented yet.
}
Reorganize ComputeContext and ComputeController. Moved transient states (current vertex buffer etc) to controller. ComputeContext becomes constant so that it's well suited for coarse-grain parallelism on cpu. The prims sharing same topology (ComputeContext) can be refined simultaneously by having mutiple compute controllers. Client facing API doesn't change.
2014-05-06 15:53:36 +00:00
context->UnbindEditShaderStorageBuffers(_deviceContext);
Batching stuffs: generalized kernel batches, table/dispatcher refactoring, multiMeshFactory, drawContext, etc. 2 client APIs are changed. - VertexBuffer::UpdateData() takes start vertex offset - ComputeController::Refine() takes FarKernelBatchVector Also, ComputeContext no longer holds farmesh. Client can free farmesh after OsdComputeContext is created. (but still need FarKernelBatchVector to apply subdivision kernels)
2013-03-08 01:50:15 +00:00
}
Added DirectCompute backend for Osd Compute
2012-12-13 18:22:30 +00:00
} // end namespace OPENSUBDIV_VERSION
} // end namespace OpenSubdiv
Reference in New Issue Copy Permalink