OpenSubdiv/opensubdiv/osd/glslComputeController.cpp
manuelk a16d558a3e Fix GLSL compute code version defintion
- remove #version declaration from the kernel code
- move it in front of shader sources before compiling to prevent some drivers from throwing errors

fixes #360
2014-10-20 18:09:30 -07:00

287 lines
8.6 KiB
C++

//
// 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.
//
#include "../osd/glslComputeController.h"
#include "../osd/vertexDescriptor.h"
//#include "../osd/debug.h"
#include "../osd/error.h"
#include "../osd/opengl.h"
#include <algorithm>
#include <cassert>
#include <iostream>
#include <sstream>
namespace OpenSubdiv {
namespace OPENSUBDIV_VERSION {
namespace Osd {
static const char *shaderSource =
#include "../osd/glslComputeKernel.gen.h"
;
// ----------------------------------------------------------------------------
class GLSLComputeController::KernelBundle :
NonCopyable<GLSLComputeController::KernelBundle> {
public:
KernelBundle() :
_program(0),
_uniformSizes(0),
_uniformOffsets(0),
_uniformIndices(0),
_uniformWeights(0),
_uniformStart(0),
_uniformEnd(0),
_uniformOffset(0),
_uniformNumCVs(0),
_workGroupSize(64) { }
~KernelBundle() {
if (_program) {
glDeleteProgram(_program);
}
}
void UseProgram(int primvarOffset) const {
glUseProgram(_program);
glUniform1i(_uniformOffset, primvarOffset);
//OSD_DEBUG_CHECK_GL_ERROR("UseProgram");
}
bool Compile(VertexBufferDescriptor const & desc) {
_desc = VertexBufferDescriptor(0, desc.length, desc.stride);
if (_program) {
glDeleteProgram(_program);
_program=0;
}
_program = glCreateProgram();
GLuint shader = glCreateShader(GL_COMPUTE_SHADER);
std::ostringstream defines;
defines << "#define OFFSET " << _desc.offset << "\n"
<< "#define LENGTH " << _desc.length << "\n"
<< "#define STRIDE " << _desc.stride << "\n"
<< "#define WORK_GROUP_SIZE " << _workGroupSize << "\n";
std::string defineStr = defines.str();
const char *shaderSources[3] = {"#version 430\n", 0, 0};
shaderSources[1] = defineStr.c_str();
shaderSources[2] = shaderSource;
glShaderSource(shader, 3, shaderSources, NULL);
glCompileShader(shader);
glAttachShader(_program, shader);
GLint linked = 0;
glLinkProgram(_program);
glGetProgramiv(_program, GL_LINK_STATUS, &linked);
if (linked == GL_FALSE) {
char buffer[1024];
glGetShaderInfoLog(shader, 1024, NULL, buffer);
Error(OSD_GLSL_LINK_ERROR, buffer);
glGetProgramInfoLog(_program, 1024, NULL, buffer);
Error(OSD_GLSL_LINK_ERROR, buffer);
glDeleteProgram(_program);
_program = 0;
return false;
}
glDeleteShader(shader);
_subStencilKernel = glGetSubroutineIndex(_program, GL_COMPUTE_SHADER, "computeStencil");
// set uniform locations for compute kernels
_uniformSizes = glGetUniformLocation(_program, "sterncilSizes");
_uniformOffsets = glGetUniformLocation(_program, "sterncilOffsets");
_uniformIndices = glGetUniformLocation(_program, "sterncilIndices");
_uniformWeights = glGetUniformLocation(_program, "sterncilIWeights");
_uniformStart = glGetUniformLocation(_program, "batchStart");
_uniformEnd = glGetUniformLocation(_program, "batchEnd");
_uniformOffset = glGetUniformLocation(_program, "primvarOffset");
_uniformNumCVs = glGetUniformLocation(_program, "numCVs");
//OSD_DEBUG_CHECK_GL_ERROR("Compile");
return true;
}
void ApplyStencilTableKernel(Far::KernelBatch const &batch, int offset, int numCVs) const {
// select stencil GLSL subroutine
glUniformSubroutinesuiv(GL_COMPUTE_SHADER, 1, &_subStencilKernel);
dispatchCompute(offset, numCVs, batch.start, batch.end);
}
struct Match {
Match(VertexBufferDescriptor const & d) : desc(d) { }
bool operator() (KernelBundle const * kernel) {
return (desc.length==kernel->_desc.length and
desc.stride==kernel->_desc.stride);
}
VertexBufferDescriptor desc;
};
protected:
void dispatchCompute(int offset, int numCVs, int start, int end) const {
int count = end - start;
if (count<=0) {
return;
}
glUniform1i(_uniformStart, start);
glUniform1i(_uniformEnd, end);
glUniform1i(_uniformOffset, offset);
glUniform1i(_uniformNumCVs, numCVs);
glDispatchCompute(count/_workGroupSize + 1, 1, 1);
// sync for later reading.
// XXX: in theory, just SHADER_STORAGE_BARRIER is needed here. However
// we found a problem (issue #295) with nvidia driver 331.49 / Quadro4000
// resulting in invalid vertices.
// Apparently adding TEXTURE_FETCH_BARRIER after a kernel fixes it.
// The workaroud is commented out, since it looks fixed as of driver 334.xx.
glMemoryBarrier(GL_SHADER_STORAGE_BARRIER_BIT);
//OSD_DEBUG_CHECK_GL_ERROR("dispatchCompute");
}
private:
GLuint _program;
GLuint _subStencilKernel; // stencil compute kernel GLSL subroutine
GLuint _uniformSizes, // uniform paramaeters for kernels
_uniformOffsets,
_uniformIndices,
_uniformWeights,
_uniformStart, // batch
_uniformEnd,
_uniformOffset, // GL primvar buffer descriptor
_uniformNumCVs; // number of const control vertices padded at
// the beginning of the buffer
VertexBufferDescriptor _desc; // primvar buffer descriptor
int _workGroupSize;
};
// ----------------------------------------------------------------------------
void
GLSLComputeController::ApplyStencilTableKernel(
Far::KernelBatch const &batch, ComputeContext const *context) const {
assert(context);
_currentBindState.kernelBundle->ApplyStencilTableKernel(
batch, _currentBindState.desc.offset, context->GetNumControlVertices());
}
// ----------------------------------------------------------------------------
GLSLComputeController::GLSLComputeController() { }
GLSLComputeController::~GLSLComputeController() {
for (KernelRegistry::iterator it = _kernelRegistry.begin();
it != _kernelRegistry.end(); ++it) {
delete *it;
}
}
// ----------------------------------------------------------------------------
void
GLSLComputeController::Synchronize() {
glFinish();
}
// ----------------------------------------------------------------------------
GLSLComputeController::KernelBundle const *
GLSLComputeController::getKernel(VertexBufferDescriptor const &desc) {
KernelRegistry::iterator it =
std::find_if(_kernelRegistry.begin(), _kernelRegistry.end(),
KernelBundle::Match(desc));
if (it != _kernelRegistry.end()) {
return *it;
} else {
KernelBundle * kernelBundle = new KernelBundle();
kernelBundle->Compile(desc);
_kernelRegistry.push_back(kernelBundle);
return kernelBundle;
}
}
void
GLSLComputeController::bindBufferAndProgram() {
if (_currentBindState.buffer)
glBindBufferBase(GL_SHADER_STORAGE_BUFFER, 0, _currentBindState.buffer);
_currentBindState.kernelBundle->UseProgram(/*primvarOffset*/0);
glMemoryBarrier(GL_TEXTURE_FETCH_BARRIER_BIT);
}
void
GLSLComputeController::unbindBufferAndProgram() {
glMemoryBarrier(GL_TEXTURE_FETCH_BARRIER_BIT);
glBindBufferBase(GL_SHADER_STORAGE_BUFFER, 0, 0);
glUseProgram(0);
}
// ----------------------------------------------------------------------------
} // end namespace Osd
} // end namespace OPENSUBDIV_VERSION
} // end namespace OpenSubdiv