mirror-github-bspeice/fractorium
1
0
Fork 0
mirror of https://bitbucket.org/mfeemster/fractorium.git synced 2025-07-01 13:56:06 -04:00
Code Issues Actions Packages Projects Releases Wiki Activity

More Linux work. This has Render, Animate and Genome building and running on Linux.

Browse Source
This commit is contained in:
mfeemster
2014-12-05 18:30:46 -08:00
parent 4777ab52bc
commit a15f6d6b32
44 changed files with 2614 additions and 1021 deletions
Download Patch File Download Diff File Expand all files Collapse all files

32
Source/EmberCL/DEOpenCLKernelCreator.cpp
View File

@ -1,7 +1,7 @@
#include "EmberCLPch.h"
#include "DEOpenCLKernelCreator.h"
namespace EmberCLns
namespace EmberCLns
{
/// <summary>
/// Empty constructor that does nothing. The user must call the one which takes a bool
@ -219,7 +219,7 @@ template <typename T>
string DEOpenCLKernelCreator<T>::CreateLogScaleAssignDEKernelString()
{
ostringstream os;
os <<
ConstantDefinesString(typeid(T) == typeid(double)) <<
DensityFilterCLStructString <<
@ -243,7 +243,7 @@ string DEOpenCLKernelCreator<T>::CreateLogScaleAssignDEKernelString()
" barrier(CLK_GLOBAL_MEM_FENCE);\n"//Just to be safe. Makes no speed difference to do all of the time or only when there's a hit.
" }\n"
"}\n";
return os.str();
}
@ -354,7 +354,7 @@ string DEOpenCLKernelCreator<T>::CreateGaussianDEKernel(size_t ss)
" if (bucket.w != 0)\n"
" cacheLog = (densityFilter->m_K1 * log(1.0 + bucket.w * densityFilter->m_K2)) / bucket.w;\n"
"\n";
if (doSS)
{
os <<
@ -626,7 +626,7 @@ string DEOpenCLKernelCreator<T>::CreateGaussianDEKernel(size_t ss)
" if (bucket.w != 0)\n"
" {\n"
" cacheLog = (densityFilter->m_K1 * log(1.0 + bucket.w * densityFilter->m_K2)) / bucket.w;\n";
if (doSS)
{
os <<
@ -644,7 +644,7 @@ string DEOpenCLKernelCreator<T>::CreateGaussianDEKernel(size_t ss)
" }\n"
" }\n"
"\n";
if (doScf)
os << " filterSelect *= scfact;\n";
}
@ -653,7 +653,7 @@ string DEOpenCLKernelCreator<T>::CreateGaussianDEKernel(size_t ss)
os
<< " filterSelect = bucket.w;\n";
}
os <<
"\n"
" if (filterSelect > densityFilter->m_MaxFilteredCounts)\n"
@ -764,7 +764,7 @@ string DEOpenCLKernelCreator<T>::CreateGaussianDEKernelNoLocalCache(size_t ss)
" ((((BLOCK_ID_Y * chunkSizeH) + chunkH) * BLOCK_SIZE_Y) + THREAD_ID_Y >= densityFilter->m_SuperRasH))\n"
" return;\n"
"\n";
if (doSS)
{
os <<
@ -777,7 +777,7 @@ string DEOpenCLKernelCreator<T>::CreateGaussianDEKernelNoLocalCache(size_t ss)
if (doScf)
os << " real_t scfact = pow((real_t)densityFilter->m_Supersample / ((real_t)densityFilter->m_Supersample + 1.0), 2.0);\n";
}
os <<
//Compute the bounds of the area to be sampled, which is just the ends minus the super sample minus 1.
" uint leftBound = densityFilter->m_Supersample - 1;\n"
@ -807,7 +807,7 @@ string DEOpenCLKernelCreator<T>::CreateGaussianDEKernelNoLocalCache(size_t ss)
" if (bucket.w != 0)\n"
" {\n"
" cacheLog = (densityFilter->m_K1 * log(1.0 + bucket.w * densityFilter->m_K2)) / bucket.w;\n";
if (doSS)
{
os <<
@ -825,7 +825,7 @@ string DEOpenCLKernelCreator<T>::CreateGaussianDEKernelNoLocalCache(size_t ss)
" }\n"
" }\n"
"\n";
if (doScf)
os << " filterSelect *= scfact;\n";
}
@ -834,7 +834,7 @@ string DEOpenCLKernelCreator<T>::CreateGaussianDEKernelNoLocalCache(size_t ss)
os
<< " filterSelect = bucket.w;\n";
}
os <<
"\n"
" if (filterSelect > densityFilter->m_MaxFilteredCounts)\n"
@ -876,4 +876,10 @@ string DEOpenCLKernelCreator<T>::CreateGaussianDEKernelNoLocalCache(size_t ss)
return os.str();
}
}
template EMBERCL_API class DEOpenCLKernelCreator<float>;
#ifdef DO_DOUBLE
template EMBERCL_API class DEOpenCLKernelCreator<double>;
#endif
}

14
Source/EmberCL/DEOpenCLKernelCreator.h
View File

@ -1,5 +1,5 @@
#pragma once
#include "EmberCLPch.h"
#include "EmberCLStructs.h"
#include "EmberCLFunctions.h"
@ -52,7 +52,7 @@ private:
string CreateLogScaleAssignDEKernelString();
string CreateGaussianDEKernel(size_t ss);
string CreateGaussianDEKernelNoLocalCache(size_t ss);
string m_LogScaleAssignDEKernel;
string m_LogScaleAssignDEEntryPoint;
@ -77,9 +77,9 @@ private:
bool m_NVidia;
};
template EMBERCL_API class DEOpenCLKernelCreator<float>;
#ifdef DO_DOUBLE
template EMBERCL_API class DEOpenCLKernelCreator<double>;
#endif
//template EMBERCL_API class DEOpenCLKernelCreator<float>;
//
//#ifdef DO_DOUBLE
// template EMBERCL_API class DEOpenCLKernelCreator<double>;
//#endif
}

20
Source/EmberCL/EmberCLPch.h
View File

@ -8,9 +8,14 @@
#define WIN32_LEAN_AND_MEAN//Exclude rarely-used stuff from Windows headers.
#define _USE_MATH_DEFINES
#include "Timing.h"
#include "Renderer.h"
#ifdef _WIN32
#include <windows.h>
#include <SDKDDKVer.h>
#else
#include "GL/glx.h"
#endif
#include <utility>
@ -25,16 +30,17 @@
#include <iterator>
#include <time.h>
#include "Timing.h"
#include "Renderer.h"
#if defined(BUILDING_EMBERCL)
#define EMBERCL_API __declspec(dllexport)
#ifdef _WIN32
#if defined(BUILDING_EMBERCL)
#define EMBERCL_API __declspec(dllexport)
#else
#define EMBERCL_API __declspec(dllimport)
#endif
#else
#define EMBERCL_API __declspec(dllimport)
#define EMBERCL_API
#endif
using namespace std;
using namespace EmberNs;
//#define TEST_CL 1
//#define TEST_CL_BUFFERS 1
//#define TEST_CL_BUFFERS 1

10
Source/EmberCL/EmberCLStructs.h
View File

@ -13,7 +13,13 @@
namespace EmberCLns
{
#define ALIGN __declspec(align(16))//These two must always match.
//These two must always match.
#ifdef WIN32
#define ALIGN __declspec(align(16))
#else
#define ALIGN __attribute__ ((aligned (16)))
#endif
#define ALIGN_CL "((aligned (16)))"//The extra parens are necessary.
/// <summary>
@ -378,4 +384,4 @@ static const char* UnionCLStructString =
" real_t m_Reals[4];\n"
"} real4reals;\n"
"\n";
}
}

14
Source/EmberCL/FinalAccumOpenCLKernelCreator.cpp
View File

@ -275,7 +275,7 @@ string FinalAccumOpenCLKernelCreator<T>::CreateFinalAccumKernelString(bool early
" finalColor.m_Float4.x = (float)newBucket.m_Real4.x;\n"//CPU side clamps, skip here because write_imagef() does the clamping for us.
" finalColor.m_Float4.y = (float)newBucket.m_Real4.y;\n"
" finalColor.m_Float4.z = (float)newBucket.m_Real4.z;\n";
if (alphaAccum)
{
if (alphaCalc)
@ -374,13 +374,13 @@ string FinalAccumOpenCLKernelCreator<T>::CreateGammaCorrectionFunctionString(boo
<< " else\n"
<< " a = 0;\n";
}
os <<
"\n"
" correctedChannels[rgbi] = (" << dataType << ")clamp(a, 0.0, 255.0);\n"
" }\n"
"\n";
//The CPU code has 3 cases for assigning alpha:
//[3] = alpha.//Early clip.
//[3] = alpha * 255.//Final Rgba with transparency.
@ -513,4 +513,10 @@ string FinalAccumOpenCLKernelCreator<T>::CreateGammaCorrectionKernelString(bool
return os.str();
}
}
template EMBERCL_API class FinalAccumOpenCLKernelCreator<float>;
#ifdef DO_DOUBLE
template EMBERCL_API class FinalAccumOpenCLKernelCreator<double>;
#endif
}

16
Source/EmberCL/FinalAccumOpenCLKernelCreator.h
View File

@ -1,5 +1,5 @@
#pragma once
#include "EmberCLPch.h"
#include "EmberCLStructs.h"
#include "EmberCLFunctions.h"
@ -26,7 +26,7 @@ class EMBERCL_API FinalAccumOpenCLKernelCreator
{
public:
FinalAccumOpenCLKernelCreator();
string GammaCorrectionWithAlphaCalcKernel();
string GammaCorrectionWithAlphaCalcEntryPoint();
@ -79,9 +79,9 @@ private:
string m_FinalAccumLateClipWithoutAlphaCalcWithAlphaAccumEntryPoint;
};
template EMBERCL_API class FinalAccumOpenCLKernelCreator<float>;
#ifdef DO_DOUBLE
template EMBERCL_API class FinalAccumOpenCLKernelCreator<double>;
#endif
}
//template EMBERCL_API class FinalAccumOpenCLKernelCreator<float>;
//
//#ifdef DO_DOUBLE
// template EMBERCL_API class FinalAccumOpenCLKernelCreator<double>;
//#endif
}

32
Source/EmberCL/IterOpenCLKernelCreator.cpp
View File

@ -60,7 +60,7 @@ string IterOpenCLKernelCreator<T>::CreateIterKernelString(Ember<T>& ember, strin
for (i = 0; i < totalXformCount; i++)
{
Xform<T>* xform = ember.GetTotalXform(i);
size_t totalVarCount = xform->TotalVariationCount();
//size_t totalVarCount = xform->TotalVariationCount();
bool needPrecalcSumSquares = false;
bool needPrecalcSqrtSumSquares = false;
bool needPrecalcAngles = false;
@ -73,7 +73,7 @@ string IterOpenCLKernelCreator<T>::CreateIterKernelString(Ember<T>& ember, strin
"{\n"
" real_t transX, transY, transZ;\n"
" real4 vIn, vOut = 0.0;\n";
//Determine if any variations, regular, pre, or post need precalcs.
while (Variation<T>* var = xform->GetVariation(v++))
{
@ -204,7 +204,7 @@ string IterOpenCLKernelCreator<T>::CreateIterKernelString(Ember<T>& ember, strin
}
}
}
if (xform->HasPost())
{
xformFuncs <<
@ -267,7 +267,7 @@ string IterOpenCLKernelCreator<T>::CreateIterKernelString(Ember<T>& ember, strin
" uint threadXDivRows = (THREAD_ID_X / (NTHREADS / THREADS_PER_WARP));\n"
" uint threadsMinus1 = NTHREADS - 1;\n"
;
os <<
"\n"
#ifndef STRAIGHT_RAND
@ -294,7 +294,7 @@ string IterOpenCLKernelCreator<T>::CreateIterKernelString(Ember<T>& ember, strin
" firstPoint = points[pointsIndex];\n"
" }\n"
"\n";
//This is done once initially here and then again after each swap-sync in the main loop.
//This along with the randomness that the point shuffle provides gives sufficient randomness
//to produce results identical to those produced on the CPU.
@ -346,7 +346,7 @@ string IterOpenCLKernelCreator<T>::CreateIterKernelString(Ember<T>& ember, strin
" switch (secondPoint.m_LastXfUsed)\n"
" {\n";
}
os <<
" case " << i << ":\n"
" {\n" <<
@ -430,7 +430,7 @@ string IterOpenCLKernelCreator<T>::CreateIterKernelString(Ember<T>& ember, strin
" continue;\n"
" }\n"
"\n";
if (ember.UseFinalXform())
{
size_t finalIndex = ember.TotalXformCount() - 1;
@ -446,7 +446,7 @@ string IterOpenCLKernelCreator<T>::CreateIterKernelString(Ember<T>& ember, strin
" }\n"
"\n";
}
os << CreateProjectionString(ember);
if (doAccum)
@ -549,7 +549,7 @@ string IterOpenCLKernelCreator<T>::CreateIterKernelString(Ember<T>& ember, strin
"\n"
" barrier(CLK_GLOBAL_MEM_FENCE);\n";//Barrier every time, whether or not the point was in bounds, else artifacts will occur when doing strips.
}
os <<
" }\n"//Main for loop.
"\n"
@ -578,7 +578,7 @@ string IterOpenCLKernelCreator<T>::CreateIterKernelString(Ember<T>& ember, strin
/// the length of the parametric values is unknown.
/// This is solved by passing a separate buffer of values dedicated specifically
/// to parametric variations.
/// In OpenCL, a series of #define constants are declared which specify the indices in
/// In OpenCL, a series of #define constants are declared which specify the indices in
/// the buffer where the various values are stored.
/// The possibility of a parametric variation type being present in multiple xforms is taken
/// into account by appending the xform index to the #define, thus making each unique.
@ -622,7 +622,7 @@ void IterOpenCLKernelCreator<T>::ParVarIndexDefines(Ember<T>& ember, pair<string
for (i = 0; i < xformCount; i++)
{
if (xform = ember.GetTotalXform(i))
if ((xform = ember.GetTotalXform(i)))
{
size_t varCount = xform->TotalVariationCount();
@ -672,7 +672,7 @@ template <typename T>
bool IterOpenCLKernelCreator<T>::IsBuildRequired(Ember<T>& ember1, Ember<T>& ember2)
{
size_t i, j, xformCount = ember1.TotalXformCount();
if (xformCount != ember2.TotalXformCount())
return true;
@ -834,4 +834,10 @@ string IterOpenCLKernelCreator<T>::CreateProjectionString(Ember<T>& ember)
return os.str();
}
}
template EMBERCL_API class IterOpenCLKernelCreator<float>;
#ifdef DO_DOUBLE
template EMBERCL_API class IterOpenCLKernelCreator<double>;
#endif
}

12
Source/EmberCL/IterOpenCLKernelCreator.h
View File

@ -41,12 +41,12 @@ private:
string m_ZeroizeEntryPoint;
bool m_NVidia;
};
template EMBERCL_API class IterOpenCLKernelCreator<float>;
#ifdef DO_DOUBLE
template EMBERCL_API class IterOpenCLKernelCreator<double>;
#endif
//
//template EMBERCL_API class IterOpenCLKernelCreator<float>;
//
//#ifdef DO_DOUBLE
// template EMBERCL_API class IterOpenCLKernelCreator<double>;
//#endif
//
//template EMBERCL_API string IterOpenCLKernelCreator::CreateIterKernelString<float>(Ember<float>& ember, string& parVarDefines, bool lockAccum, bool doAccum);

94
Source/EmberCL/OpenCLWrapper.cpp
View File

@ -48,13 +48,13 @@ bool OpenCLWrapper::Init(unsigned int platform, unsigned int device, bool shared
m_Init = false;
m_ErrorReport.clear();
if (m_Platforms.size() > 0)
{
if (platform < m_Platforms.size() && platform < m_Devices.size())
{
m_PlatformIndex = platform;//Platform is ok, now do context.
if (CreateContext(shared))
{
//Context is ok, now do device.
@ -90,10 +90,10 @@ bool OpenCLWrapper::Init(unsigned int platform, unsigned int device, bool shared
/// <param name="program">The program source</param>
/// <param name="entryPoint">The name of the entry point kernel function in the program</param>
/// <returns>True if success, else false.</returns>
bool OpenCLWrapper::AddProgram(std::string name, std::string& program, std::string& entryPoint, bool doublePrecision)
bool OpenCLWrapper::AddProgram(const string& name, const string& program, const string& entryPoint, bool doublePrecision)
{
Spk spk;
if (CreateSPK(name, program, entryPoint, spk, doublePrecision))
{
for (size_t i = 0; i < m_Programs.size(); i++)
@ -109,7 +109,7 @@ bool OpenCLWrapper::AddProgram(std::string name, std::string& program, std::stri
m_Programs.push_back(spk);
return true;
}
return false;
}
@ -135,14 +135,14 @@ void OpenCLWrapper::ClearPrograms()
bool OpenCLWrapper::AddBuffer(const string& name, size_t size, cl_mem_flags flags)
{
cl_int err;
if (m_Init)
{
int bufferIndex = FindBufferIndex(name);
if (bufferIndex == -1)//If the buffer didn't exist, create and add.
{
cl::Buffer buff(m_Context, flags, size, NULL, &err);
cl::Buffer buff(m_Context, flags, size, nullptr, &err);
if (!CheckCL(err, "cl::Buffer()"))
return false;
@ -153,9 +153,9 @@ bool OpenCLWrapper::AddBuffer(const string& name, size_t size, cl_mem_flags flag
}
else if (GetBufferSize(bufferIndex) != size)//If it did exist, only create and add if the sizes were different.
{
m_Buffers[bufferIndex] = NamedBuffer(cl::Buffer(m_Context, flags, 0, NULL, &err), "emptybuffer");//First clear out the original so the two don't exist in memory at once.
m_Buffers[bufferIndex] = NamedBuffer(cl::Buffer(m_Context, flags, 0, nullptr, &err), "emptybuffer");//First clear out the original so the two don't exist in memory at once.
cl::Buffer buff(m_Context, flags, size, NULL, &err);//Create the new buffer.
cl::Buffer buff(m_Context, flags, size, nullptr, &err);//Create the new buffer.
if (!CheckCL(err, "cl::Buffer()"))
return false;
@ -205,7 +205,7 @@ bool OpenCLWrapper::WriteBuffer(const string& name, void* data, size_t size)
{
int bufferIndex = FindBufferIndex(name);
return bufferIndex != -1 ? WriteBuffer(bufferIndex, data, size) : false;
return bufferIndex != -1 ? WriteBuffer(bufferIndex, data, size) : false;
}
/// <summary>
@ -220,7 +220,7 @@ bool OpenCLWrapper::WriteBuffer(unsigned int bufferIndex, void* data, size_t siz
if (m_Init && (bufferIndex < m_Buffers.size()) && (GetBufferSize(bufferIndex) == size))
{
cl::Event e;
cl_int err = m_Queue.enqueueWriteBuffer(m_Buffers[bufferIndex].m_Buffer, CL_TRUE, 0, size, data, NULL, &e);
cl_int err = m_Queue.enqueueWriteBuffer(m_Buffers[bufferIndex].m_Buffer, CL_TRUE, 0, size, data, nullptr, &e);
e.wait();
m_Queue.finish();
@ -243,7 +243,7 @@ bool OpenCLWrapper::ReadBuffer(const string& name, void* data, size_t size)
{
int bufferIndex = FindBufferIndex(name);
return bufferIndex != -1 ? ReadBuffer(bufferIndex, data, size) : false;
return bufferIndex != -1 ? ReadBuffer(bufferIndex, data, size) : false;
}
/// <summary>
@ -258,7 +258,7 @@ bool OpenCLWrapper::ReadBuffer(unsigned int bufferIndex, void* data, size_t size
if (m_Init && (bufferIndex < m_Buffers.size()) && (GetBufferSize(bufferIndex) == size))
{
cl::Event e;
cl_int err = m_Queue.enqueueReadBuffer(m_Buffers[bufferIndex].m_Buffer, CL_TRUE, 0, size, data, NULL, &e);
cl_int err = m_Queue.enqueueReadBuffer(m_Buffers[bufferIndex].m_Buffer, CL_TRUE, 0, size, data, nullptr, &e);
e.wait();
m_Queue.finish();
@ -291,9 +291,9 @@ int OpenCLWrapper::FindBufferIndex(const string& name)
/// <returns>The size of the buffer if found, else 0.</returns>
unsigned int OpenCLWrapper::GetBufferSize(const string& name)
{
unsigned int bufferIndex = FindBufferIndex(name);
int bufferIndex = FindBufferIndex(name);
return bufferIndex != -1 ? GetBufferSize(bufferIndex) : 0;
return bufferIndex != -1 ? GetBufferSize(bufferIndex) : 0;
}
/// <summary>
@ -304,7 +304,7 @@ unsigned int OpenCLWrapper::GetBufferSize(const string& name)
unsigned int OpenCLWrapper::GetBufferSize(unsigned int bufferIndex)
{
if (m_Init && bufferIndex < m_Buffers.size())
return (unsigned int)m_Buffers[bufferIndex].m_Buffer.getInfo<CL_MEM_SIZE>();
return (unsigned int)m_Buffers[bufferIndex].m_Buffer.getInfo<CL_MEM_SIZE>(nullptr);
return 0;
}
@ -437,7 +437,7 @@ bool OpenCLWrapper::WriteImage2D(unsigned int index, bool shared, ::size_t width
cl_int err;
cl::Event e;
cl::size_t<3> origin, region;
origin[0] = 0;
origin[1] = 0;
origin[2] = 0;
@ -452,7 +452,7 @@ bool OpenCLWrapper::WriteImage2D(unsigned int index, bool shared, ::size_t width
if (EnqueueAcquireGLObjects(imageGL))
{
err = m_Queue.enqueueWriteImage(imageGL, CL_TRUE, origin, region, row_pitch, 0, data, NULL, &e);
err = m_Queue.enqueueWriteImage(imageGL, CL_TRUE, origin, region, row_pitch, 0, data, nullptr, &e);
e.wait();
m_Queue.finish();
@ -462,7 +462,7 @@ bool OpenCLWrapper::WriteImage2D(unsigned int index, bool shared, ::size_t width
}
else if (!shared && index < m_Images.size())
{
err = m_Queue.enqueueWriteImage(m_Images[index].m_Image, CL_TRUE, origin, region, row_pitch, 0, data, NULL, &e);
err = m_Queue.enqueueWriteImage(m_Images[index].m_Image, CL_TRUE, origin, region, row_pitch, 0, data, nullptr, &e);
e.wait();
m_Queue.finish();
return CheckCL(err, "cl::enqueueWriteImage()");
@ -512,7 +512,7 @@ bool OpenCLWrapper::ReadImage(unsigned int imageIndex, ::size_t width, ::size_t
cl_int err;
cl::Event e;
cl::size_t<3> origin, region;
origin[0] = 0;
origin[1] = 0;
origin[2] = 0;
@ -595,16 +595,16 @@ unsigned int OpenCLWrapper::GetImageSize(unsigned int imageIndex, bool shared)
vector<cl::Memory> images;
images.push_back(m_GLImages[imageIndex].m_Image);
IMAGEGL2D image = m_GLImages[imageIndex].m_Image;
if (EnqueueAcquireGLObjects(&images))
size = image.getImageInfo<CL_IMAGE_WIDTH>() * image.getImageInfo<CL_IMAGE_HEIGHT>() * image.getImageInfo<CL_IMAGE_ELEMENT_SIZE>();//Should pitch be checked here?
size = image.getImageInfo<CL_IMAGE_WIDTH>(nullptr) * image.getImageInfo<CL_IMAGE_HEIGHT>(nullptr) * image.getImageInfo<CL_IMAGE_ELEMENT_SIZE>(nullptr);//Should pitch be checked here?
EnqueueReleaseGLObjects(&images);
}
else if (!shared && imageIndex < m_Images.size())
{
cl::Image2D image = m_Images[imageIndex].m_Image;
size = image.getImageInfo<CL_IMAGE_WIDTH>() * image.getImageInfo<CL_IMAGE_HEIGHT>() * image.getImageInfo<CL_IMAGE_ELEMENT_SIZE>();//Should pitch be checked here?
size = image.getImageInfo<CL_IMAGE_WIDTH>(nullptr) * image.getImageInfo<CL_IMAGE_HEIGHT>(nullptr) * image.getImageInfo<CL_IMAGE_ELEMENT_SIZE>(nullptr);//Should pitch be checked here?
}
}
@ -623,13 +623,13 @@ unsigned int OpenCLWrapper::GetImageSize(unsigned int imageIndex, bool shared)
/// <returns>True if all parameters matched, else false.</returns>
bool OpenCLWrapper::CompareImageParams(cl::Image& image, cl_mem_flags flags, const cl::ImageFormat& format, ::size_t width, ::size_t height, ::size_t row_pitch)
{
cl_image_format tempFormat = image.getImageInfo<CL_IMAGE_FORMAT>();
cl_image_format tempFormat = image.getImageInfo<CL_IMAGE_FORMAT>(nullptr);
return (/*image.getImageInfo<CL_MEM_FLAGS>() == flags &&*/
tempFormat.image_channel_data_type == format.image_channel_data_type &&
tempFormat.image_channel_order == format.image_channel_order &&
image.getImageInfo<CL_IMAGE_WIDTH>() == width &&
image.getImageInfo<CL_IMAGE_HEIGHT>() == height/* &&
tempFormat.image_channel_data_type == format.image_channel_data_type &&
tempFormat.image_channel_order == format.image_channel_order &&
image.getImageInfo<CL_IMAGE_WIDTH>(nullptr) == width &&
image.getImageInfo<CL_IMAGE_HEIGHT>(nullptr) == height/* &&
image.getImageInfo<CL_IMAGE_ROW_PITCH>() == row_pitch*/);//Pitch will be (width * bytes per pixel) + padding.
}
@ -822,7 +822,7 @@ bool OpenCLWrapper::EnqueueReleaseGLObjects(const VECTOR_CLASS<cl::Memory>* memO
bool OpenCLWrapper::CreateSampler(cl::Sampler& sampler, cl_bool normalizedCoords, cl_addressing_mode addressingMode, cl_filter_mode filterMode)
{
cl_int err;
sampler = cl::Sampler(m_Context,
normalizedCoords,
addressingMode,
@ -840,7 +840,7 @@ bool OpenCLWrapper::CreateSampler(cl::Sampler& sampler, cl_bool normalizedCoords
/// <param name="argIndex">Index of the argument</param>
/// <param name="name">The name of the buffer</param>
/// <returns>True if success, else false.</returns>
bool OpenCLWrapper::SetBufferArg(unsigned int kernelIndex, unsigned int argIndex, string name)
bool OpenCLWrapper::SetBufferArg(unsigned int kernelIndex, unsigned int argIndex, const string& name)
{
int bufferIndex = OpenCLWrapper::FindBufferIndex(name);
@ -872,7 +872,7 @@ bool OpenCLWrapper::SetBufferArg(unsigned int kernelIndex, unsigned int argIndex
/// <param name="shared">True if shared with an OpenGL texture, else false</param>
/// <param name="name">The name of the 2D image</param>
/// <returns>True if success, else false.</returns>
bool OpenCLWrapper::SetImageArg(unsigned int kernelIndex, unsigned int argIndex, bool shared, string name)
bool OpenCLWrapper::SetImageArg(unsigned int kernelIndex, unsigned int argIndex, bool shared, const string& name)
{
if (m_Init)
{
@ -948,7 +948,7 @@ bool OpenCLWrapper::RunKernel(unsigned int kernelIndex, unsigned int totalGridWi
cl::NullRange,
cl::NDRange(totalGridWidth, totalGridHeight, totalGridDepth),
cl::NDRange(blockWidth, blockHeight, blockDepth),
NULL,
nullptr,
&e);
e.wait();
@ -984,7 +984,7 @@ T OpenCLWrapper::GetInfo(size_t platform, size_t device, cl_device_info name)
string OpenCLWrapper::PlatformName(size_t platform)
{
if (platform < m_Platforms.size())
return m_Platforms[platform].getInfo<CL_PLATFORM_VENDOR>() + " " + m_Platforms[platform].getInfo<CL_PLATFORM_NAME>() + " " + m_Platforms[platform].getInfo<CL_PLATFORM_VERSION>();
return m_Platforms[platform].getInfo<CL_PLATFORM_VENDOR>(nullptr) + " " + m_Platforms[platform].getInfo<CL_PLATFORM_NAME>(nullptr) + " " + m_Platforms[platform].getInfo<CL_PLATFORM_VERSION>(nullptr);
else
return "";
}
@ -1018,7 +1018,7 @@ string OpenCLWrapper::DeviceName(size_t platform, size_t device)
if (platform < m_Platforms.size() && platform < m_Devices.size())
if (device < m_Devices[platform].size())
s = m_Devices[platform][device].getInfo<CL_DEVICE_VENDOR>() + " " + m_Devices[platform][device].getInfo<CL_DEVICE_NAME>();// + " " + m_Devices[platform][device].getInfo<CL_DEVICE_VERSION>();
s = m_Devices[platform][device].getInfo<CL_DEVICE_VENDOR>(nullptr) + " " + m_Devices[platform][device].getInfo<CL_DEVICE_NAME>(nullptr);// + " " + m_Devices[platform][device].getInfo<CL_DEVICE_VERSION>();
return s;
}
@ -1043,7 +1043,7 @@ vector<string> OpenCLWrapper::DeviceNames(size_t platform)
i++;
} while (s != "");
return devices;
}
@ -1094,13 +1094,13 @@ string OpenCLWrapper::DumpInfo()
os << "CL_DEVICE_MAX_READ_IMAGE_ARGS: " << GetInfo<cl_uint> (platform, device, CL_DEVICE_MAX_READ_IMAGE_ARGS) << endl;
os << "CL_DEVICE_MAX_WRITE_IMAGE_ARGS: " << GetInfo<cl_uint> (platform, device, CL_DEVICE_MAX_WRITE_IMAGE_ARGS) << endl;
os << "CL_DEVICE_MAX_MEM_ALLOC_SIZE: " << GetInfo<cl_ulong>(platform, device, CL_DEVICE_MAX_MEM_ALLOC_SIZE) << endl;
os << "CL_DEVICE_GLOBAL_MEM_CACHE_TYPE: " << GetInfo<cl_uint> (platform, device, CL_DEVICE_GLOBAL_MEM_CACHE_TYPE) << endl;
os << "CL_DEVICE_GLOBAL_MEM_CACHELINE_SIZE: " << GetInfo<cl_uint> (platform, device, CL_DEVICE_GLOBAL_MEM_CACHELINE_SIZE) << endl;
os << "CL_DEVICE_GLOBAL_MEM_CACHE_SIZE: " << GetInfo<cl_ulong>(platform, device, CL_DEVICE_GLOBAL_MEM_CACHE_SIZE) << endl;
os << "CL_DEVICE_GLOBAL_MEM_SIZE: " << GetInfo<cl_ulong>(platform, device, CL_DEVICE_GLOBAL_MEM_SIZE) << endl;
os << "CL_DEVICE_MAX_CONSTANT_BUFFER_SIZE: " << GetInfo<cl_ulong>(platform, device, CL_DEVICE_MAX_CONSTANT_BUFFER_SIZE) << endl;
os << "CL_DEVICE_MAX_CONSTANT_ARGS: " << GetInfo<cl_uint> (platform, device, CL_DEVICE_MAX_CONSTANT_ARGS) << endl;
os << "CL_DEVICE_MAX_WORK_ITEM_DIMENSIONS: " << GetInfo<cl_uint> (platform, device, CL_DEVICE_MAX_WORK_ITEM_DIMENSIONS) << endl;
os << "CL_DEVICE_MAX_WORK_GROUP_SIZE: " << GetInfo<::size_t>(platform, device, CL_DEVICE_MAX_WORK_GROUP_SIZE) << endl;
@ -1165,7 +1165,7 @@ bool OpenCLWrapper::CreateContext(bool shared)
0
};
m_Context = cl::Context(CL_DEVICE_TYPE_GPU, props, NULL, NULL, &err);//May need to tinker with this on Mac.
m_Context = cl::Context(CL_DEVICE_TYPE_GPU, props, nullptr, nullptr, &err);//May need to tinker with this on Mac.
#else
#if defined WIN32
cl_context_properties props[] =
@ -1175,18 +1175,18 @@ bool OpenCLWrapper::CreateContext(bool shared)
CL_CONTEXT_PLATFORM, (cl_context_properties)(m_Platforms[m_PlatformIndex])(),
0
};
m_Context = cl::Context(CL_DEVICE_TYPE_GPU, props, NULL, NULL, &err);
m_Context = cl::Context(CL_DEVICE_TYPE_GPU, props, nullptr, nullptr, &err);
#else
cl_context_properties props[] =
{
CL_GL_CONTEXT_KHR, (cl_context_properties)glXGetCurrentContext(),
CL_GLX_DISPLAY_KHR, (cl_context_properties)glXGetCurrentDisplay(),
CL_CONTEXT_PLATFORM, (cl_context_properties)(m_Platforms[m_Platform])(),
CL_CONTEXT_PLATFORM, (cl_context_properties)(m_Platforms[m_PlatformIndex])(),
0
};
m_Context = cl::Context(CL_DEVICE_TYPE_GPU, props, NULL, NULL, &err);
m_Context = cl::Context(CL_DEVICE_TYPE_GPU, props, nullptr, nullptr, &err);
#endif
#endif
}
@ -1199,7 +1199,7 @@ bool OpenCLWrapper::CreateContext(bool shared)
0
};
m_Context = cl::Context(CL_DEVICE_TYPE_ALL, props, NULL, NULL, &err);
m_Context = cl::Context(CL_DEVICE_TYPE_ALL, props, nullptr, nullptr, &err);
}
return CheckCL(err, "cl::Context()");
@ -1213,7 +1213,7 @@ bool OpenCLWrapper::CreateContext(bool shared)
/// <param name="entryPoint">The name of the entry point kernel function in the program</param>
/// <param name="spk">The Spk object to store the resulting compiled program in</param>
/// <returns>True if success, else false.</returns>
bool OpenCLWrapper::CreateSPK(std::string& name, std::string& program, std::string& entryPoint, Spk& spk, bool doublePrecision)
bool OpenCLWrapper::CreateSPK(const string& name, const string& program, const string& entryPoint, Spk& spk, bool doublePrecision)
{
if (m_Init)
{
@ -1231,12 +1231,12 @@ bool OpenCLWrapper::CreateSPK(std::string& name, std::string& program, std::stri
//err = spk.m_Program.build(m_DeviceVec, "-cl-mad-enable -cl-single-precision-constant");
//err = spk.m_Program.build(m_DeviceVec, "-cl-mad-enable -cl-no-signed-zeros -cl-fast-relaxed-math -cl-single-precision-constant");//This can cause some rounding.
//err = spk.m_Program.build(m_DeviceVec, "-cl-mad-enable -cl-single-precision-constant");
if (CheckCL(err, "cl::Program::build()"))
{
//Building of program is ok, now create kernel with the specified entry point.
spk.m_Kernel = cl::Kernel(spk.m_Program, entryPoint.c_str(), &err);
if (CheckCL(err, "cl::Kernel()"))
return true;//Everything is ok.
}
@ -1330,4 +1330,4 @@ std::string OpenCLWrapper::ErrorToStringCL(cl_int err)
}
}
}
}
}

20
Source/EmberCL/OpenCLWrapper.h
View File

@ -37,7 +37,7 @@ public:
{
}
NamedBuffer(cl::Buffer& buff, string name)
NamedBuffer(const cl::Buffer& buff, const string& name)
{
m_Buffer = buff;
m_Name = name;
@ -57,7 +57,7 @@ public:
{
}
NamedImage2D(cl::Image2D& image, string name)
NamedImage2D(const cl::Image2D& image, const string& name)
{
m_Image = image;
m_Name = name;
@ -78,7 +78,7 @@ public:
{
}
NamedImage2DGL(IMAGEGL2D& image, string name)
NamedImage2DGL(const IMAGEGL2D& image, const string& name)
{
m_Image = image;
m_Name = name;
@ -107,7 +107,7 @@ public:
bool Init(unsigned int platform, unsigned int device, bool shared = false);
//Programs.
bool AddProgram(std::string name, std::string& program, std::string& entryPoint, bool doublePrecision);
bool AddProgram(const string& name, const string& program, const string& entryPoint, bool doublePrecision);
void ClearPrograms();
//Buffers.
@ -141,13 +141,13 @@ public:
bool EnqueueAcquireGLObjects(const VECTOR_CLASS<cl::Memory>* memObjects = NULL);
bool EnqueueReleaseGLObjects(const VECTOR_CLASS<cl::Memory>* memObjects = NULL);
bool CreateSampler(cl::Sampler& sampler, cl_bool normalizedCoords, cl_addressing_mode addressingMode, cl_filter_mode filterMode);
//Arguments.
bool SetBufferArg(unsigned int kernelIndex, unsigned int argIndex, string name);
bool SetBufferArg(unsigned int kernelIndex, unsigned int argIndex, const string& name);
bool SetBufferArg(unsigned int kernelIndex, unsigned int argIndex, unsigned int bufferIndex);
bool SetImageArg(unsigned int kernelIndex, unsigned int argIndex, bool shared, string name);
bool SetImageArg(unsigned int kernelIndex, unsigned int argIndex, bool shared, const string& name);
bool SetImageArg(unsigned int kernelIndex, unsigned int argIndex, bool shared, unsigned int imageIndex);
/// <summary>
/// Set an argument in the specified kernel, at the specified argument index.
/// Must keep this here in the .h because it's templated.
@ -195,7 +195,7 @@ public:
private:
bool CreateContext(bool shared);
bool CreateSPK(std::string& name, std::string& program, std::string& entryPoint, Spk& spk, bool doublePrecision);
bool CreateSPK(const string& name, const string& program, const string& entryPoint, Spk& spk, bool doublePrecision);
bool CheckCL(cl_int err, const char* name);
std::string ErrorToStringCL(cl_int err);
@ -216,4 +216,4 @@ private:
std::vector<NamedImage2D> m_Images;
std::vector<NamedImage2DGL> m_GLImages;
};
}
}

73
Source/EmberCL/RendererCL.cpp
View File

@ -52,7 +52,7 @@ RendererCL<T>::RendererCL(unsigned int platform, unsigned int device, bool share
m_PaletteFormat.image_channel_data_type = CL_FLOAT;
m_FinalFormat.image_channel_order = CL_RGBA;
m_FinalFormat.image_channel_data_type = CL_UNORM_INT8;//Change if this ever supports 2BPC outputs for PNG.
FillSeeds();
Init(platform, device, shared, outputTexID);//Init OpenCL upon construction and create programs that will not change.
}
@ -73,7 +73,7 @@ RendererCL<T>::~RendererCL()
/// Initialize OpenCL.
/// In addition to initializing, this function will create the zeroization program,
/// as well as the basic log scale filtering programs. This is done to ensure basic
/// compilation works. Further compilation will be done later for iteration, density filtering,
/// compilation works. Further compilation will be done later for iteration, density filtering,
/// and final accumulation.
/// </summary>
/// <param name="platform">The index platform of the platform to use</param>
@ -146,7 +146,7 @@ bool RendererCL<T>::SetOutputTexture(GLuint outputTexID)
m_ErrorReport.push_back(loc);
success = false;
}
LeaveResize();
return success;
}
@ -312,7 +312,7 @@ bool RendererCL<T>::ClearFinal()
vector<unsigned char> v;
unsigned int index = m_Wrapper.FindImageIndex(m_FinalImageName, m_Wrapper.Shared());
if (PrepFinalAccumVector(v))
if (this->PrepFinalAccumVector(v))
{
bool b = m_Wrapper.WriteImage2D(index, m_Wrapper.Shared(), FinalRasW(), FinalRasH(), 0, v.data());
@ -472,7 +472,7 @@ eRendererType RendererCL<T>::RendererType() const
/// </summary>
/// <returns>The concatenated error report string</returns>
template <typename T>
string RendererCL<T>::ErrorReportString()
string RendererCL<T>::ErrorReportString()
{
return EmberReport::ErrorReportString() + m_Wrapper.ErrorReportString();
}
@ -527,9 +527,11 @@ bool RendererCL<T>::RandVec(vector<QTIsaac<ISAAC_SIZE, ISAAC_INT>>& randVec)
template <typename T>
void RendererCL<T>::MakeDmap(T colorScalar)
{
m_Ember.m_Palette.MakeDmap<float>(m_Dmap, colorScalar);
//m_Ember.m_Palette.MakeDmap<float>(m_DmapCL, colorScalar);
m_Ember.m_Palette.MakeDmap(m_DmapCL, colorScalar);
}
/// <summary>
/// Allocate all buffers required for running as well as the final
/// 2D image.
@ -561,9 +563,10 @@ bool RendererCL<T>::Alloc()
if (b && !(b = m_Wrapper.AddBuffer(m_AccumBufferName, accumLength))) { m_ErrorReport.push_back(loc); }//Accum buffer.
if (b && !(b = m_Wrapper.AddBuffer(m_PointsBufferName, IterGridKernelCount() * sizeof(PointCL<T>)))) { m_ErrorReport.push_back(loc); }//Points between iter calls.
LeaveResize();
if (b && !(b = SetOutputTexture(m_OutputTexID))) { m_ErrorReport.push_back(loc); }
LeaveResize();
return b;
}
@ -706,7 +709,7 @@ EmberStats RendererCL<T>::Iterate(size_t iterCount, size_t temporalSample)
else
{
m_Abort = true;
m_ErrorReport.push_back(loc);
m_ErrorReport.push_back(loc);
}
return stats;
@ -787,9 +790,9 @@ bool RendererCL<T>::RunIter(size_t iterCount, size_t temporalSample, size_t& ite
if (b && !(b = m_Wrapper.WriteBuffer (m_XformsBufferName, (void*)m_XformsCL.data(), sizeof(m_XformsCL[0]) * m_XformsCL.size()))) { m_ErrorReport.push_back(loc); }
if (b && !(b = m_Wrapper.AddAndWriteBuffer(m_DistBufferName, (void*)XformDistributions(), XformDistributionsSize()))) { m_ErrorReport.push_back(loc); }//Will be resized for xaos.
if (b && !(b = m_Wrapper.WriteBuffer (m_CarToRasBufferName, (void*)&m_CarToRasCL, sizeof(m_CarToRasCL)))) { m_ErrorReport.push_back(loc); }
if (b && !(b = m_Wrapper.AddAndWriteImage("Palette", CL_MEM_READ_ONLY, m_PaletteFormat, m_Dmap.m_Entries.size(), 1, 0, m_Dmap.m_Entries.data()))) { m_ErrorReport.push_back(loc); }
if (b && !(b = m_Wrapper.AddAndWriteImage("Palette", CL_MEM_READ_ONLY, m_PaletteFormat, m_DmapCL.m_Entries.size(), 1, 0, m_DmapCL.m_Entries.data()))) { m_ErrorReport.push_back(loc); }
//If animating, treat each temporal sample as a newly started render for fusing purposes.
if (temporalSample > 0)
m_Calls = 0;
@ -830,7 +833,7 @@ bool RendererCL<T>::RunIter(size_t iterCount, size_t temporalSample, size_t& ite
if (b && !(b = m_Wrapper.SetArg (kernelIndex, argIndex++, supersize))) { m_ErrorReport.push_back(loc); }//Histogram size.
if (b && !(b = m_Wrapper.SetImageArg (kernelIndex, argIndex++, false, "Palette"))) { m_ErrorReport.push_back(loc); }//Palette.
if (b && !(b = m_Wrapper.SetBufferArg(kernelIndex, argIndex++, m_PointsBufferName))) { m_ErrorReport.push_back(loc); }//Random start points.
if (b && !(b = m_Wrapper.RunKernel(kernelIndex,
gridW * IterBlockKernelWidth(),//Total grid dims.
gridH * IterBlockKernelHeight(),
@ -840,10 +843,10 @@ bool RendererCL<T>::RunIter(size_t iterCount, size_t temporalSample, size_t& ite
1)))
{
m_Abort = true;
m_ErrorReport.push_back(loc);
m_ErrorReport.push_back(loc);
break;
}
itersRan += iterCountThisLaunch;
m_Calls++;
@ -863,14 +866,14 @@ bool RendererCL<T>::RunIter(size_t iterCount, size_t temporalSample, size_t& ite
double percentDiff = percent - m_LastIterPercent;
double toc = m_ProgressTimer.Toc();
if (percentDiff >= 10 || (toc > 1000 && percentDiff >= 1))//Call callback function if either 10% has passed, or one second (and 1%).
{
etaMs = ((100.0 - percent) / percent) * m_RenderTimer.Toc();
if (!m_Callback->ProgressFunc(m_Ember, m_ProgressParameter, percent, 0, etaMs))
Abort();
m_LastIterPercent = percent;
m_ProgressTimer.Tic();
}
@ -901,7 +904,7 @@ eRenderStatus RendererCL<T>::RunLogScaleFilter()
eRenderStatus status = RENDER_OK;
if (kernelIndex != -1)
{
{
m_DensityFilterCL = ConvertDensityFilter();
unsigned int argIndex = 0;
unsigned int blockW = m_WarpSize;
@ -910,7 +913,7 @@ eRenderStatus RendererCL<T>::RunLogScaleFilter()
unsigned int gridH = m_DensityFilterCL.m_SuperRasH;
OpenCLWrapper::MakeEvenGridDims(blockW, blockH, gridW, gridH);
if (b && !(b = m_Wrapper.AddAndWriteBuffer(m_DEFilterParamsBufferName, (void*)&m_DensityFilterCL, sizeof(m_DensityFilterCL)))) { m_ErrorReport.push_back(loc); }
if (b && !(b = m_Wrapper.SetBufferArg(kernelIndex, argIndex++, m_HistBufferName))) { m_ErrorReport.push_back(loc); }//Histogram.
@ -926,7 +929,7 @@ eRenderStatus RendererCL<T>::RunLogScaleFilter()
b = false;
m_ErrorReport.push_back(loc);
}
if (b && m_Callback)
m_Callback->ProgressFunc(m_Ember, m_ProgressParameter, 100.0, 1, 0.0);
@ -1030,7 +1033,7 @@ eRenderStatus RendererCL<T>::RunDensityFilter()
{
double percent = (double((rowChunk * chunkSizeW) + (colChunk + 1)) / totalChunks) * 100.0;
double etaMs = ((100.0 - percent) / percent) * t.Toc();
if (!m_Callback->ProgressFunc(m_Ember, m_ProgressParameter, percent, 1, etaMs))
Abort();
}
@ -1040,7 +1043,7 @@ eRenderStatus RendererCL<T>::RunDensityFilter()
if (b && m_Callback)
m_Callback->ProgressFunc(m_Ember, m_ProgressParameter, 100.0, 1, 0.0);
//t2.Toc(__FUNCTION__ " all passes");
}
else
@ -1048,10 +1051,10 @@ eRenderStatus RendererCL<T>::RunDensityFilter()
b = false;
m_ErrorReport.push_back(loc);
}
return m_Abort ? RENDER_ABORT : (b ? RENDER_OK : RENDER_ERROR);
}
/// <summary>
/// Run final accumulation to the 2D output image.
/// </summary>
@ -1096,13 +1099,13 @@ eRenderStatus RendererCL<T>::RunFinalAccum()
if (b && !(b = m_Wrapper.SetBufferArg(gammaCorrectKernelIndex, argIndex++, m_AccumBufferName))) { m_ErrorReport.push_back(loc); }//Accumulator.
if (b && !(b = m_Wrapper.SetBufferArg(gammaCorrectKernelIndex, argIndex++, m_SpatialFilterParamsBufferName))) { m_ErrorReport.push_back(loc); }//SpatialFilterCL.
if (b && !(b = m_Wrapper.RunKernel(gammaCorrectKernelIndex, gridW, gridH, 1, blockW, blockH, 1))) { m_ErrorReport.push_back(loc); }
}
else
{
b = false;
m_ErrorReport.push_back(loc);
m_ErrorReport.push_back(loc);
}
}
@ -1122,9 +1125,9 @@ eRenderStatus RendererCL<T>::RunFinalAccum()
if (b && m_Wrapper.Shared())
if (b && !(b = m_Wrapper.EnqueueAcquireGLObjects(m_FinalImageName))) { m_ErrorReport.push_back(loc); }
if (b && !(b = m_Wrapper.RunKernel(accumKernelIndex, gridW, gridH, 1, blockW, blockH, 1))) { m_ErrorReport.push_back(loc); }
if (b && m_Wrapper.Shared())
if (b && !(b = m_Wrapper.EnqueueReleaseGLObjects(m_FinalImageName))) { m_ErrorReport.push_back(loc); }
@ -1135,7 +1138,7 @@ eRenderStatus RendererCL<T>::RunFinalAccum()
b = false;
m_ErrorReport.push_back(loc);
}
return b ? RENDER_OK : RENDER_ERROR;
}
@ -1350,7 +1353,7 @@ SpatialFilterCL<T> RendererCL<T>::ConvertSpatialFilter()
Color<T> background;
SpatialFilterCL<T> filterCL;
PrepFinalAccumVals(background, g, linRange, vibrancy);
this->PrepFinalAccumVals(background, g, linRange, vibrancy);
filterCL.m_SuperRasW = (unsigned int)SuperRasW();
filterCL.m_SuperRasH = (unsigned int)SuperRasH();
@ -1368,7 +1371,7 @@ SpatialFilterCL<T> RendererCL<T>::ConvertSpatialFilter()
filterCL.m_Gamma = g;
filterCL.m_LinRange = linRange;
filterCL.m_Background = background;
return filterCL;
}
@ -1467,4 +1470,10 @@ void RendererCL<T>::FillSeeds()
m_Seeds[i].y = m_Rand[0].Rand();
}
}
}
template EMBERCL_API class RendererCL<float>;
#ifdef DO_DOUBLE
template EMBERCL_API class RendererCL<double>;
#endif
}

36
Source/EmberCL/RendererCL.h
View File

@ -18,6 +18,7 @@ namespace EmberCLns
class EMBERCL_API RendererCLBase
{
public:
virtual ~RendererCLBase() { }
virtual bool ReadFinal(unsigned char* pixels) = 0;
virtual bool ClearFinal() = 0;
};
@ -33,8 +34,25 @@ public:
/// and uses both for the base.
/// </summary>
template <typename T>
class EMBERCL_API RendererCL : public RendererCLBase, public Renderer<T, T>
class EMBERCL_API RendererCL : public Renderer<T, T>, public RendererCLBase
{
using EmberNs::Renderer<T, T>::RendererBase::EmberReport::m_ErrorReport;
using EmberNs::Renderer<T, T>::RendererBase::m_ProgressParameter;
using EmberNs::Renderer<T, T>::RendererBase::m_YAxisUp;
using EmberNs::Renderer<T, T>::RendererBase::m_LockAccum;
using EmberNs::Renderer<T, T>::RendererBase::m_Abort;
using EmberNs::Renderer<T, T>::RendererBase::m_NumChannels;
using EmberNs::Renderer<T, T>::RendererBase::m_LastIter;
using EmberNs::Renderer<T, T>::RendererBase::m_LastIterPercent;
using EmberNs::Renderer<T, T>::RendererBase::m_Stats;
using EmberNs::Renderer<T, T>::RendererBase::m_Callback;
using EmberNs::Renderer<T, T>::RendererBase::m_Rand;
using EmberNs::Renderer<T, T>::RendererBase::m_RenderTimer;
using EmberNs::Renderer<T, T>::RendererBase::m_IterTimer;
using EmberNs::Renderer<T, T>::RendererBase::m_ProgressTimer;
using EmberNs::Renderer<T, T>::m_RotMat;
using EmberNs::Renderer<T, T>::m_Ember;
public:
RendererCL(unsigned int platform = 0, unsigned int device = 0, bool shared = false, GLuint outputTexID = 0);
~RendererCL();
@ -158,7 +176,7 @@ private:
//Kernels.
string m_IterKernel;
OpenCLWrapper m_Wrapper;
cl::ImageFormat m_PaletteFormat;
cl::ImageFormat m_FinalFormat;
@ -168,7 +186,7 @@ private:
EmberCL<T> m_EmberCL;
vector<XformCL<T>> m_XformsCL;
vector<glm::highp_uvec2> m_Seeds;
Palette<float> m_Dmap;//Used instead of the base class' m_Dmap because OpenCL only supports float textures.
Palette<float> m_DmapCL;//Used instead of the base class' m_Dmap because OpenCL only supports float textures.
CarToRasCL<T> m_CarToRasCL;
DensityFilterCL<T> m_DensityFilterCL;
SpatialFilterCL<T> m_SpatialFilterCL;
@ -179,9 +197,9 @@ private:
Ember<T> m_LastBuiltEmber;
};
template EMBERCL_API class RendererCL<float>;
#ifdef DO_DOUBLE
template EMBERCL_API class RendererCL<double>;
#endif
}
//template EMBERCL_API class RendererCL<float>;
//
//#ifdef DO_DOUBLE
// template EMBERCL_API class RendererCL<double>;
//#endif
}