fractorium/Source/EmberCL/RendererCL.h

235 lines
8.9 KiB
C
Raw Normal View History

#pragma once
#include "EmberCLPch.h"
#include "OpenCLWrapper.h"
#include "IterOpenCLKernelCreator.h"
#include "DEOpenCLKernelCreator.h"
#include "FinalAccumOpenCLKernelCreator.h"
/// <summary>
/// RendererCLBase and RendererCL classes.
/// </summary>
namespace EmberCLns
{
/// <summary>
/// Serves only as an interface for OpenCL specific rendering functions.
/// </summary>
class EMBERCL_API RendererCLBase
{
public:
virtual ~RendererCLBase() { }
virtual bool ReadFinal(byte* pixels) = 0;
0.4.1.3 Beta 10/14/2014 --User Changes Size is no longer fixed to the window size. Size scaling is done differently in the final render dialog. This fixes several bugs. Remove Xml saving size from settings and options dialog, it no longer applies. Final render can be broken into strips. Set default save path to the desktop if none is found in the settings file. Set default output size to 1920x1080 if none is found in the settings file. --Bug Fixes Better memory size reporting in final render dialog. --Code Changes Migrate to C++11, Qt 5.3.1, and Visual Studio 2013. Change most instances of unsigned int to size_t, and int to intmax_t. Add m_OrigPixPerUnit and m_ScaleType to Ember for scaling purposes. Replace some sprintf_s() calls in XmlToEmber with ostringstream. Move more non-templated members into RendererBase. Add CopyVec() overload that takes a per element function pointer. Add vector Memset(). Replace '&' with '+' instead of "&amp;" in XmlToEmber for much faster parsing. Break strips rendering out into EmberCommon and call from EmberRender and Fractorium. Make AddAndWriteBuffer() just call WriteBuffer(). Make AddAndWriteImage() delete the existing image first before replacing it. Add SetOutputTexture() to RendererCL to support making new textures in response to resize events. Remove multiple return statements in RendererCL, and replace with a bool that tracks results. Add ToDouble(), MakeEnd(), ToString() and Exists() wrappers in Fractorium. Add Size() wrapper in EmberFile. Make QString function arguments const QString&, and string with const string&. Make ShowCritical() wrapper for invoking a message box from another thread. Add combo box to TwoButtonWidget and rename.
2014-10-14 11:53:15 -04:00
virtual bool ClearFinal() = 0;
};
/// <summary>
/// RendererCL is a derivation of the basic CPU renderer which
/// overrides various functions to render on the GPU using OpenCL.
/// Since this class derives from EmberReport and also contains an
/// OpenCLWrapper member which also derives from EmberReport, the
/// reporting functions are overridden to aggregate the errors from
/// both sources.
/// It does not support different types for T and bucketT, so it only has one template argument
/// and uses both for the base.
/// </summary>
template <typename T, typename bucketT>
class EMBERCL_API RendererCL : public Renderer<T, bucketT>, public RendererCLBase
{
using EmberNs::Renderer<T, bucketT>::RendererBase::Abort;
using EmberNs::Renderer<T, bucketT>::RendererBase::EarlyClip;
using EmberNs::Renderer<T, bucketT>::RendererBase::Transparency;
using EmberNs::Renderer<T, bucketT>::RendererBase::EnterResize;
using EmberNs::Renderer<T, bucketT>::RendererBase::LeaveResize;
using EmberNs::Renderer<T, bucketT>::RendererBase::FinalRasW;
using EmberNs::Renderer<T, bucketT>::RendererBase::FinalRasH;
using EmberNs::Renderer<T, bucketT>::RendererBase::SuperRasW;
using EmberNs::Renderer<T, bucketT>::RendererBase::SuperRasH;
using EmberNs::Renderer<T, bucketT>::RendererBase::SuperSize;
using EmberNs::Renderer<T, bucketT>::RendererBase::BytesPerChannel;
using EmberNs::Renderer<T, bucketT>::RendererBase::TemporalSamples;
using EmberNs::Renderer<T, bucketT>::RendererBase::ItersPerTemporalSample;
using EmberNs::Renderer<T, bucketT>::RendererBase::FuseCount;
using EmberNs::Renderer<T, bucketT>::RendererBase::DensityFilterOffset;
using EmberNs::Renderer<T, bucketT>::RendererBase::m_ProgressParameter;
using EmberNs::Renderer<T, bucketT>::RendererBase::m_YAxisUp;
using EmberNs::Renderer<T, bucketT>::RendererBase::m_LockAccum;
using EmberNs::Renderer<T, bucketT>::RendererBase::m_Abort;
using EmberNs::Renderer<T, bucketT>::RendererBase::m_NumChannels;
using EmberNs::Renderer<T, bucketT>::RendererBase::m_LastIter;
using EmberNs::Renderer<T, bucketT>::RendererBase::m_LastIterPercent;
using EmberNs::Renderer<T, bucketT>::RendererBase::m_Stats;
using EmberNs::Renderer<T, bucketT>::RendererBase::m_Callback;
using EmberNs::Renderer<T, bucketT>::RendererBase::m_Rand;
using EmberNs::Renderer<T, bucketT>::RendererBase::m_RenderTimer;
using EmberNs::Renderer<T, bucketT>::RendererBase::m_IterTimer;
using EmberNs::Renderer<T, bucketT>::RendererBase::m_ProgressTimer;
using EmberNs::Renderer<T, bucketT>::RendererBase::EmberReport::m_ErrorReport;
using EmberNs::Renderer<T, bucketT>::m_RotMat;
using EmberNs::Renderer<T, bucketT>::m_Ember;
using EmberNs::Renderer<T, bucketT>::m_Csa;
using EmberNs::Renderer<T, bucketT>::m_CurvesSet;
using EmberNs::Renderer<T, bucketT>::CenterX;
using EmberNs::Renderer<T, bucketT>::CenterY;
using EmberNs::Renderer<T, bucketT>::K1;
using EmberNs::Renderer<T, bucketT>::K2;
using EmberNs::Renderer<T, bucketT>::Supersample;
using EmberNs::Renderer<T, bucketT>::HighlightPower;
using EmberNs::Renderer<T, bucketT>::HistBuckets;
using EmberNs::Renderer<T, bucketT>::AccumulatorBuckets;
using EmberNs::Renderer<T, bucketT>::GetDensityFilter;
using EmberNs::Renderer<T, bucketT>::GetSpatialFilter;
using EmberNs::Renderer<T, bucketT>::CoordMap;
using EmberNs::Renderer<T, bucketT>::XformDistributions;
using EmberNs::Renderer<T, bucketT>::XformDistributionsSize;
public:
RendererCL(uint platform = 0, uint device = 0, bool shared = false, GLuint outputTexID = 0);
~RendererCL();
0.4.1.3 Beta 10/14/2014 --User Changes Size is no longer fixed to the window size. Size scaling is done differently in the final render dialog. This fixes several bugs. Remove Xml saving size from settings and options dialog, it no longer applies. Final render can be broken into strips. Set default save path to the desktop if none is found in the settings file. Set default output size to 1920x1080 if none is found in the settings file. --Bug Fixes Better memory size reporting in final render dialog. --Code Changes Migrate to C++11, Qt 5.3.1, and Visual Studio 2013. Change most instances of unsigned int to size_t, and int to intmax_t. Add m_OrigPixPerUnit and m_ScaleType to Ember for scaling purposes. Replace some sprintf_s() calls in XmlToEmber with ostringstream. Move more non-templated members into RendererBase. Add CopyVec() overload that takes a per element function pointer. Add vector Memset(). Replace '&' with '+' instead of "&amp;" in XmlToEmber for much faster parsing. Break strips rendering out into EmberCommon and call from EmberRender and Fractorium. Make AddAndWriteBuffer() just call WriteBuffer(). Make AddAndWriteImage() delete the existing image first before replacing it. Add SetOutputTexture() to RendererCL to support making new textures in response to resize events. Remove multiple return statements in RendererCL, and replace with a bool that tracks results. Add ToDouble(), MakeEnd(), ToString() and Exists() wrappers in Fractorium. Add Size() wrapper in EmberFile. Make QString function arguments const QString&, and string with const string&. Make ShowCritical() wrapper for invoking a message box from another thread. Add combo box to TwoButtonWidget and rename.
2014-10-14 11:53:15 -04:00
//Non-virtual member functions for OpenCL specific tasks.
bool Init(uint platform, uint device, bool shared, GLuint outputTexID);
0.4.1.3 Beta 10/14/2014 --User Changes Size is no longer fixed to the window size. Size scaling is done differently in the final render dialog. This fixes several bugs. Remove Xml saving size from settings and options dialog, it no longer applies. Final render can be broken into strips. Set default save path to the desktop if none is found in the settings file. Set default output size to 1920x1080 if none is found in the settings file. --Bug Fixes Better memory size reporting in final render dialog. --Code Changes Migrate to C++11, Qt 5.3.1, and Visual Studio 2013. Change most instances of unsigned int to size_t, and int to intmax_t. Add m_OrigPixPerUnit and m_ScaleType to Ember for scaling purposes. Replace some sprintf_s() calls in XmlToEmber with ostringstream. Move more non-templated members into RendererBase. Add CopyVec() overload that takes a per element function pointer. Add vector Memset(). Replace '&' with '+' instead of "&amp;" in XmlToEmber for much faster parsing. Break strips rendering out into EmberCommon and call from EmberRender and Fractorium. Make AddAndWriteBuffer() just call WriteBuffer(). Make AddAndWriteImage() delete the existing image first before replacing it. Add SetOutputTexture() to RendererCL to support making new textures in response to resize events. Remove multiple return statements in RendererCL, and replace with a bool that tracks results. Add ToDouble(), MakeEnd(), ToString() and Exists() wrappers in Fractorium. Add Size() wrapper in EmberFile. Make QString function arguments const QString&, and string with const string&. Make ShowCritical() wrapper for invoking a message box from another thread. Add combo box to TwoButtonWidget and rename.
2014-10-14 11:53:15 -04:00
bool SetOutputTexture(GLuint outputTexID);
0.4.1.5 Beta 11/28/2014 --User Changes Remove limit on the number of xforms allowable on the GPU. This was previously 21. Show actual strips count to be used in parens outside of user specified strips count on final render dialog. Allow for adjustment of iteration depth and fuse count per ember and save/read these values with the xml. Iteration optimizations on both CPU and GPU. Automatically adjust default quality spinner value when using CPU/GPU to 10/30, respectively. --Bug Fixes Fix severe randomization bug with OpenCL. Fix undo list off by one error when doing a new edit anywhere but the end of the undo list. Make integer variation parameters use 4 decimal places in the variations list like all the others. New build of the latest Qt to fix scroll bar drawing bug. Prevent grid from showing as much when pressing control to increase a spinner's increment speed. Still shows sometimes, but better than before. --Code Changes Pass count and fuse to iterator as a structure now to allow for passing more params in the future. Slightly different grid/block logic when running DE filtering on the GPU. Attempt a different way of doing DE, but #define out because it ended up not being faster. Restructure some things to allow for a variable length xforms buffer to be passed to the GPU. Add sub batch size and fuse count as ember members, and remove them from the renderer classes. Remove m_LastPass from Renderer. It should have been removed with passes. Pass seeds as a buffer to the OpenCL iteration kernel, rather than a single seed that gets modified. Slight optimization on CPU accum. Use case statement instead of if/else for xform chosing in OpenCL for a 2% speedup on params with large numbers of xforms. Add SizeOf() wrapper around sizeof(vec[0]) * vec.size(). Remove LogScaleSum() functions from the CPU and GPU because they're no longer used since passes were removed. Make some OpenCLWrapper getters const. Better ogranize RendererCL methods that return grid dimensions.
2014-11-28 04:37:51 -05:00
//Iters per kernel/block/grid.
inline uint IterCountPerKernel() const;
inline uint IterCountPerBlock() const;
inline uint IterCountPerGrid() const;
0.4.1.5 Beta 11/28/2014 --User Changes Remove limit on the number of xforms allowable on the GPU. This was previously 21. Show actual strips count to be used in parens outside of user specified strips count on final render dialog. Allow for adjustment of iteration depth and fuse count per ember and save/read these values with the xml. Iteration optimizations on both CPU and GPU. Automatically adjust default quality spinner value when using CPU/GPU to 10/30, respectively. --Bug Fixes Fix severe randomization bug with OpenCL. Fix undo list off by one error when doing a new edit anywhere but the end of the undo list. Make integer variation parameters use 4 decimal places in the variations list like all the others. New build of the latest Qt to fix scroll bar drawing bug. Prevent grid from showing as much when pressing control to increase a spinner's increment speed. Still shows sometimes, but better than before. --Code Changes Pass count and fuse to iterator as a structure now to allow for passing more params in the future. Slightly different grid/block logic when running DE filtering on the GPU. Attempt a different way of doing DE, but #define out because it ended up not being faster. Restructure some things to allow for a variable length xforms buffer to be passed to the GPU. Add sub batch size and fuse count as ember members, and remove them from the renderer classes. Remove m_LastPass from Renderer. It should have been removed with passes. Pass seeds as a buffer to the OpenCL iteration kernel, rather than a single seed that gets modified. Slight optimization on CPU accum. Use case statement instead of if/else for xform chosing in OpenCL for a 2% speedup on params with large numbers of xforms. Add SizeOf() wrapper around sizeof(vec[0]) * vec.size(). Remove LogScaleSum() functions from the CPU and GPU because they're no longer used since passes were removed. Make some OpenCLWrapper getters const. Better ogranize RendererCL methods that return grid dimensions.
2014-11-28 04:37:51 -05:00
//Kernels per block.
inline uint IterBlockKernelWidth() const;
inline uint IterBlockKernelHeight() const;
inline uint IterBlockKernelCount() const;
0.4.1.5 Beta 11/28/2014 --User Changes Remove limit on the number of xforms allowable on the GPU. This was previously 21. Show actual strips count to be used in parens outside of user specified strips count on final render dialog. Allow for adjustment of iteration depth and fuse count per ember and save/read these values with the xml. Iteration optimizations on both CPU and GPU. Automatically adjust default quality spinner value when using CPU/GPU to 10/30, respectively. --Bug Fixes Fix severe randomization bug with OpenCL. Fix undo list off by one error when doing a new edit anywhere but the end of the undo list. Make integer variation parameters use 4 decimal places in the variations list like all the others. New build of the latest Qt to fix scroll bar drawing bug. Prevent grid from showing as much when pressing control to increase a spinner's increment speed. Still shows sometimes, but better than before. --Code Changes Pass count and fuse to iterator as a structure now to allow for passing more params in the future. Slightly different grid/block logic when running DE filtering on the GPU. Attempt a different way of doing DE, but #define out because it ended up not being faster. Restructure some things to allow for a variable length xforms buffer to be passed to the GPU. Add sub batch size and fuse count as ember members, and remove them from the renderer classes. Remove m_LastPass from Renderer. It should have been removed with passes. Pass seeds as a buffer to the OpenCL iteration kernel, rather than a single seed that gets modified. Slight optimization on CPU accum. Use case statement instead of if/else for xform chosing in OpenCL for a 2% speedup on params with large numbers of xforms. Add SizeOf() wrapper around sizeof(vec[0]) * vec.size(). Remove LogScaleSum() functions from the CPU and GPU because they're no longer used since passes were removed. Make some OpenCLWrapper getters const. Better ogranize RendererCL methods that return grid dimensions.
2014-11-28 04:37:51 -05:00
//Kernels per grid.
inline uint IterGridKernelWidth() const;
inline uint IterGridKernelHeight() const;
inline uint IterGridKernelCount() const;
0.4.1.5 Beta 11/28/2014 --User Changes Remove limit on the number of xforms allowable on the GPU. This was previously 21. Show actual strips count to be used in parens outside of user specified strips count on final render dialog. Allow for adjustment of iteration depth and fuse count per ember and save/read these values with the xml. Iteration optimizations on both CPU and GPU. Automatically adjust default quality spinner value when using CPU/GPU to 10/30, respectively. --Bug Fixes Fix severe randomization bug with OpenCL. Fix undo list off by one error when doing a new edit anywhere but the end of the undo list. Make integer variation parameters use 4 decimal places in the variations list like all the others. New build of the latest Qt to fix scroll bar drawing bug. Prevent grid from showing as much when pressing control to increase a spinner's increment speed. Still shows sometimes, but better than before. --Code Changes Pass count and fuse to iterator as a structure now to allow for passing more params in the future. Slightly different grid/block logic when running DE filtering on the GPU. Attempt a different way of doing DE, but #define out because it ended up not being faster. Restructure some things to allow for a variable length xforms buffer to be passed to the GPU. Add sub batch size and fuse count as ember members, and remove them from the renderer classes. Remove m_LastPass from Renderer. It should have been removed with passes. Pass seeds as a buffer to the OpenCL iteration kernel, rather than a single seed that gets modified. Slight optimization on CPU accum. Use case statement instead of if/else for xform chosing in OpenCL for a 2% speedup on params with large numbers of xforms. Add SizeOf() wrapper around sizeof(vec[0]) * vec.size(). Remove LogScaleSum() functions from the CPU and GPU because they're no longer used since passes were removed. Make some OpenCLWrapper getters const. Better ogranize RendererCL methods that return grid dimensions.
2014-11-28 04:37:51 -05:00
//Blocks per grid.
inline uint IterGridBlockWidth() const;
inline uint IterGridBlockHeight() const;
inline uint IterGridBlockCount() const;
0.4.1.5 Beta 11/28/2014 --User Changes Remove limit on the number of xforms allowable on the GPU. This was previously 21. Show actual strips count to be used in parens outside of user specified strips count on final render dialog. Allow for adjustment of iteration depth and fuse count per ember and save/read these values with the xml. Iteration optimizations on both CPU and GPU. Automatically adjust default quality spinner value when using CPU/GPU to 10/30, respectively. --Bug Fixes Fix severe randomization bug with OpenCL. Fix undo list off by one error when doing a new edit anywhere but the end of the undo list. Make integer variation parameters use 4 decimal places in the variations list like all the others. New build of the latest Qt to fix scroll bar drawing bug. Prevent grid from showing as much when pressing control to increase a spinner's increment speed. Still shows sometimes, but better than before. --Code Changes Pass count and fuse to iterator as a structure now to allow for passing more params in the future. Slightly different grid/block logic when running DE filtering on the GPU. Attempt a different way of doing DE, but #define out because it ended up not being faster. Restructure some things to allow for a variable length xforms buffer to be passed to the GPU. Add sub batch size and fuse count as ember members, and remove them from the renderer classes. Remove m_LastPass from Renderer. It should have been removed with passes. Pass seeds as a buffer to the OpenCL iteration kernel, rather than a single seed that gets modified. Slight optimization on CPU accum. Use case statement instead of if/else for xform chosing in OpenCL for a 2% speedup on params with large numbers of xforms. Add SizeOf() wrapper around sizeof(vec[0]) * vec.size(). Remove LogScaleSum() functions from the CPU and GPU because they're no longer used since passes were removed. Make some OpenCLWrapper getters const. Better ogranize RendererCL methods that return grid dimensions.
2014-11-28 04:37:51 -05:00
uint PlatformIndex();
uint DeviceIndex();
bool ReadHist();
bool ReadAccum();
bool ReadPoints(vector<PointCL<T>>& vec);
bool ClearHist();
bool ClearAccum();
bool WritePoints(vector<PointCL<T>>& vec);
0.4.1.5 Beta 11/28/2014 --User Changes Remove limit on the number of xforms allowable on the GPU. This was previously 21. Show actual strips count to be used in parens outside of user specified strips count on final render dialog. Allow for adjustment of iteration depth and fuse count per ember and save/read these values with the xml. Iteration optimizations on both CPU and GPU. Automatically adjust default quality spinner value when using CPU/GPU to 10/30, respectively. --Bug Fixes Fix severe randomization bug with OpenCL. Fix undo list off by one error when doing a new edit anywhere but the end of the undo list. Make integer variation parameters use 4 decimal places in the variations list like all the others. New build of the latest Qt to fix scroll bar drawing bug. Prevent grid from showing as much when pressing control to increase a spinner's increment speed. Still shows sometimes, but better than before. --Code Changes Pass count and fuse to iterator as a structure now to allow for passing more params in the future. Slightly different grid/block logic when running DE filtering on the GPU. Attempt a different way of doing DE, but #define out because it ended up not being faster. Restructure some things to allow for a variable length xforms buffer to be passed to the GPU. Add sub batch size and fuse count as ember members, and remove them from the renderer classes. Remove m_LastPass from Renderer. It should have been removed with passes. Pass seeds as a buffer to the OpenCL iteration kernel, rather than a single seed that gets modified. Slight optimization on CPU accum. Use case statement instead of if/else for xform chosing in OpenCL for a 2% speedup on params with large numbers of xforms. Add SizeOf() wrapper around sizeof(vec[0]) * vec.size(). Remove LogScaleSum() functions from the CPU and GPU because they're no longer used since passes were removed. Make some OpenCLWrapper getters const. Better ogranize RendererCL methods that return grid dimensions.
2014-11-28 04:37:51 -05:00
#ifdef TEST_CL
bool WriteRandomPoints();
#endif
string IterKernel();
string DEKernel();
string FinalAccumKernel();
0.4.1.3 Beta 10/14/2014 --User Changes Size is no longer fixed to the window size. Size scaling is done differently in the final render dialog. This fixes several bugs. Remove Xml saving size from settings and options dialog, it no longer applies. Final render can be broken into strips. Set default save path to the desktop if none is found in the settings file. Set default output size to 1920x1080 if none is found in the settings file. --Bug Fixes Better memory size reporting in final render dialog. --Code Changes Migrate to C++11, Qt 5.3.1, and Visual Studio 2013. Change most instances of unsigned int to size_t, and int to intmax_t. Add m_OrigPixPerUnit and m_ScaleType to Ember for scaling purposes. Replace some sprintf_s() calls in XmlToEmber with ostringstream. Move more non-templated members into RendererBase. Add CopyVec() overload that takes a per element function pointer. Add vector Memset(). Replace '&' with '+' instead of "&amp;" in XmlToEmber for much faster parsing. Break strips rendering out into EmberCommon and call from EmberRender and Fractorium. Make AddAndWriteBuffer() just call WriteBuffer(). Make AddAndWriteImage() delete the existing image first before replacing it. Add SetOutputTexture() to RendererCL to support making new textures in response to resize events. Remove multiple return statements in RendererCL, and replace with a bool that tracks results. Add ToDouble(), MakeEnd(), ToString() and Exists() wrappers in Fractorium. Add Size() wrapper in EmberFile. Make QString function arguments const QString&, and string with const string&. Make ShowCritical() wrapper for invoking a message box from another thread. Add combo box to TwoButtonWidget and rename.
2014-10-14 11:53:15 -04:00
//Virtual functions overridden from RendererCLBase.
virtual bool ReadFinal(byte* pixels);
0.4.1.3 Beta 10/14/2014 --User Changes Size is no longer fixed to the window size. Size scaling is done differently in the final render dialog. This fixes several bugs. Remove Xml saving size from settings and options dialog, it no longer applies. Final render can be broken into strips. Set default save path to the desktop if none is found in the settings file. Set default output size to 1920x1080 if none is found in the settings file. --Bug Fixes Better memory size reporting in final render dialog. --Code Changes Migrate to C++11, Qt 5.3.1, and Visual Studio 2013. Change most instances of unsigned int to size_t, and int to intmax_t. Add m_OrigPixPerUnit and m_ScaleType to Ember for scaling purposes. Replace some sprintf_s() calls in XmlToEmber with ostringstream. Move more non-templated members into RendererBase. Add CopyVec() overload that takes a per element function pointer. Add vector Memset(). Replace '&' with '+' instead of "&amp;" in XmlToEmber for much faster parsing. Break strips rendering out into EmberCommon and call from EmberRender and Fractorium. Make AddAndWriteBuffer() just call WriteBuffer(). Make AddAndWriteImage() delete the existing image first before replacing it. Add SetOutputTexture() to RendererCL to support making new textures in response to resize events. Remove multiple return statements in RendererCL, and replace with a bool that tracks results. Add ToDouble(), MakeEnd(), ToString() and Exists() wrappers in Fractorium. Add Size() wrapper in EmberFile. Make QString function arguments const QString&, and string with const string&. Make ShowCritical() wrapper for invoking a message box from another thread. Add combo box to TwoButtonWidget and rename.
2014-10-14 11:53:15 -04:00
virtual bool ClearFinal();
//Public virtual functions overridden from Renderer or RendererBase.
virtual size_t MemoryAvailable() override;
virtual bool Ok() const override;
virtual void NumChannels(size_t numChannels) override;
virtual void DumpErrorReport() override;
virtual void ClearErrorReport() override;
virtual size_t SubBatchSize() const override;
virtual size_t ThreadCount() const override;
virtual bool CreateDEFilter(bool& newAlloc) override;
virtual bool CreateSpatialFilter(bool& newAlloc) override;
virtual eRendererType RendererType() const override;
virtual string ErrorReportString() override;
virtual vector<string> ErrorReport() override;
virtual bool RandVec(vector<QTIsaac<ISAAC_SIZE, ISAAC_INT>>& randVec) override;
#ifndef TEST_CL
protected:
#endif
0.4.1.3 Beta 10/14/2014 --User Changes Size is no longer fixed to the window size. Size scaling is done differently in the final render dialog. This fixes several bugs. Remove Xml saving size from settings and options dialog, it no longer applies. Final render can be broken into strips. Set default save path to the desktop if none is found in the settings file. Set default output size to 1920x1080 if none is found in the settings file. --Bug Fixes Better memory size reporting in final render dialog. --Code Changes Migrate to C++11, Qt 5.3.1, and Visual Studio 2013. Change most instances of unsigned int to size_t, and int to intmax_t. Add m_OrigPixPerUnit and m_ScaleType to Ember for scaling purposes. Replace some sprintf_s() calls in XmlToEmber with ostringstream. Move more non-templated members into RendererBase. Add CopyVec() overload that takes a per element function pointer. Add vector Memset(). Replace '&' with '+' instead of "&amp;" in XmlToEmber for much faster parsing. Break strips rendering out into EmberCommon and call from EmberRender and Fractorium. Make AddAndWriteBuffer() just call WriteBuffer(). Make AddAndWriteImage() delete the existing image first before replacing it. Add SetOutputTexture() to RendererCL to support making new textures in response to resize events. Remove multiple return statements in RendererCL, and replace with a bool that tracks results. Add ToDouble(), MakeEnd(), ToString() and Exists() wrappers in Fractorium. Add Size() wrapper in EmberFile. Make QString function arguments const QString&, and string with const string&. Make ShowCritical() wrapper for invoking a message box from another thread. Add combo box to TwoButtonWidget and rename.
2014-10-14 11:53:15 -04:00
//Protected virtual functions overridden from Renderer.
virtual void MakeDmap(T colorScalar) override;
virtual bool Alloc() override;
virtual bool ResetBuckets(bool resetHist = true, bool resetAccum = true) override;
virtual eRenderStatus LogScaleDensityFilter() override;
virtual eRenderStatus GaussianDensityFilter() override;
virtual eRenderStatus AccumulatorToFinalImage(byte* pixels, size_t finalOffset) override;
virtual EmberStats Iterate(size_t iterCount, size_t temporalSample) override;
#ifndef TEST_CL
private:
#endif
//Private functions for making and running OpenCL programs.
bool BuildIterProgramForEmber(bool doAccum = true);
bool RunIter(size_t iterCount, size_t temporalSample, size_t& itersRan);
eRenderStatus RunLogScaleFilter();
eRenderStatus RunDensityFilter();
eRenderStatus RunFinalAccum();
bool ClearBuffer(const string& bufferName, uint width, uint height, uint elementSize);
bool RunDensityFilterPrivate(uint kernelIndex, uint gridW, uint gridH, uint blockW, uint blockH, uint chunkSizeW, uint chunkSizeH, uint chunkW, uint chunkH);
int MakeAndGetDensityFilterProgram(size_t ss, uint filterWidth);
int MakeAndGetFinalAccumProgram(double& alphaBase, double& alphaScale);
int MakeAndGetGammaCorrectionProgram();
void FillSeeds();
//Private functions passing data to OpenCL programs.
void ConvertDensityFilter();
void ConvertSpatialFilter();
0.4.1.5 Beta 11/28/2014 --User Changes Remove limit on the number of xforms allowable on the GPU. This was previously 21. Show actual strips count to be used in parens outside of user specified strips count on final render dialog. Allow for adjustment of iteration depth and fuse count per ember and save/read these values with the xml. Iteration optimizations on both CPU and GPU. Automatically adjust default quality spinner value when using CPU/GPU to 10/30, respectively. --Bug Fixes Fix severe randomization bug with OpenCL. Fix undo list off by one error when doing a new edit anywhere but the end of the undo list. Make integer variation parameters use 4 decimal places in the variations list like all the others. New build of the latest Qt to fix scroll bar drawing bug. Prevent grid from showing as much when pressing control to increase a spinner's increment speed. Still shows sometimes, but better than before. --Code Changes Pass count and fuse to iterator as a structure now to allow for passing more params in the future. Slightly different grid/block logic when running DE filtering on the GPU. Attempt a different way of doing DE, but #define out because it ended up not being faster. Restructure some things to allow for a variable length xforms buffer to be passed to the GPU. Add sub batch size and fuse count as ember members, and remove them from the renderer classes. Remove m_LastPass from Renderer. It should have been removed with passes. Pass seeds as a buffer to the OpenCL iteration kernel, rather than a single seed that gets modified. Slight optimization on CPU accum. Use case statement instead of if/else for xform chosing in OpenCL for a 2% speedup on params with large numbers of xforms. Add SizeOf() wrapper around sizeof(vec[0]) * vec.size(). Remove LogScaleSum() functions from the CPU and GPU because they're no longer used since passes were removed. Make some OpenCLWrapper getters const. Better ogranize RendererCL methods that return grid dimensions.
2014-11-28 04:37:51 -05:00
void ConvertEmber(Ember<T>& ember, EmberCL<T>& emberCL, vector<XformCL<T>>& xformsCL);
static CarToRasCL<T> ConvertCarToRas(const CarToRas<T>& carToRas);
bool m_Init;
bool m_NVidia;
bool m_DoublePrecision;
uint m_IterCountPerKernel;
uint m_IterBlocksWide, m_IterBlockWidth;
uint m_IterBlocksHigh, m_IterBlockHeight;
uint m_MaxDEBlockSizeW;
uint m_MaxDEBlockSizeH;
uint m_WarpSize;
0.4.1.3 Beta 10/14/2014 --User Changes Size is no longer fixed to the window size. Size scaling is done differently in the final render dialog. This fixes several bugs. Remove Xml saving size from settings and options dialog, it no longer applies. Final render can be broken into strips. Set default save path to the desktop if none is found in the settings file. Set default output size to 1920x1080 if none is found in the settings file. --Bug Fixes Better memory size reporting in final render dialog. --Code Changes Migrate to C++11, Qt 5.3.1, and Visual Studio 2013. Change most instances of unsigned int to size_t, and int to intmax_t. Add m_OrigPixPerUnit and m_ScaleType to Ember for scaling purposes. Replace some sprintf_s() calls in XmlToEmber with ostringstream. Move more non-templated members into RendererBase. Add CopyVec() overload that takes a per element function pointer. Add vector Memset(). Replace '&' with '+' instead of "&amp;" in XmlToEmber for much faster parsing. Break strips rendering out into EmberCommon and call from EmberRender and Fractorium. Make AddAndWriteBuffer() just call WriteBuffer(). Make AddAndWriteImage() delete the existing image first before replacing it. Add SetOutputTexture() to RendererCL to support making new textures in response to resize events. Remove multiple return statements in RendererCL, and replace with a bool that tracks results. Add ToDouble(), MakeEnd(), ToString() and Exists() wrappers in Fractorium. Add Size() wrapper in EmberFile. Make QString function arguments const QString&, and string with const string&. Make ShowCritical() wrapper for invoking a message box from another thread. Add combo box to TwoButtonWidget and rename.
2014-10-14 11:53:15 -04:00
size_t m_Calls;
0.4.1.3 Beta 10/14/2014 --User Changes Size is no longer fixed to the window size. Size scaling is done differently in the final render dialog. This fixes several bugs. Remove Xml saving size from settings and options dialog, it no longer applies. Final render can be broken into strips. Set default save path to the desktop if none is found in the settings file. Set default output size to 1920x1080 if none is found in the settings file. --Bug Fixes Better memory size reporting in final render dialog. --Code Changes Migrate to C++11, Qt 5.3.1, and Visual Studio 2013. Change most instances of unsigned int to size_t, and int to intmax_t. Add m_OrigPixPerUnit and m_ScaleType to Ember for scaling purposes. Replace some sprintf_s() calls in XmlToEmber with ostringstream. Move more non-templated members into RendererBase. Add CopyVec() overload that takes a per element function pointer. Add vector Memset(). Replace '&' with '+' instead of "&amp;" in XmlToEmber for much faster parsing. Break strips rendering out into EmberCommon and call from EmberRender and Fractorium. Make AddAndWriteBuffer() just call WriteBuffer(). Make AddAndWriteImage() delete the existing image first before replacing it. Add SetOutputTexture() to RendererCL to support making new textures in response to resize events. Remove multiple return statements in RendererCL, and replace with a bool that tracks results. Add ToDouble(), MakeEnd(), ToString() and Exists() wrappers in Fractorium. Add Size() wrapper in EmberFile. Make QString function arguments const QString&, and string with const string&. Make ShowCritical() wrapper for invoking a message box from another thread. Add combo box to TwoButtonWidget and rename.
2014-10-14 11:53:15 -04:00
//Buffer names.
string m_EmberBufferName;
0.4.1.5 Beta 11/28/2014 --User Changes Remove limit on the number of xforms allowable on the GPU. This was previously 21. Show actual strips count to be used in parens outside of user specified strips count on final render dialog. Allow for adjustment of iteration depth and fuse count per ember and save/read these values with the xml. Iteration optimizations on both CPU and GPU. Automatically adjust default quality spinner value when using CPU/GPU to 10/30, respectively. --Bug Fixes Fix severe randomization bug with OpenCL. Fix undo list off by one error when doing a new edit anywhere but the end of the undo list. Make integer variation parameters use 4 decimal places in the variations list like all the others. New build of the latest Qt to fix scroll bar drawing bug. Prevent grid from showing as much when pressing control to increase a spinner's increment speed. Still shows sometimes, but better than before. --Code Changes Pass count and fuse to iterator as a structure now to allow for passing more params in the future. Slightly different grid/block logic when running DE filtering on the GPU. Attempt a different way of doing DE, but #define out because it ended up not being faster. Restructure some things to allow for a variable length xforms buffer to be passed to the GPU. Add sub batch size and fuse count as ember members, and remove them from the renderer classes. Remove m_LastPass from Renderer. It should have been removed with passes. Pass seeds as a buffer to the OpenCL iteration kernel, rather than a single seed that gets modified. Slight optimization on CPU accum. Use case statement instead of if/else for xform chosing in OpenCL for a 2% speedup on params with large numbers of xforms. Add SizeOf() wrapper around sizeof(vec[0]) * vec.size(). Remove LogScaleSum() functions from the CPU and GPU because they're no longer used since passes were removed. Make some OpenCLWrapper getters const. Better ogranize RendererCL methods that return grid dimensions.
2014-11-28 04:37:51 -05:00
string m_XformsBufferName;
string m_ParVarsBufferName;
0.4.1.5 Beta 11/28/2014 --User Changes Remove limit on the number of xforms allowable on the GPU. This was previously 21. Show actual strips count to be used in parens outside of user specified strips count on final render dialog. Allow for adjustment of iteration depth and fuse count per ember and save/read these values with the xml. Iteration optimizations on both CPU and GPU. Automatically adjust default quality spinner value when using CPU/GPU to 10/30, respectively. --Bug Fixes Fix severe randomization bug with OpenCL. Fix undo list off by one error when doing a new edit anywhere but the end of the undo list. Make integer variation parameters use 4 decimal places in the variations list like all the others. New build of the latest Qt to fix scroll bar drawing bug. Prevent grid from showing as much when pressing control to increase a spinner's increment speed. Still shows sometimes, but better than before. --Code Changes Pass count and fuse to iterator as a structure now to allow for passing more params in the future. Slightly different grid/block logic when running DE filtering on the GPU. Attempt a different way of doing DE, but #define out because it ended up not being faster. Restructure some things to allow for a variable length xforms buffer to be passed to the GPU. Add sub batch size and fuse count as ember members, and remove them from the renderer classes. Remove m_LastPass from Renderer. It should have been removed with passes. Pass seeds as a buffer to the OpenCL iteration kernel, rather than a single seed that gets modified. Slight optimization on CPU accum. Use case statement instead of if/else for xform chosing in OpenCL for a 2% speedup on params with large numbers of xforms. Add SizeOf() wrapper around sizeof(vec[0]) * vec.size(). Remove LogScaleSum() functions from the CPU and GPU because they're no longer used since passes were removed. Make some OpenCLWrapper getters const. Better ogranize RendererCL methods that return grid dimensions.
2014-11-28 04:37:51 -05:00
string m_SeedsBufferName;
string m_DistBufferName;
string m_CarToRasBufferName;
string m_DEFilterParamsBufferName;
string m_SpatialFilterParamsBufferName;
string m_CurvesCsaName;
string m_DECoefsBufferName;
string m_DEWidthsBufferName;
string m_DECoefIndicesBufferName;
string m_SpatialFilterCoefsBufferName;
string m_HistBufferName;
string m_AccumBufferName;
string m_FinalImageName;
string m_PointsBufferName;
0.4.1.3 Beta 10/14/2014 --User Changes Size is no longer fixed to the window size. Size scaling is done differently in the final render dialog. This fixes several bugs. Remove Xml saving size from settings and options dialog, it no longer applies. Final render can be broken into strips. Set default save path to the desktop if none is found in the settings file. Set default output size to 1920x1080 if none is found in the settings file. --Bug Fixes Better memory size reporting in final render dialog. --Code Changes Migrate to C++11, Qt 5.3.1, and Visual Studio 2013. Change most instances of unsigned int to size_t, and int to intmax_t. Add m_OrigPixPerUnit and m_ScaleType to Ember for scaling purposes. Replace some sprintf_s() calls in XmlToEmber with ostringstream. Move more non-templated members into RendererBase. Add CopyVec() overload that takes a per element function pointer. Add vector Memset(). Replace '&' with '+' instead of "&amp;" in XmlToEmber for much faster parsing. Break strips rendering out into EmberCommon and call from EmberRender and Fractorium. Make AddAndWriteBuffer() just call WriteBuffer(). Make AddAndWriteImage() delete the existing image first before replacing it. Add SetOutputTexture() to RendererCL to support making new textures in response to resize events. Remove multiple return statements in RendererCL, and replace with a bool that tracks results. Add ToDouble(), MakeEnd(), ToString() and Exists() wrappers in Fractorium. Add Size() wrapper in EmberFile. Make QString function arguments const QString&, and string with const string&. Make ShowCritical() wrapper for invoking a message box from another thread. Add combo box to TwoButtonWidget and rename.
2014-10-14 11:53:15 -04:00
//Kernels.
string m_IterKernel;
OpenCLWrapper m_Wrapper;
cl::ImageFormat m_PaletteFormat;
cl::ImageFormat m_FinalFormat;
cl::Image2D m_Palette;
IMAGEGL2D m_AccumImage;
GLuint m_OutputTexID;
EmberCL<T> m_EmberCL;
0.4.1.5 Beta 11/28/2014 --User Changes Remove limit on the number of xforms allowable on the GPU. This was previously 21. Show actual strips count to be used in parens outside of user specified strips count on final render dialog. Allow for adjustment of iteration depth and fuse count per ember and save/read these values with the xml. Iteration optimizations on both CPU and GPU. Automatically adjust default quality spinner value when using CPU/GPU to 10/30, respectively. --Bug Fixes Fix severe randomization bug with OpenCL. Fix undo list off by one error when doing a new edit anywhere but the end of the undo list. Make integer variation parameters use 4 decimal places in the variations list like all the others. New build of the latest Qt to fix scroll bar drawing bug. Prevent grid from showing as much when pressing control to increase a spinner's increment speed. Still shows sometimes, but better than before. --Code Changes Pass count and fuse to iterator as a structure now to allow for passing more params in the future. Slightly different grid/block logic when running DE filtering on the GPU. Attempt a different way of doing DE, but #define out because it ended up not being faster. Restructure some things to allow for a variable length xforms buffer to be passed to the GPU. Add sub batch size and fuse count as ember members, and remove them from the renderer classes. Remove m_LastPass from Renderer. It should have been removed with passes. Pass seeds as a buffer to the OpenCL iteration kernel, rather than a single seed that gets modified. Slight optimization on CPU accum. Use case statement instead of if/else for xform chosing in OpenCL for a 2% speedup on params with large numbers of xforms. Add SizeOf() wrapper around sizeof(vec[0]) * vec.size(). Remove LogScaleSum() functions from the CPU and GPU because they're no longer used since passes were removed. Make some OpenCLWrapper getters const. Better ogranize RendererCL methods that return grid dimensions.
2014-11-28 04:37:51 -05:00
vector<XformCL<T>> m_XformsCL;
vector<glm::highp_uvec2> m_Seeds;
Palette<float> m_DmapCL;//Used instead of the base class' m_Dmap because OpenCL only supports float textures. Likely not needed if we switch to float only hist.
CarToRasCL<T> m_CarToRasCL;
DensityFilterCL<bucketT> m_DensityFilterCL;
SpatialFilterCL<bucketT> m_SpatialFilterCL;
IterOpenCLKernelCreator<T> m_IterOpenCLKernelCreator;
DEOpenCLKernelCreator m_DEOpenCLKernelCreator;
FinalAccumOpenCLKernelCreator m_FinalAccumOpenCLKernelCreator;
pair<string, vector<T>> m_Params;
Ember<T> m_LastBuiltEmber;
};
}