2014-07-08 03:11:14 -04:00
|
|
|
#pragma once
|
|
|
|
|
|
2014-10-14 11:53:15 -04:00
|
|
|
#include "RendererBase.h"
|
2014-07-08 03:11:14 -04:00
|
|
|
#include "Iterator.h"
|
|
|
|
|
#include "SpatialFilter.h"
|
|
|
|
|
#include "TemporalFilter.h"
|
|
|
|
|
#include "Interpolate.h"
|
|
|
|
|
#include "CarToRas.h"
|
|
|
|
|
#include "EmberToXml.h"
|
|
|
|
|
|
|
|
|
|
/// <summary>
|
2014-10-14 11:53:15 -04:00
|
|
|
/// Renderer.
|
2014-07-08 03:11:14 -04:00
|
|
|
/// </summary>
|
|
|
|
|
|
|
|
|
|
namespace EmberNs
|
|
|
|
|
{
|
|
|
|
|
/// <summary>
|
|
|
|
|
/// Renderer is the main class where all of the execution takes place.
|
|
|
|
|
/// It is intended that the program have one instance of it that it
|
|
|
|
|
/// keeps around for its duration. After a user sets up an ember, it's passed
|
|
|
|
|
/// in to be rendered.
|
|
|
|
|
/// This class derives from EmberReport, so the caller is able
|
|
|
|
|
/// to retrieve a text dump of error information if any errors occur.
|
|
|
|
|
/// The final image output vector is also passed in because the calling code has more
|
|
|
|
|
/// use for it than this class does.
|
|
|
|
|
/// Several functions are made virtual and have a default CPU-based implementation
|
2014-10-14 11:53:15 -04:00
|
|
|
/// that roughly matches what flam3 did. However they can be overridden in derived classes
|
2014-07-08 03:11:14 -04:00
|
|
|
/// to provide alternative rendering implementations, such as using the GPU.
|
|
|
|
|
/// Since this is a templated class, it's supposed to be entirely implemented in this .h file.
|
|
|
|
|
/// However, VC++ 2010 has very crippled support for lambdas, which Renderer makes use of.
|
|
|
|
|
/// If too many lambdas are used in a .h file, it will crash the compiler when another library
|
|
|
|
|
/// tries to link to it. To work around the bug, only declarations are here and all implementations
|
|
|
|
|
/// are in the .cpp file. It's unclear at what point it starts/stops working. But it seems that once
|
|
|
|
|
/// enough code is placed in the .h file, the compiler crashes. So for the sake of consistency, everything
|
|
|
|
|
/// is moved to the .cpp, even simple getters. One drawback however, is that the class must be
|
|
|
|
|
/// explicitly exported at the bottom of the file.
|
|
|
|
|
/// Also, despite explicitly doing this, the compiler throws a C4661 warning
|
|
|
|
|
/// for every single function in this class, saying it can't find the implementation. This warning
|
|
|
|
|
/// can be safely ignored.
|
|
|
|
|
/// Template argument T expected to be float or double.
|
--User changes
-Add support for multiple GPU devices.
--These options are present in the command line and in Fractorium.
-Change scheme of specifying devices from platform,device to just total device index.
--Single number on the command line.
--Change from combo boxes for device selection to a table of all devices in Fractorium.
-Temporal samples defaults to 100 instead of 1000 which was needless overkill.
--Bug fixes
-EmberAnimate, EmberRender, FractoriumSettings, FinalRenderDialog: Fix wrong order of arguments to Clamp() when assigning thread priority.
-VariationsDC.h: Fix NVidia OpenCL compilation error in DCTriangleVariation.
-FractoriumXformsColor.cpp: Checking for null pixmap pointer is not enough, must also check if the underlying buffer is null via call to QPixmap::isNull().
--Code changes
-Ember.h: Add case for FLAME_MOTION_NONE and default in ApplyFlameMotion().
-EmberMotion.h: Call base constructor.
-EmberPch.h: #pragma once only on Windows.
-EmberToXml.h:
--Handle different types of exceptions.
--Add default cases to ToString().
-Isaac.h: Remove unused variable in constructor.
-Point.h: Call base constructor in Color().
-Renderer.h/cpp:
--Add bool to Alloc() to only allocate memory for the histogram. Needed for multi-GPU.
--Make CoordMap() return a const ref, not a pointer.
-SheepTools.h:
--Use 64-bit types like the rest of the code already does.
--Fix some comment misspellings.
-Timing.h: Make BeginTime(), EndTime(), ElapsedTime() and Format() be const functions.
-Utils.h:
--Add new functions Equal() and Split().
--Handle more exception types in ReadFile().
--Get rid of most legacy blending of C and C++ argument parsing.
-XmlToEmber.h:
--Get rid of most legacy blending of C and C++ code from flam3.
--Remove some unused variables.
-EmberAnimate:
--Support multi-GPU processing that alternates full frames between devices.
--Use OpenCLInfo instead of OpenCLWrapper for --openclinfo option.
--Remove bucketT template parameter, and hard code float in its place.
--If a render fails, exit since there is no point in continuing an animation with a missing frame.
--Pass variables to threaded save better, which most likely fixes a very subtle bug that existed before.
--Remove some unused variables.
-EmberGenome, EmberRender:
--Support multi-GPU processing that alternates full frames between devices.
--Use OpenCLInfo instead of OpenCLWrapper for --openclinfo option.
--Remove bucketT template parameter, and hard code float in its place.
-EmberRender:
--Support multi-GPU processing that alternates full frames between devices.
--Use OpenCLInfo instead of OpenCLWrapper for --openclinfo option.
--Remove bucketT template parameter, and hard code float in its place.
--Only print values when not rendering with OpenCL, since they're always 0 in that case.
-EmberCLPch.h:
--#pragma once only on Windows.
--#include <atomic>.
-IterOpenCLKernelCreator.h: Add new kernel for summing two histograms. This is needed for multi-GPU.
-OpenCLWrapper.h:
--Move all OpenCL info related code into its own class OpenCLInfo.
--Add members to cache the values of global memory size and max allocation size.
-RendererCL.h/cpp:
--Redesign to accomodate multi-GPU.
--Constructor now takes a vector of devices.
--Remove DumpErrorReport() function, it's handled in the base.
--ClearBuffer(), ReadPoints(), WritePoints(), ReadHist() and WriteHist() now optionally take a device index as a parameter.
--MakeDmap() override and m_DmapCL member removed because it no longer applies since the histogram is always float since the last commit.
--Add new function SumDeviceHist() to sum histograms from two devices by first copying to a temporary on the host, then a temporary on the device, then summing.
--m_Calls member removed, as it's now per-device.
--OpenCLWrapper removed.
--m_Seeds member is now a vector of vector of seeds, to accomodate a separate and different array of seeds for each device.
--Added member m_Devices, a vector of unique_ptr of RendererCLDevice.
-EmberCommon.h
--Added Devices() function to convert from a vector of device indices to a vector of platform,device indices.
--Changed CreateRenderer() to accept a vector of devices to create a single RendererCL which will split work across multiple devices.
--Added CreateRenderers() function to accept a vector of devices to create multiple RendererCL, each which will render on a single device.
--Add more comments to some existing functions.
-EmberCommonPch.h: #pragma once only on Windows.
-EmberOptions.h:
--Remove --platform option, it's just sequential device number now with the --device option.
--Make --out be OPT_USE_RENDER instead of OPT_RENDER_ANIM since it's an error condition when animating. It makes no sense to write all frames to a single image.
--Add Devices() function to parse comma separated --device option string and return a vector of device indices.
--Make int and uint types be 64-bit, so intmax_t and size_t.
--Make better use of macros.
-JpegUtils.h: Make string parameters to WriteJpeg() and WritePng() be const ref.
-All project files: Turn off buffer security check option in Visual Studio (/Gs-)
-deployment.pri: Remove the line OTHER_FILES +=, it's pointless and was causing problems.
-Ember.pro, EmberCL.pro: Add CONFIG += plugin, otherwise it wouldn't link.
-EmberCL.pro: Add new files for multi-GPU support.
-build_all.sh: use -j4 and QMAKE=${QMAKE:/usr/bin/qmake}
-shared_settings.pri:
-Add version string.
-Remove old DESTDIR definitions.
-Add the following lines or else nothing would build:
CONFIG(release, debug|release) {
CONFIG += warn_off
DESTDIR = ../../../Bin/release
}
CONFIG(debug, debug|release) {
DESTDIR = ../../../Bin/debug
}
QMAKE_POST_LINK += $$quote(cp --update ../../../Data/flam3-palettes.xml $${DESTDIR}$$escape_expand(\n\t))
LIBS += -L/usr/lib -lpthread
-AboutDialog.ui: Another futile attempt to make it look correct on Linux.
-FinalRenderDialog.h/cpp:
--Add support for multi-GPU.
--Change from combo boxes for device selection to a table of all devices.
--Ensure device selection makes sense.
--Remove "FinalRender" prefix of various function names, it's implied given the context.
-FinalRenderEmberController.h/cpp:
--Add support for multi-GPU.
--Change m_FinishedImageCount to be atomic.
--Move CancelRender() from the base to FinalRenderEmberController<T>.
--Refactor RenderComplete() to omit any progress related functionality or image saving since it can be potentially ran in a thread.
--Consolidate setting various renderer fields into SyncGuiToRenderer().
-Fractorium.cpp: Allow for resizing of the options dialog to show the entire device table.
-FractoriumCommon.h: Add various functions to handle a table showing the available OpenCL devices on the system.
-FractoriumEmberController.h/cpp: Remove m_FinalImageIndex, it's no longer needed.
-FractoriumRender.cpp: Scale the interactive sub batch count and quality by the number of devices used.
-FractoriumSettings.h/cpp:
--Temporal samples defaults to 100 instead of 1000 which was needless overkill.
--Add multi-GPU support, remove old device,platform pair.
-FractoriumToolbar.cpp: Disable OpenCL toolbar button if there are no devices present on the system.
-FractoriumOptionsDialog.h/cpp:
--Add support for multi-GPU.
--Consolidate more assignments in DataToGui().
--Enable/disable CPU/OpenCL items in response to OpenCL checkbox event.
-Misc: Convert almost everything to size_t for unsigned, intmax_t for signed.
2015-09-12 21:33:45 -04:00
|
|
|
/// Template argument bucketT must always be float.
|
2014-07-08 03:11:14 -04:00
|
|
|
/// </summary>
|
|
|
|
|
template <typename T, typename bucketT>
|
|
|
|
|
class EMBER_API Renderer : public RendererBase
|
|
|
|
|
{
|
|
|
|
|
public:
|
2016-04-03 21:55:12 -04:00
|
|
|
|
|
|
|
|
Renderer() = default;
|
|
|
|
|
Renderer(const Renderer<T, bucketT>& renderer) = delete;
|
|
|
|
|
Renderer<T, bucketT>& operator = (const Renderer<T, bucketT>& renderer) = delete;
|
|
|
|
|
virtual ~Renderer() = default;
|
2014-07-08 03:11:14 -04:00
|
|
|
|
2014-10-14 11:53:15 -04:00
|
|
|
//Non-virtual processing functions.
|
2014-07-08 03:11:14 -04:00
|
|
|
void AddEmber(Ember<T>& ember);
|
|
|
|
|
bool AssignIterator();
|
|
|
|
|
|
2014-10-14 11:53:15 -04:00
|
|
|
//Virtual processing functions overriden from RendererBase.
|
|
|
|
|
virtual void ComputeBounds() override;
|
2015-03-21 18:27:37 -04:00
|
|
|
virtual void ComputeQuality() override;
|
2014-10-18 15:56:37 -04:00
|
|
|
virtual void ComputeCamera() override;
|
2017-05-12 18:31:48 -04:00
|
|
|
virtual void SetEmber(const Ember<T>& ember, eProcessAction action = eProcessAction::FULL_RENDER, bool prep = false) override;
|
2016-04-03 21:55:12 -04:00
|
|
|
template <typename C>
|
|
|
|
|
void SetEmber(const C& embers);
|
|
|
|
|
void MoveEmbers(vector<Ember<T>>& embers);
|
|
|
|
|
void SetExternalEmbersPointer(vector<Ember<T>>* embers);
|
2014-10-14 11:53:15 -04:00
|
|
|
virtual bool CreateDEFilter(bool& newAlloc) override;
|
|
|
|
|
virtual bool CreateSpatialFilter(bool& newAlloc) override;
|
|
|
|
|
virtual bool CreateTemporalFilter(bool& newAlloc) override;
|
2015-03-21 18:27:37 -04:00
|
|
|
virtual size_t HistBucketSize() const override { return sizeof(tvec4<bucketT, glm::defaultp>); }
|
2014-12-06 00:05:09 -05:00
|
|
|
virtual eRenderStatus Run(vector<byte>& finalImage, double time = 0, size_t subBatchCountOverride = 0, bool forceOutput = false, size_t finalOffset = 0) override;
|
2016-04-03 21:55:12 -04:00
|
|
|
virtual EmberImageComments ImageComments(const EmberStats& stats, size_t printEditDepth = 0, bool hexPalette = true) override;
|
2014-07-08 03:11:14 -04:00
|
|
|
|
|
|
|
|
protected:
|
2014-10-14 11:53:15 -04:00
|
|
|
//New virtual functions to be overridden in derived renderers that use the GPU, but not accessed outside.
|
2014-07-08 03:11:14 -04:00
|
|
|
virtual void MakeDmap(T colorScalar);
|
--User changes
-Add support for multiple GPU devices.
--These options are present in the command line and in Fractorium.
-Change scheme of specifying devices from platform,device to just total device index.
--Single number on the command line.
--Change from combo boxes for device selection to a table of all devices in Fractorium.
-Temporal samples defaults to 100 instead of 1000 which was needless overkill.
--Bug fixes
-EmberAnimate, EmberRender, FractoriumSettings, FinalRenderDialog: Fix wrong order of arguments to Clamp() when assigning thread priority.
-VariationsDC.h: Fix NVidia OpenCL compilation error in DCTriangleVariation.
-FractoriumXformsColor.cpp: Checking for null pixmap pointer is not enough, must also check if the underlying buffer is null via call to QPixmap::isNull().
--Code changes
-Ember.h: Add case for FLAME_MOTION_NONE and default in ApplyFlameMotion().
-EmberMotion.h: Call base constructor.
-EmberPch.h: #pragma once only on Windows.
-EmberToXml.h:
--Handle different types of exceptions.
--Add default cases to ToString().
-Isaac.h: Remove unused variable in constructor.
-Point.h: Call base constructor in Color().
-Renderer.h/cpp:
--Add bool to Alloc() to only allocate memory for the histogram. Needed for multi-GPU.
--Make CoordMap() return a const ref, not a pointer.
-SheepTools.h:
--Use 64-bit types like the rest of the code already does.
--Fix some comment misspellings.
-Timing.h: Make BeginTime(), EndTime(), ElapsedTime() and Format() be const functions.
-Utils.h:
--Add new functions Equal() and Split().
--Handle more exception types in ReadFile().
--Get rid of most legacy blending of C and C++ argument parsing.
-XmlToEmber.h:
--Get rid of most legacy blending of C and C++ code from flam3.
--Remove some unused variables.
-EmberAnimate:
--Support multi-GPU processing that alternates full frames between devices.
--Use OpenCLInfo instead of OpenCLWrapper for --openclinfo option.
--Remove bucketT template parameter, and hard code float in its place.
--If a render fails, exit since there is no point in continuing an animation with a missing frame.
--Pass variables to threaded save better, which most likely fixes a very subtle bug that existed before.
--Remove some unused variables.
-EmberGenome, EmberRender:
--Support multi-GPU processing that alternates full frames between devices.
--Use OpenCLInfo instead of OpenCLWrapper for --openclinfo option.
--Remove bucketT template parameter, and hard code float in its place.
-EmberRender:
--Support multi-GPU processing that alternates full frames between devices.
--Use OpenCLInfo instead of OpenCLWrapper for --openclinfo option.
--Remove bucketT template parameter, and hard code float in its place.
--Only print values when not rendering with OpenCL, since they're always 0 in that case.
-EmberCLPch.h:
--#pragma once only on Windows.
--#include <atomic>.
-IterOpenCLKernelCreator.h: Add new kernel for summing two histograms. This is needed for multi-GPU.
-OpenCLWrapper.h:
--Move all OpenCL info related code into its own class OpenCLInfo.
--Add members to cache the values of global memory size and max allocation size.
-RendererCL.h/cpp:
--Redesign to accomodate multi-GPU.
--Constructor now takes a vector of devices.
--Remove DumpErrorReport() function, it's handled in the base.
--ClearBuffer(), ReadPoints(), WritePoints(), ReadHist() and WriteHist() now optionally take a device index as a parameter.
--MakeDmap() override and m_DmapCL member removed because it no longer applies since the histogram is always float since the last commit.
--Add new function SumDeviceHist() to sum histograms from two devices by first copying to a temporary on the host, then a temporary on the device, then summing.
--m_Calls member removed, as it's now per-device.
--OpenCLWrapper removed.
--m_Seeds member is now a vector of vector of seeds, to accomodate a separate and different array of seeds for each device.
--Added member m_Devices, a vector of unique_ptr of RendererCLDevice.
-EmberCommon.h
--Added Devices() function to convert from a vector of device indices to a vector of platform,device indices.
--Changed CreateRenderer() to accept a vector of devices to create a single RendererCL which will split work across multiple devices.
--Added CreateRenderers() function to accept a vector of devices to create multiple RendererCL, each which will render on a single device.
--Add more comments to some existing functions.
-EmberCommonPch.h: #pragma once only on Windows.
-EmberOptions.h:
--Remove --platform option, it's just sequential device number now with the --device option.
--Make --out be OPT_USE_RENDER instead of OPT_RENDER_ANIM since it's an error condition when animating. It makes no sense to write all frames to a single image.
--Add Devices() function to parse comma separated --device option string and return a vector of device indices.
--Make int and uint types be 64-bit, so intmax_t and size_t.
--Make better use of macros.
-JpegUtils.h: Make string parameters to WriteJpeg() and WritePng() be const ref.
-All project files: Turn off buffer security check option in Visual Studio (/Gs-)
-deployment.pri: Remove the line OTHER_FILES +=, it's pointless and was causing problems.
-Ember.pro, EmberCL.pro: Add CONFIG += plugin, otherwise it wouldn't link.
-EmberCL.pro: Add new files for multi-GPU support.
-build_all.sh: use -j4 and QMAKE=${QMAKE:/usr/bin/qmake}
-shared_settings.pri:
-Add version string.
-Remove old DESTDIR definitions.
-Add the following lines or else nothing would build:
CONFIG(release, debug|release) {
CONFIG += warn_off
DESTDIR = ../../../Bin/release
}
CONFIG(debug, debug|release) {
DESTDIR = ../../../Bin/debug
}
QMAKE_POST_LINK += $$quote(cp --update ../../../Data/flam3-palettes.xml $${DESTDIR}$$escape_expand(\n\t))
LIBS += -L/usr/lib -lpthread
-AboutDialog.ui: Another futile attempt to make it look correct on Linux.
-FinalRenderDialog.h/cpp:
--Add support for multi-GPU.
--Change from combo boxes for device selection to a table of all devices.
--Ensure device selection makes sense.
--Remove "FinalRender" prefix of various function names, it's implied given the context.
-FinalRenderEmberController.h/cpp:
--Add support for multi-GPU.
--Change m_FinishedImageCount to be atomic.
--Move CancelRender() from the base to FinalRenderEmberController<T>.
--Refactor RenderComplete() to omit any progress related functionality or image saving since it can be potentially ran in a thread.
--Consolidate setting various renderer fields into SyncGuiToRenderer().
-Fractorium.cpp: Allow for resizing of the options dialog to show the entire device table.
-FractoriumCommon.h: Add various functions to handle a table showing the available OpenCL devices on the system.
-FractoriumEmberController.h/cpp: Remove m_FinalImageIndex, it's no longer needed.
-FractoriumRender.cpp: Scale the interactive sub batch count and quality by the number of devices used.
-FractoriumSettings.h/cpp:
--Temporal samples defaults to 100 instead of 1000 which was needless overkill.
--Add multi-GPU support, remove old device,platform pair.
-FractoriumToolbar.cpp: Disable OpenCL toolbar button if there are no devices present on the system.
-FractoriumOptionsDialog.h/cpp:
--Add support for multi-GPU.
--Consolidate more assignments in DataToGui().
--Enable/disable CPU/OpenCL items in response to OpenCL checkbox event.
-Misc: Convert almost everything to size_t for unsigned, intmax_t for signed.
2015-09-12 21:33:45 -04:00
|
|
|
virtual bool Alloc(bool histOnly = false);
|
2014-07-08 03:11:14 -04:00
|
|
|
virtual bool ResetBuckets(bool resetHist = true, bool resetAccum = true);
|
--User changes
-Add new variations: bubbleT3D, crob, hexaplay3D, hexcrop, hexes, hexnix3D, loonie2, loonie3, nBlur, octapol and synth.
-Allow for pre/post versions of dc_bubble, dc_cylinder and dc_linear whereas before they were omitted.
-When saving a file with multiple embers in it, detect if time values are all the same and if so, start them at zero and increment by 1 for each ember.
-Allow for numerous quality increases to be coalesced into one. It will pick up at the end of the current render.
-Show selection highlight on variations tree in response to mouse hover. This makes it easier to see for which variation or param the current mouse wheel action will apply.
-Make default temporal samples be 100, whereas before it was 1000 which was overkill.
-Require the shift key to be held with delete for deleting an ember to prevent it from triggering when the user enters delete in the edit box.
-This wasn't otherwise fixable without writing a lot more code.
--Bug fixes
-EmberGenome was crashing when generating a sequence from a source file with more than 2 embers in it.
-EmberGenome was improperly handling the first frame of a merge after the last frame of the loop.
-These bugs were due to a previous commit. Revert parts of that commit.
-Prevent a zoom value of less than 0 when reading from xml.
-Slight optimization of the crescents, and mask variations, if the compiler wasn't doing it already.
-Unique file naming was broken because it was looking for _# and the default names ended with -#.
-Disallow renaming of an ember in the library tree to an empty string.
-Severe bug that prevented some variations from being read correctly from params generated outside this program.
-Severe OpenCL randomization bug. The first x coordinates of the first points in the first kernel call of the first ember of a render since the OpenCL renderer object was created were not random and were mostly -1.
-Severe bug when populating xform selection distributions that could sometimes cause a crash due to roundoff error. Fix by using double.
-Limit the max number of variations in a random ember to MAX_CL_VARS, which is 8. This ensures they'll look the same on CPU and GPU.
-Prevent user from saving stylesheet to default.qss, it's a special reserved filename.
--Code changes
-Generalize using the running sum output point inside of a variation for all cases: pre, reg and post.
-Allow for array variables in variations where the address of each element is stored in m_Params.
-Qualify all math functions with std::
-No longer use our own Clamp() in OpenCL, instead use the standard clamp().
-Redesign how functions are used in the variations OpenCL code.
-Add tests to EmberTester to verify some of the new functionality.
-Place more const and override qualifiers on functions where appropriate.
-Add a global rand with a lock to be used very sparingly.
-Use a map instead of a vector for bad param names in Xml parsing.
-Prefix affine interpolation mode defines with "AFFINE_" to make their purpose more clear.
-Allow for variations that change state during iteration by sending a separate copy of the ember to each rendering thread.
-Implement this same functionality with a local struct in OpenCL. It's members are the total of all variables that need to change state within an ember.
-Add Contains() function to Utils.h.
-EmberRender: print names of kernels being printed with --dump_kernel option.
-Clean up EmberTester to handle some of the recent changes.
-Fix various casts.
-Replace % 2 with & 1, even though the compiler was likely doing this already.
-Add new file Variations06.h to accommodate new variations.
-General cleanup.
2015-11-22 17:15:07 -05:00
|
|
|
virtual eRenderStatus LogScaleDensityFilter(bool forceOutput = false);
|
2014-07-08 03:11:14 -04:00
|
|
|
virtual eRenderStatus GaussianDensityFilter();
|
2014-12-06 00:05:09 -05:00
|
|
|
virtual eRenderStatus AccumulatorToFinalImage(vector<byte>& pixels, size_t finalOffset);
|
|
|
|
|
virtual eRenderStatus AccumulatorToFinalImage(byte* pixels, size_t finalOffset);
|
2014-11-03 02:16:34 -05:00
|
|
|
virtual EmberStats Iterate(size_t iterCount, size_t temporalSample);
|
2014-07-08 03:11:14 -04:00
|
|
|
|
|
|
|
|
public:
|
2014-10-14 11:53:15 -04:00
|
|
|
//Non-virtual render properties, getters and setters.
|
2014-07-08 03:11:14 -04:00
|
|
|
inline T PixelAspectRatio() const;
|
|
|
|
|
void PixelAspectRatio(T pixelAspectRatio);
|
|
|
|
|
|
2014-10-14 11:53:15 -04:00
|
|
|
//Non-virtual renderer properties, getters only.
|
2015-03-21 18:27:37 -04:00
|
|
|
inline T Scale() const;
|
|
|
|
|
inline T PixelsPerUnitX() const;
|
|
|
|
|
inline T PixelsPerUnitY() const;
|
2015-08-10 23:10:23 -04:00
|
|
|
inline bucketT K1() const;
|
|
|
|
|
inline bucketT K2() const;
|
--User changes
-Add support for multiple GPU devices.
--These options are present in the command line and in Fractorium.
-Change scheme of specifying devices from platform,device to just total device index.
--Single number on the command line.
--Change from combo boxes for device selection to a table of all devices in Fractorium.
-Temporal samples defaults to 100 instead of 1000 which was needless overkill.
--Bug fixes
-EmberAnimate, EmberRender, FractoriumSettings, FinalRenderDialog: Fix wrong order of arguments to Clamp() when assigning thread priority.
-VariationsDC.h: Fix NVidia OpenCL compilation error in DCTriangleVariation.
-FractoriumXformsColor.cpp: Checking for null pixmap pointer is not enough, must also check if the underlying buffer is null via call to QPixmap::isNull().
--Code changes
-Ember.h: Add case for FLAME_MOTION_NONE and default in ApplyFlameMotion().
-EmberMotion.h: Call base constructor.
-EmberPch.h: #pragma once only on Windows.
-EmberToXml.h:
--Handle different types of exceptions.
--Add default cases to ToString().
-Isaac.h: Remove unused variable in constructor.
-Point.h: Call base constructor in Color().
-Renderer.h/cpp:
--Add bool to Alloc() to only allocate memory for the histogram. Needed for multi-GPU.
--Make CoordMap() return a const ref, not a pointer.
-SheepTools.h:
--Use 64-bit types like the rest of the code already does.
--Fix some comment misspellings.
-Timing.h: Make BeginTime(), EndTime(), ElapsedTime() and Format() be const functions.
-Utils.h:
--Add new functions Equal() and Split().
--Handle more exception types in ReadFile().
--Get rid of most legacy blending of C and C++ argument parsing.
-XmlToEmber.h:
--Get rid of most legacy blending of C and C++ code from flam3.
--Remove some unused variables.
-EmberAnimate:
--Support multi-GPU processing that alternates full frames between devices.
--Use OpenCLInfo instead of OpenCLWrapper for --openclinfo option.
--Remove bucketT template parameter, and hard code float in its place.
--If a render fails, exit since there is no point in continuing an animation with a missing frame.
--Pass variables to threaded save better, which most likely fixes a very subtle bug that existed before.
--Remove some unused variables.
-EmberGenome, EmberRender:
--Support multi-GPU processing that alternates full frames between devices.
--Use OpenCLInfo instead of OpenCLWrapper for --openclinfo option.
--Remove bucketT template parameter, and hard code float in its place.
-EmberRender:
--Support multi-GPU processing that alternates full frames between devices.
--Use OpenCLInfo instead of OpenCLWrapper for --openclinfo option.
--Remove bucketT template parameter, and hard code float in its place.
--Only print values when not rendering with OpenCL, since they're always 0 in that case.
-EmberCLPch.h:
--#pragma once only on Windows.
--#include <atomic>.
-IterOpenCLKernelCreator.h: Add new kernel for summing two histograms. This is needed for multi-GPU.
-OpenCLWrapper.h:
--Move all OpenCL info related code into its own class OpenCLInfo.
--Add members to cache the values of global memory size and max allocation size.
-RendererCL.h/cpp:
--Redesign to accomodate multi-GPU.
--Constructor now takes a vector of devices.
--Remove DumpErrorReport() function, it's handled in the base.
--ClearBuffer(), ReadPoints(), WritePoints(), ReadHist() and WriteHist() now optionally take a device index as a parameter.
--MakeDmap() override and m_DmapCL member removed because it no longer applies since the histogram is always float since the last commit.
--Add new function SumDeviceHist() to sum histograms from two devices by first copying to a temporary on the host, then a temporary on the device, then summing.
--m_Calls member removed, as it's now per-device.
--OpenCLWrapper removed.
--m_Seeds member is now a vector of vector of seeds, to accomodate a separate and different array of seeds for each device.
--Added member m_Devices, a vector of unique_ptr of RendererCLDevice.
-EmberCommon.h
--Added Devices() function to convert from a vector of device indices to a vector of platform,device indices.
--Changed CreateRenderer() to accept a vector of devices to create a single RendererCL which will split work across multiple devices.
--Added CreateRenderers() function to accept a vector of devices to create multiple RendererCL, each which will render on a single device.
--Add more comments to some existing functions.
-EmberCommonPch.h: #pragma once only on Windows.
-EmberOptions.h:
--Remove --platform option, it's just sequential device number now with the --device option.
--Make --out be OPT_USE_RENDER instead of OPT_RENDER_ANIM since it's an error condition when animating. It makes no sense to write all frames to a single image.
--Add Devices() function to parse comma separated --device option string and return a vector of device indices.
--Make int and uint types be 64-bit, so intmax_t and size_t.
--Make better use of macros.
-JpegUtils.h: Make string parameters to WriteJpeg() and WritePng() be const ref.
-All project files: Turn off buffer security check option in Visual Studio (/Gs-)
-deployment.pri: Remove the line OTHER_FILES +=, it's pointless and was causing problems.
-Ember.pro, EmberCL.pro: Add CONFIG += plugin, otherwise it wouldn't link.
-EmberCL.pro: Add new files for multi-GPU support.
-build_all.sh: use -j4 and QMAKE=${QMAKE:/usr/bin/qmake}
-shared_settings.pri:
-Add version string.
-Remove old DESTDIR definitions.
-Add the following lines or else nothing would build:
CONFIG(release, debug|release) {
CONFIG += warn_off
DESTDIR = ../../../Bin/release
}
CONFIG(debug, debug|release) {
DESTDIR = ../../../Bin/debug
}
QMAKE_POST_LINK += $$quote(cp --update ../../../Data/flam3-palettes.xml $${DESTDIR}$$escape_expand(\n\t))
LIBS += -L/usr/lib -lpthread
-AboutDialog.ui: Another futile attempt to make it look correct on Linux.
-FinalRenderDialog.h/cpp:
--Add support for multi-GPU.
--Change from combo boxes for device selection to a table of all devices.
--Ensure device selection makes sense.
--Remove "FinalRender" prefix of various function names, it's implied given the context.
-FinalRenderEmberController.h/cpp:
--Add support for multi-GPU.
--Change m_FinishedImageCount to be atomic.
--Move CancelRender() from the base to FinalRenderEmberController<T>.
--Refactor RenderComplete() to omit any progress related functionality or image saving since it can be potentially ran in a thread.
--Consolidate setting various renderer fields into SyncGuiToRenderer().
-Fractorium.cpp: Allow for resizing of the options dialog to show the entire device table.
-FractoriumCommon.h: Add various functions to handle a table showing the available OpenCL devices on the system.
-FractoriumEmberController.h/cpp: Remove m_FinalImageIndex, it's no longer needed.
-FractoriumRender.cpp: Scale the interactive sub batch count and quality by the number of devices used.
-FractoriumSettings.h/cpp:
--Temporal samples defaults to 100 instead of 1000 which was needless overkill.
--Add multi-GPU support, remove old device,platform pair.
-FractoriumToolbar.cpp: Disable OpenCL toolbar button if there are no devices present on the system.
-FractoriumOptionsDialog.h/cpp:
--Add support for multi-GPU.
--Consolidate more assignments in DataToGui().
--Enable/disable CPU/OpenCL items in response to OpenCL checkbox event.
-Misc: Convert almost everything to size_t for unsigned, intmax_t for signed.
2015-09-12 21:33:45 -04:00
|
|
|
inline const CarToRas<T>& CoordMap() const;
|
2015-03-21 18:27:37 -04:00
|
|
|
inline tvec4<bucketT, glm::defaultp>* HistBuckets();
|
|
|
|
|
inline tvec4<bucketT, glm::defaultp>* AccumulatorBuckets();
|
2015-08-10 23:10:23 -04:00
|
|
|
inline SpatialFilter<bucketT>* GetSpatialFilter();
|
2015-03-21 18:27:37 -04:00
|
|
|
inline TemporalFilter<T>* GetTemporalFilter();
|
2014-09-01 01:03:45 -04:00
|
|
|
|
2014-10-14 11:53:15 -04:00
|
|
|
//Virtual renderer properties overridden from RendererBase, getters only.
|
|
|
|
|
virtual double ScaledQuality() const override;
|
|
|
|
|
virtual double LowerLeftX(bool gutter = true) const override;
|
|
|
|
|
virtual double LowerLeftY(bool gutter = true) const override;
|
|
|
|
|
virtual double UpperRightX(bool gutter = true) const override;
|
|
|
|
|
virtual double UpperRightY(bool gutter = true) const override;
|
|
|
|
|
virtual DensityFilterBase* GetDensityFilter() override;
|
|
|
|
|
|
|
|
|
|
//Non-virtual ember wrappers, getters only.
|
--User changes
-Add a palette editor.
-Add support for reading .ugr/.gradient/.gradients palette files.
-Allow toggling on spinners whose minimum value is not zero.
-Allow toggling display of image, affines and grid.
-Add new variations: cylinder2, circlesplit, tile_log, truchet_fill, waves2_radial.
--Bug fixes
-cpow2 was wrong.
-Palettes with rapid changes in color would produce slightly different outputs from Apo/Chaotica. This was due to a long standing bug from flam3.
-Use exec() on Apple and show() on all other OSes for dialog boxes.
-Trying to render a sequence with no frames would crash.
-Selecting multiple xforms and rotating them would produce the wrong rotation.
-Better handling when parsing flames using different encoding, such as unicode and UTF-8.
-Switching between SP/DP didn't reselect the selected flame in the Library tab.
--Code changes
-Make all types concerning palettes be floats, including PaletteTableWidgetItem.
-PaletteTableWidgetItem is no longer templated because all palettes are float.
-Include the source colors for user created gradients.
-Change parallel_for() calls to work with very old versions of TBB which are lingering on some systems.
-Split conditional out of accumulation loop on the CPU for better performance.
-Vectorize summing when doing density filter for better performance.
-Make all usage of palettes be of type float, double is pointless.
-Allow palettes to reside in multiple folders, while ensuring only one of each name is added.
-Refactor some palette path searching code.
-Make ReadFile() throw and catch an exception if the file operation fails.
-A little extra safety in foci and foci3D with a call to Zeps().
-Cast to (real_t) in the OpenCL string for the w variation, which was having trouble compiling on Mac.
-Fixing missing comma between paths in InitPaletteList().
-Move Xml and PaletteList classes into cpp to shorten build times when working on them.
-Remove default param values for IterOpenCLKernelCreator<T>::SharedDataIndexDefines in cpp file.
-Change more NULL to nullptr.
2017-02-26 03:02:21 -05:00
|
|
|
inline bool XaosPresent() const;
|
|
|
|
|
inline size_t Supersample() const;
|
|
|
|
|
inline size_t PaletteIndex() const;
|
|
|
|
|
inline T Time() const;
|
|
|
|
|
inline T Quality() const;
|
|
|
|
|
inline T SpatialFilterRadius() const;
|
|
|
|
|
inline T PixelsPerUnit() const;
|
|
|
|
|
inline T Zoom() const;
|
|
|
|
|
inline T CenterX() const;
|
|
|
|
|
inline T CenterY() const;
|
|
|
|
|
inline T Rotate() const;
|
|
|
|
|
inline T Hue() const;
|
|
|
|
|
inline bucketT Brightness() const;
|
|
|
|
|
inline bucketT Contrast() const;
|
|
|
|
|
inline bucketT Gamma() const;
|
|
|
|
|
inline bucketT Vibrancy() const;
|
|
|
|
|
inline bucketT GammaThresh() const;
|
|
|
|
|
inline bucketT HighlightPower() const;
|
|
|
|
|
inline Color<T> Background() const;
|
|
|
|
|
inline const Xform<T>* Xforms() const;
|
|
|
|
|
inline Xform<T>* NonConstXforms();
|
|
|
|
|
inline size_t XformCount() const;
|
|
|
|
|
inline const Xform<T>* FinalXform() const;
|
|
|
|
|
inline Xform<T>* NonConstFinalXform();
|
|
|
|
|
inline bool UseFinalXform() const;
|
|
|
|
|
inline const Palette<float>* GetPalette() const;
|
|
|
|
|
inline ePaletteMode PaletteMode() const;
|
2014-10-14 11:53:15 -04:00
|
|
|
|
|
|
|
|
//Virtual ember wrappers overridden from RendererBase, getters only.
|
|
|
|
|
virtual size_t TemporalSamples() const override;
|
|
|
|
|
virtual size_t FinalRasW() const override;
|
|
|
|
|
virtual size_t FinalRasH() const override;
|
2014-11-28 04:37:51 -05:00
|
|
|
virtual size_t SubBatchSize() const override;
|
|
|
|
|
virtual size_t FuseCount() const override;
|
2014-10-14 11:53:15 -04:00
|
|
|
|
|
|
|
|
//Non-virtual iterator wrappers.
|
2014-12-07 02:51:44 -05:00
|
|
|
const byte* XformDistributions() const;
|
|
|
|
|
size_t XformDistributionsSize() const;
|
|
|
|
|
Point<T>* Samples(size_t threadIndex) const;
|
2014-07-08 03:11:14 -04:00
|
|
|
|
|
|
|
|
protected:
|
|
|
|
|
//Non-virtual functions that might be needed by a derived class.
|
2015-08-10 23:10:23 -04:00
|
|
|
void PrepFinalAccumVals(Color<bucketT>& background, bucketT& g, bucketT& linRange, bucketT& vibrancy);
|
2014-07-08 03:11:14 -04:00
|
|
|
|
2015-12-31 19:00:36 -05:00
|
|
|
private:
|
2014-10-14 11:53:15 -04:00
|
|
|
//Miscellaneous non-virtual functions used only in this class.
|
2014-11-28 04:37:51 -05:00
|
|
|
void Accumulate(QTIsaac<ISAAC_SIZE, ISAAC_INT>& rand, Point<T>* samples, size_t sampleCount, const Palette<bucketT>* palette);
|
2015-03-21 18:27:37 -04:00
|
|
|
/*inline*/ void AddToAccum(const tvec4<bucketT, glm::defaultp>& bucket, intmax_t i, intmax_t ii, intmax_t j, intmax_t jj);
|
2015-08-10 23:10:23 -04:00
|
|
|
template <typename accumT> void GammaCorrection(tvec4<bucketT, glm::defaultp>& bucket, Color<bucketT>& background, bucketT g, bucketT linRange, bucketT vibrancy, bool doAlpha, bool scale, accumT* correctedChannels);
|
|
|
|
|
void CurveAdjust(bucketT& a, const glm::length_t& index);
|
--User changes
-Add new variations: bubbleT3D, crob, hexaplay3D, hexcrop, hexes, hexnix3D, loonie2, loonie3, nBlur, octapol and synth.
-Allow for pre/post versions of dc_bubble, dc_cylinder and dc_linear whereas before they were omitted.
-When saving a file with multiple embers in it, detect if time values are all the same and if so, start them at zero and increment by 1 for each ember.
-Allow for numerous quality increases to be coalesced into one. It will pick up at the end of the current render.
-Show selection highlight on variations tree in response to mouse hover. This makes it easier to see for which variation or param the current mouse wheel action will apply.
-Make default temporal samples be 100, whereas before it was 1000 which was overkill.
-Require the shift key to be held with delete for deleting an ember to prevent it from triggering when the user enters delete in the edit box.
-This wasn't otherwise fixable without writing a lot more code.
--Bug fixes
-EmberGenome was crashing when generating a sequence from a source file with more than 2 embers in it.
-EmberGenome was improperly handling the first frame of a merge after the last frame of the loop.
-These bugs were due to a previous commit. Revert parts of that commit.
-Prevent a zoom value of less than 0 when reading from xml.
-Slight optimization of the crescents, and mask variations, if the compiler wasn't doing it already.
-Unique file naming was broken because it was looking for _# and the default names ended with -#.
-Disallow renaming of an ember in the library tree to an empty string.
-Severe bug that prevented some variations from being read correctly from params generated outside this program.
-Severe OpenCL randomization bug. The first x coordinates of the first points in the first kernel call of the first ember of a render since the OpenCL renderer object was created were not random and were mostly -1.
-Severe bug when populating xform selection distributions that could sometimes cause a crash due to roundoff error. Fix by using double.
-Limit the max number of variations in a random ember to MAX_CL_VARS, which is 8. This ensures they'll look the same on CPU and GPU.
-Prevent user from saving stylesheet to default.qss, it's a special reserved filename.
--Code changes
-Generalize using the running sum output point inside of a variation for all cases: pre, reg and post.
-Allow for array variables in variations where the address of each element is stored in m_Params.
-Qualify all math functions with std::
-No longer use our own Clamp() in OpenCL, instead use the standard clamp().
-Redesign how functions are used in the variations OpenCL code.
-Add tests to EmberTester to verify some of the new functionality.
-Place more const and override qualifiers on functions where appropriate.
-Add a global rand with a lock to be used very sparingly.
-Use a map instead of a vector for bad param names in Xml parsing.
-Prefix affine interpolation mode defines with "AFFINE_" to make their purpose more clear.
-Allow for variations that change state during iteration by sending a separate copy of the ember to each rendering thread.
-Implement this same functionality with a local struct in OpenCL. It's members are the total of all variables that need to change state within an ember.
-Add Contains() function to Utils.h.
-EmberRender: print names of kernels being printed with --dump_kernel option.
-Clean up EmberTester to handle some of the recent changes.
-Fix various casts.
-Replace % 2 with & 1, even though the compiler was likely doing this already.
-Add new file Variations06.h to accommodate new variations.
-General cleanup.
2015-11-22 17:15:07 -05:00
|
|
|
void VectorizedLogScale(size_t row, size_t rowEnd);
|
2014-07-08 03:11:14 -04:00
|
|
|
|
|
|
|
|
protected:
|
|
|
|
|
T m_Scale;
|
|
|
|
|
T m_PixelsPerUnitX;
|
|
|
|
|
T m_PixelsPerUnitY;
|
2016-04-03 21:55:12 -04:00
|
|
|
T m_PixelAspectRatio = 1;
|
2014-07-08 03:11:14 -04:00
|
|
|
T m_LowerLeftX;
|
|
|
|
|
T m_LowerLeftY;
|
|
|
|
|
T m_UpperRightX;
|
|
|
|
|
T m_UpperRightY;
|
2015-08-10 23:10:23 -04:00
|
|
|
bucketT m_K1;
|
|
|
|
|
bucketT m_K2;
|
|
|
|
|
bucketT m_Vibrancy;//Accumulate these after each temporal sample.
|
|
|
|
|
bucketT m_Gamma;
|
2014-10-14 11:53:15 -04:00
|
|
|
T m_ScaledQuality;
|
2015-08-10 23:10:23 -04:00
|
|
|
Color<bucketT> m_Background;//This is a scaled copy of the m_Background member of m_Ember, but with a type of bucketT.
|
2014-07-08 03:11:14 -04:00
|
|
|
Affine2D<T> m_RotMat;
|
|
|
|
|
Ember<T> m_Ember;
|
|
|
|
|
Ember<T> m_TempEmber;
|
|
|
|
|
Ember<T> m_LastEmber;
|
2016-04-03 21:55:12 -04:00
|
|
|
private:
|
2014-07-08 03:11:14 -04:00
|
|
|
vector<Ember<T>> m_Embers;
|
2016-04-03 21:55:12 -04:00
|
|
|
|
|
|
|
|
protected:
|
|
|
|
|
vector<Ember<T>>* m_EmbersP = &m_Embers;
|
--User changes
-Add new variations: bubbleT3D, crob, hexaplay3D, hexcrop, hexes, hexnix3D, loonie2, loonie3, nBlur, octapol and synth.
-Allow for pre/post versions of dc_bubble, dc_cylinder and dc_linear whereas before they were omitted.
-When saving a file with multiple embers in it, detect if time values are all the same and if so, start them at zero and increment by 1 for each ember.
-Allow for numerous quality increases to be coalesced into one. It will pick up at the end of the current render.
-Show selection highlight on variations tree in response to mouse hover. This makes it easier to see for which variation or param the current mouse wheel action will apply.
-Make default temporal samples be 100, whereas before it was 1000 which was overkill.
-Require the shift key to be held with delete for deleting an ember to prevent it from triggering when the user enters delete in the edit box.
-This wasn't otherwise fixable without writing a lot more code.
--Bug fixes
-EmberGenome was crashing when generating a sequence from a source file with more than 2 embers in it.
-EmberGenome was improperly handling the first frame of a merge after the last frame of the loop.
-These bugs were due to a previous commit. Revert parts of that commit.
-Prevent a zoom value of less than 0 when reading from xml.
-Slight optimization of the crescents, and mask variations, if the compiler wasn't doing it already.
-Unique file naming was broken because it was looking for _# and the default names ended with -#.
-Disallow renaming of an ember in the library tree to an empty string.
-Severe bug that prevented some variations from being read correctly from params generated outside this program.
-Severe OpenCL randomization bug. The first x coordinates of the first points in the first kernel call of the first ember of a render since the OpenCL renderer object was created were not random and were mostly -1.
-Severe bug when populating xform selection distributions that could sometimes cause a crash due to roundoff error. Fix by using double.
-Limit the max number of variations in a random ember to MAX_CL_VARS, which is 8. This ensures they'll look the same on CPU and GPU.
-Prevent user from saving stylesheet to default.qss, it's a special reserved filename.
--Code changes
-Generalize using the running sum output point inside of a variation for all cases: pre, reg and post.
-Allow for array variables in variations where the address of each element is stored in m_Params.
-Qualify all math functions with std::
-No longer use our own Clamp() in OpenCL, instead use the standard clamp().
-Redesign how functions are used in the variations OpenCL code.
-Add tests to EmberTester to verify some of the new functionality.
-Place more const and override qualifiers on functions where appropriate.
-Add a global rand with a lock to be used very sparingly.
-Use a map instead of a vector for bad param names in Xml parsing.
-Prefix affine interpolation mode defines with "AFFINE_" to make their purpose more clear.
-Allow for variations that change state during iteration by sending a separate copy of the ember to each rendering thread.
-Implement this same functionality with a local struct in OpenCL. It's members are the total of all variables that need to change state within an ember.
-Add Contains() function to Utils.h.
-EmberRender: print names of kernels being printed with --dump_kernel option.
-Clean up EmberTester to handle some of the recent changes.
-Fix various casts.
-Replace % 2 with & 1, even though the compiler was likely doing this already.
-Add new file Variations06.h to accommodate new variations.
-General cleanup.
2015-11-22 17:15:07 -05:00
|
|
|
vector<Ember<T>> m_ThreadEmbers;
|
2016-04-23 12:02:09 -04:00
|
|
|
Interpolater<T> m_Interpolater;
|
2014-07-08 03:11:14 -04:00
|
|
|
CarToRas<T> m_CarToRas;
|
2016-04-03 21:55:12 -04:00
|
|
|
unique_ptr<StandardIterator<T>> m_StandardIterator = make_unique<StandardIterator<T>>();
|
|
|
|
|
unique_ptr<XaosIterator<T>> m_XaosIterator = make_unique<XaosIterator<T>>();
|
|
|
|
|
Iterator<T>* m_Iterator = m_StandardIterator.get();
|
2015-03-21 18:27:37 -04:00
|
|
|
Palette<bucketT> m_Dmap, m_Csa;
|
|
|
|
|
vector<tvec4<bucketT, glm::defaultp>> m_HistBuckets;
|
|
|
|
|
vector<tvec4<bucketT, glm::defaultp>> m_AccumulatorBuckets;
|
2015-08-10 23:10:23 -04:00
|
|
|
unique_ptr<SpatialFilter<bucketT>> m_SpatialFilter;
|
2014-09-10 01:41:26 -04:00
|
|
|
unique_ptr<TemporalFilter<T>> m_TemporalFilter;
|
2015-08-10 23:10:23 -04:00
|
|
|
unique_ptr<DensityFilter<bucketT>> m_DensityFilter;
|
2014-07-08 03:11:14 -04:00
|
|
|
vector<vector<Point<T>>> m_Samples;
|
|
|
|
|
EmberToXml<T> m_EmberToXml;
|
|
|
|
|
};
|
|
|
|
|
|
|
|
|
|
//This class had to be implemented in a cpp file because the compiler was breaking.
|
|
|
|
|
//So the explicit instantiation must be declared here rather than in Ember.cpp where
|
|
|
|
|
//all of the other classes are done.
|
2014-09-01 01:03:45 -04:00
|
|
|
}
|