mirror of
https://bitbucket.org/mfeemster/fractorium.git
synced 2025-01-22 05:30:06 -05:00
47b6614c8a
-Fix improper usage of rand() in cpow2, cpow3, hypertile1, hypertile3D1, hypertile3D2, juliac, juliaq. -Fix program crashing during density filtering on some Nvidia cards. -hypertile3D1 was wrong. -Parsing phoenix_julia when coming from Apophysis was wrong. -Density filtering was freezing on certain Nvidia cards. --Code changes -Optimize juliac, npolar. -Add a new function Crand() which behaves like the legacy C rand() which returns an integer between 0 and 32766, inclusive. -Use RandBit() in some places. -Remove Zeps() from vignette, it's not needed. -Restructure OpenCL code for density filtering such that it does not hang after being compiled on some Nvidia cards, such as the gtx 1660. Remove barriers from conditionals where possible.
77 lines
2.5 KiB
C++
77 lines
2.5 KiB
C++
#pragma once
|
|
|
|
#include "EmberCLPch.h"
|
|
#include "EmberCLStructs.h"
|
|
#include "EmberCLFunctions.h"
|
|
|
|
/// <summary>
|
|
/// DEOpenCLKernelCreator class.
|
|
/// </summary>
|
|
|
|
namespace EmberCLns
|
|
{
|
|
/// <summary>
|
|
/// Kernel creator for density filtering.
|
|
/// This implements both basic log scale filtering
|
|
/// as well as the full flam3 density estimation filtering
|
|
/// in OpenCL.
|
|
/// Several conditionals are present in the CPU version. They
|
|
/// are stripped out of the kernels and instead a separate kernel
|
|
/// is created for every possible case.
|
|
/// If the filter width is 9 or less, then the entire process can be
|
|
/// done in shared memory which is very fast.
|
|
/// However, if the filter width is greater than 9, shared memory is not
|
|
/// used and all filtering is done directly with main global VRAM. This
|
|
/// ends up being not much faster than doing it on the CPU.
|
|
/// String members are kept for the program source and entry points
|
|
/// for each version of the program.
|
|
/// </summary>
|
|
class EMBERCL_API DEOpenCLKernelCreator
|
|
{
|
|
public:
|
|
DEOpenCLKernelCreator();
|
|
DEOpenCLKernelCreator(bool doublePrecision, bool nVidia);
|
|
|
|
//Accessors.
|
|
const string& LogScaleAssignDEKernel() const;
|
|
const string& LogScaleAssignDEEntryPoint() const;
|
|
const string& GaussianDEKernel(size_t ss, uint filterWidth) const;
|
|
const string& GaussianDEEntryPoint(size_t ss, uint filterWidth) const;
|
|
|
|
//Miscellaneous static functions.
|
|
static uint MaxDEFilterSize();
|
|
static double SolveMaxDERad(double desiredFilterSize, double ss);
|
|
static uint SolveMaxBoxSize(uint localMem);
|
|
|
|
private:
|
|
//Kernel creators.
|
|
string CreateLogScaleAssignDEKernelString();
|
|
string CreateGaussianDEKernel(size_t ss);
|
|
string CreateGaussianDEKernelNoLocalCache(size_t ss);
|
|
|
|
string m_LogScaleAssignDEKernel;
|
|
string m_LogScaleAssignDEEntryPoint = "LogScaleAssignDensityFilterKernel";
|
|
|
|
string m_GaussianDEWithoutSsKernel;
|
|
string m_GaussianDEWithoutSsEntryPoint = "GaussianDEWithoutSsKernel";
|
|
|
|
string m_GaussianDESsWithScfKernel;
|
|
string m_GaussianDESsWithScfEntryPoint = "GaussianDESsWithScfKernel";
|
|
|
|
string m_GaussianDESsWithoutScfKernel;
|
|
string m_GaussianDESsWithoutScfEntryPoint = "GaussianDESsWithoutScfKernel";
|
|
|
|
string m_GaussianDEWithoutSsNoCacheKernel;
|
|
string m_GaussianDEWithoutSsNoCacheEntryPoint = "GaussianDEWithoutSsNoCacheKernel";
|
|
|
|
string m_GaussianDESsWithScfNoCacheKernel;
|
|
string m_GaussianDESsWithScfNoCacheEntryPoint = "GaussianDESsWithScfNoCacheKernel";
|
|
|
|
string m_GaussianDESsWithoutScfNoCacheKernel;
|
|
string m_GaussianDESsWithoutScfNoCacheEntryPoint = "GaussianDESsWithoutScfNoCacheKernel";
|
|
|
|
bool m_DoublePrecision;
|
|
bool m_NVidia;
|
|
};
|
|
}
|