Where communities thrive


  • Join over 1.5M+ people
  • Join over 100K+ communities
  • Free without limits
  • Create your own community
People
Repo info
Activity
Krzysztof Marczak
@buddhi1980
There is no guaranty that the image will look the same like in old program but with small corrections of effects should be all right.
biberino
@biberino
hi guys
i just downloaded the code from sourceforge a couple of days ago and modified it to fit my needs. And today i see that there's a github repo :) my current changes to the code are kinda hacky and it was a first test if this would work, which it does. I want to integrate the changes as clean as possible so i might have some questions about the code. Basically I need to add a new type of iteration (twin procedure), invented by a german physicist, which produces very interesting results in 2D. I also need to implement some functions, but this looks easy.
Happy holidays!
mclarekin
@mclarekin
@biberino Implementing twin procedure iterations in 3D?
mclarekin
@mclarekin
3D. I had a look at your repo's , you obviously know what you are doing:) In fractal_list.hpp use numbers starting with 800 for your formulas if you want.
Krzysztof Marczak
@buddhi1980
Your first commits on GitHub you can do as Pull Requests. Them we will be able to review changes before merging the code.
biberino
@biberino
@mclarekin yes, i want to implement it for 3D. I'll give u guys some background on what that iteration looks like before i do that.
@buddhi1980 alright thanks :) i'll start with a formula. I have a question, what does sExtendedAux.DE stand for?
I get the best results for my formula when i use
aux.DE = aux.r
Krzysztof Marczak
@buddhi1980
aux.DE is an approximarion of first derivative of fractal formula. It is used to calculate estimated distance to fractal surface. In many cases it is difficult to find correct formula for it and we are trying to find it experimentally (as you did).
biberino
@biberino
ah i see thanks for the help
biberino
@biberino
Some background: We found that if one uses z(z) +c instead of z^2+c, where z* is the complex conjugate of z, then the resulting fractal gets stable and looks like a cross section of an eye. If you use the twin procedure, which takes away some of the unnatural sharpness of the image, even more details (lense, pupil) emerges! Images here
now i want to reproduce some of the images we found in 3D. For now i use the multiplication used in the mandelbulb fractal and (x,y,z)* = (x,-y,-z), which already shows the eye. But if u have better ideas pls let me know.
biberino
@biberino
i see the .cl files are autogenerated, is it a problem if i use functioncalls in fractal_formulas.cpp?
Sebastian Jennen
@zebastian
@biberino the missing functions will break the opencl kernel for this function. I will look into this and port the required quaternion functions to opencl and let the auto-generate run. @buddhi1980 are you happy with these new added functions in algebra.* regarding quaternion algebra? or do you want to change the cpp code before the porting?
Krzysztof Marczak
@buddhi1980
These new functions are added to common_math.cpp which is all right for me. It would change all std::pow(), std::sin(), etc.. to just pow(), sin(). It will improve readability of the code and will not change how the code works (it already uses functions from std)
Sebastian Jennen
@zebastian
OK, will do...
mclarekin
@mclarekin
@biberino there is room for optimization of code but maybe get the formula working first.
Where does int n come from, the ui??
aux.DE is for the preferred analytic distance estimation (DE) method, but we can also try using the DeltaDE method which does not require any calculation within the formula.
mclarekin
@mclarekin
@buddhi1980 Historical interest, the reference
//hypercomplex multiplication taken from:
//https://www.physicsforums.com/threads/16-different-spherical-coordinate-systems.331883/
is twinbee (Daniel White) back in 2009
biberino
@biberino
@mclarekin you mean the n from hypercomplex_pow_constant? It just computes vec^n, I don't use that function yet but i might in the future
Sebastian Jennen
@zebastian
This may be a good starting point for timeline-based animation
mclarekin
@mclarekin
@zebastian
Yep, worth investigating
CurveShape { EaseInCurve, EaseOutCurve, EaseInOutCurve, LinearCurve, SineCurve, CosineCurve }
Krzysztof Marczak
@buddhi1980
Purpose of this class is to provide animations for GUI elements. I have already investigated if it is possible to use this class for morphing of parameters, but it is for different things.
TasMania17
@TasMania17
Nvidia made GPUs just for bitcoin mining. Now they are being sold cheaply online. I bought one called a P106-100 as a second GPU unit to go with a GTX-1060. It works under Ubuntu and is recognized by Mandelbulber OpenCL. It is listed by 'nvidia-smi' and 'clinfo'. I don't think they work with Windows because of 'signed driver enforcement' or something. I saved a few hundred and got lucky :)
Starmute
@Starmute
they were designed for ethOS which is a cryptocurrency mining operating system lol
Yevgeny Pats
@yevgenypats
Hi All, Wanted to contribute an integration to Fuzzit (Our Continuous Fuzzing platform - free for oss). Is this relevant? Also the idea was not only to run long running fuzzing jobs but also quick sanity tests on the generated corpus every commit which should provide more stability.
Krzysztof Marczak
@buddhi1980
@yevgenypats , Sorry for late answer. I havent't seen your message.
We already use TravisCI service for Linux builds, Appveyor for Windows and CircleCI for MacOS. They covers all of our needs. These platforms already do compilation of program and run tests.