• January 18, 2018, 10:40:40 AM

### Recent Posts

Pages: [1] 2 3 ... 10
1
##### Board Rules and Guidelines / Re: New Features of V2
« Last post by 1bryan1 on Today at 07:13:35 AM »
We decided to go without annoying ads.
All expenses of the update and server costs are currently payed by the FFteam out of our own pockets.
Any support from you is highly appreciated - paypal@fractalforums.org

Curious. How much per year is this costing the team?
2
##### Fractal Image Gallery / Captain Balls hunting for humans on the red planet Darmma
« Last post by weavers on Today at 07:00:37 AM »
Captain Balls hunting for humans on the red planet Darmma

https://fractalforums.org/index.php?action=gallery;sa=view;id=275

The Royal AI, Queen Majesty Sharkkadooms son, Captain Balls, seeking vengeance, searches the Universe for humans, to see more paintings of the evil characters, and read more about the battle royal : goto our website https://badpersonalities.wordpress.com/ or Badpersonalities.blogspot.com
3
##### Kalles Fraktaler / Re: Kalles Fraktaler 2 + GMP
« Last post by claude on Today at 06:35:14 AM »
I made some attempts at compiling smb,

Oh, Windows makes things difficult I guess.

Quote
If knighty could write an "export to kfb" (and read rotate data from kfr) option it could be used as a "final renderer" after finding want you want in KF.
good idea.

Quote
For DE coloring just feed the analytic DE numbers to whatever method now processes the numerical DE and it should work transparently, no? It's the same data except more accurate.

Sure, but so far I'm hackily storing the DE in the Trans array instead of allocating separate space for it, need to fix that....

Quote
What's the problem with size estimation for that burning ship?
It's way off in some circumstances, eg my embedded Julia set NR zoom finds near the miniset depth and the miniset NR zoom finds deep in the miniset.  Other locations are less wrong, and it definitely worked at some point...

Currently working on the Newton-Raphson zoom GUI in KF, hoping to add "capture start zoom depth and use average with minibrot size" as suggested by Dinkydau to improve zoom accuracy, also a custom factor.
4
##### Kalles Fraktaler / Re: Kalles Fraktaler 2 + GMP
« Last post by gerrit on Today at 06:17:54 AM »
Maybe you could hack superMB to add more exciting colouring?  That speed boost is incredible
I made some attempts at compiling smb, even downloaded some Qt IDE it needs, but then it needed 2 more things which I couldn't figure out how to install, after downloading, so that the IDE knows it and the config file works and the paths are ok and more stuff I hate to do. Also it's been 15 years since I did any C stuff and I can't even understand the code without running it.  $$\alpha_0$$ can learn to play world class chess in 4 hrs, and cars can drive themselves, but AI has not figured out how to compile code.

smb is not really a replacement for KF as it is missing all the important interactive features, of which NR zooming is probably the most spectacular.
If knighty could write an "export to kfb" (and read rotate data from kfr) option it could be used as a "final renderer" after finding want you want in KF.

For DE coloring just feed the analytic DE numbers to whatever method now processes the numerical DE and it should work transparently, no? It's the same data except more accurate.

What's the problem with size estimation for that burning ship?

5
##### Image Threads / Re: gerrit images
« Last post by claude on Today at 05:10:04 AM »
6
##### Kalles Fraktaler / Re: Kalles Fraktaler 2 + GMP
« Last post by claude on Today at 05:07:14 AM »
How's 2.13 going? I'm really looking forward to real DE rendering. My latest pretty picture took 12s in superMB and 1hr in KF to get the same quality as you need to oversample the numerical DE by so much.

Maybe you could hack superMB to add more exciting colouring?  That speed boost is incredible
In my renderers I upload calculated data as an RGBA32F texture to OpenGL, and use GLSL shaders to do the colouring.  I think SuperMB uses OpenGL somewhere already?

Very slow progress on 2.13.  Currently DE works for Mandelbrot power 2 and Burning Ship power 2, but the colouring is hardcoded to a black on white with no twiddleable options yet.

I have mostly been working on 'et's formula compiler instead, finding time to work on KF only while doing long renders with 'et'.  Back-porting (parts of) 'et' to KF isn't a high priority, adding manual derivative calculations to the XML even less.  Instead I want to port mandelbrot-perturbator's GTK GUI to Haskell, and make it use the 'et' engine.  So much work to do: currently to change formula I have to edit source and recompile the whole thing, I should make an API and compile to DLL/SO at runtime, then load it.  But calling out to GCC might not be very nice on Windows (thinking long term), especially as the C code I'm generating needs MPFR.

Also, I ran into a flaw in my Burning Ship size estimate, which means I can't NR zoom properly yet.  I asked on M.SE about it, in case you feel like getting imaginary internet points.
7
##### Fractal Humor / Re: Fibonacci is always right..
« Last post by gerrit on Today at 04:48:26 AM »
Trump apparently has a very high IQ which normally is a good thing? in his case it it interferes with the ability to make rational statements
Statements by Trump about his high IQ are unreliable due to his low IQ.
8
##### Kalles Fraktaler / Re: Kalles Fraktaler 2 + GMP
« Last post by gerrit on Today at 04:43:54 AM »
Thanks, fixed in 2.12.9 (to be released this month, if all goes well.)
How's 2.13 going? I'm really looking forward to real DE rendering. My latest pretty picture took 12s in superMB and 1hr in KF to get the same quality as you need to oversample the numerical DE by so much.
I don't care about NR zooming etc for burning ship but I'm sure many people do.

Thanks for all your efforts, they are greatly appreciated.
9
##### Fractal Mathematics And New Theories / Re: Fast Mandelbrot set research summary
« Last post by gerrit on Today at 04:19:49 AM »
Thanks for the feedback, I've updated the draft.

I don't want to include sample code, just focus on the algorithms.
I also want to avoid covering representation functions as there as so many, and there's not much new there I think. Numerical DE without derivative computation is the exception which I covered.
Douady theory (rays etc.) are a bit too mathematical for me, though it's fascinating. Maybe I should add at least references to it.
I think the density is about right for a numerical analysis/algorithm paper. Could use some elaborations at some points though.
Finally, I'll add my name if I get some co-authors and wrap it into a real paper for refereed journal publication. Any volunteers?
Mandelbrot pictures are generally considered trivial from a computational perspective. so I think this would be of interest.
10
##### Kalles Fraktaler / Re: Kalles Fraktaler 2 + GMP
« Last post by claude on Today at 04:19:29 AM »
If in iterations dialog you deselect "auto approximation terms", set to a lower value, it still resets it to 60 (I guess the auto value for my location).

Thanks, fixed in 2.12.9 (to be released this month, if all goes well.)
Pages: [1] 2 3 ... 10