News: Visit the official fractalforums.com Youtube Channel
 
*
Welcome, Guest. Please login or register. October 23, 2014, 06:28:42 AM


Login with username, password and session length



Pages: [1]   Go Down
  Print  
Share this topic on DiggShare this topic on FacebookShare this topic on GoogleShare this topic on RedditShare this topic on StumbleUponShare this topic on Twitter
Author Topic: Cross platform with infinite precision  (Read 1330 times)
0 Members and 1 Guest are viewing this topic.
cortexman
Guest
« on: January 17, 2010, 12:50:12 AM »

Hi all, I have been lurking for a while now. I think I've seen *all* of your amazing Mandelbulb renders, pics & vids included. Very impressive work you've all done. Unfortunately the only renderer that I've found which works on Linux (located here) doesn't produce very impressive renders and is buggy.

I have some ideas for the software which I would use to produce a cross platform renderer. Although I am capable of writing C/C++ it makes my skin crawl, therefore I would be interested in using PyCUDA.

Additionally, from my reading of the Mandelbulb forums the current limitation of all renderers is that even those that support infinite precision (such as UltraFractal) are extremely slow. I have available to me a cluster with 24 nodes, each node of which has an nVidia Quadro FX 370 LP (which is CUDA capable). Additionally each node is connected using Infiniband for very low latency. Therefore if we could identify a parallel algorithm for unlimited precision, and a Mandelbulb implemenation on CUDA, we could use my cluster to produce deep renders. We can utilize the Infiniband interconnects via PyMPI or a similar technology. Even if we can't come up with an inherently parallel infinite precision algorithm, perhaps we can make a slower one faster using the cluster.

The first step to such a project is to implement a renderer via PyCUDA. Since I am new to graphics and fractal rendering in general I would like to seek some assistance in this matter. If you're willing to work together on a PyCUDA impl, or have other ideas about how we should go about this, give me a shoutout.  grin
Logged
cortexman
Guest
« Reply #1 on: January 17, 2010, 04:41:29 AM »

I've actually had a change of heart on this. After discovering the GMP arbitrary precision arithmetic library I realized it wouldn't be that hard to implement a renderer using our own software, emergent (http://grey.colorado.edu/emergent). We have a neural network simulator with a very generic architecture (by design) that is written in C++ and optimized for speed. It even has the ability to open OpenGL canvases in new tabs. I just hadn't thought of using it as a fractal explorer until now smiley It's written in C++. I think I will call it the emergent fractal explorer. Now I will refocus my efforts around an opengl renderer using gmp if possible. After that I can easily plug it into emergent using our plugin architecture.

Another nice thing: We already package it for linux, mac and windows.
Logged
Pages: [1]   Go Down
  Print  
 
Jump to:  


Related Topics
Subject Started by Replies Views Last post
Mandelbrot Maps - A cross platform fractal explorer Fractal Programs mandelbrotmaps 4 3385 Last post March 20, 2012, 07:46:33 PM
by withad
How much more useable precision will I get using quad precision? Programming Duncan C 1 800 Last post April 15, 2010, 12:16:56 PM
by Botond Kˇsa
Platform K-2 Mandelbulb3D Gallery Kali 0 182 Last post March 12, 2011, 01:38:47 PM
by Kali
The A7 chip as a fractal Platform Announcements & News jehovajah 3 348 Last post October 11, 2013, 09:05:21 PM
by Nahee_Enterprises
Skybot Platform Images Showcase (Rate My Fractal) JoeFRAQ 0 129 Last post January 15, 2014, 01:27:50 PM
by JoeFRAQ

Powered by MySQL Powered by PHP Powered by SMF 1.1.20 | SMF © 2013, Simple Machines

Valid XHTML 1.0! Valid CSS! Dilber MC Theme by HarzeM
Page created in 0.294 seconds with 26 queries. (Pretty URLs adds 0.013s, 2q)