Logo by Trifox - Contribute your own Logo!

END OF AN ERA, FRACTALFORUMS.COM IS CONTINUED ON FRACTALFORUMS.ORG

it was a great time but no longer maintainable by c.Kleinhuis contact him for any data retrieval,
thanks and see you perhaps in 10 years again

this forum will stay online for reference
News: Support us via Flattr FLATTR Link
 
*
Welcome, Guest. Please login or register. September 23, 2019, 09:59:54 AM


Login with username, password and session length


The All New FractalForums is now in Public Beta Testing! Visit FractalForums.org and check it out!


Pages: 1 [2]   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: Mandelbulber 1.20  (Read 2865 times)
0 Members and 1 Guest are viewing this topic.
mclarekin
Fractal Senior
******
Posts: 1739



« Reply #15 on: January 03, 2014, 05:36:00 AM »

Thanks even more quaz0r and buddhi  A Beer Cup A Beer Cup A Beer Cup.  The version 1.20-32 wouldn't work on Win7-64  AMD A10-5800GHz 4795 Radeon HD 7660D but 1.20-64 does and wow, default  mandelbulb  with openCL normal setting renders in 3.2 sec on CPU and an amazing 0.68 sec on GPU, faster than I can say "one elephant". grin (For non English speakers "one elephant, two elephant .... is an approximation for counting seconds)
 
My laptop i3 2.13 takes 8 seconds with openCL using CPU.

Maybe it's time for an animation grin
« Last Edit: January 03, 2014, 06:50:01 AM by mclarekin, Reason: Clarifying \"elephants\" » Logged
lukesleftleg
Navigator
*****
Posts: 63


« Reply #16 on: January 05, 2014, 05:16:32 PM »

Hi Buddhi,

Thank you once again for another version of my favourite fractal renderer. smiley

I just thought I'd let you know that I think I'm getting close to understanding my crash problems, and all evidence points to it not being Mandelbulber's fault at all.

I've noticed previously, that the 'random crashes' I was getting in OpenCL mode seemed to happen most often when a ray was going at a very narrow angle to a surface (or in other words, had a lot more steps to march along the ray), which in turn increased the time it took to calculate that ray.

It seemed quite counter-intuitive, but I've found in versions 1.19 and 1.20, if I reduce the GPU RAM usage to something really small, say 16MB, then it was much more likely to render the problem frames without crashing, as it seemed to be rendering in smaller chunks.

Anyway, I was reading around the forum, and it appears that Fragmentarium has some similar issues, and this explanation from Syntopia seems to be entirely consistent with the problem:

The GPU is completely locked until a frame render completes in Fragmentarium. (Which is why Windows just kills the display driver if a GPU operation takes too long - it cannot show a dialog asking if you want to stop the operation).

Theoretically, the new WDDM 1.2 driver model in Windows 8 supports GPU Pre-emption (multitasking as we know it from desktop OS'es), but I don't know if the Nvidia and AMD drivers implement this yet - I don't think so.

The only way to keep the OS responsive is by rendering smaller tiles.
(I'm sure there was another post that went into more detail, where he said that you could adjust the time-out in the control panel, but that it was a bad idea if you didn't know exactly what you were doing, but I can't seem to find it now).

So it looks like we're just going to have to live with it until Nvidia and AMD start writing drivers that can take this into account.

This is actually fine for me. I've mentioned elsewhere that I've been using a macro recorder to just restart the program if it crashes, and it doesn't happen that much anyway. It seems a small price to pay, but it's good to get to the root of the problem. smiley
« Last Edit: January 05, 2014, 05:18:53 PM by lukesleftleg » Logged
magixx
Forums Newbie
*
Posts: 1


« Reply #17 on: January 28, 2014, 01:54:28 AM »

Hi, I'm currently considering putting together a rig for rendering stuff like a 4k electricsheep since I already have the hardware from bitcoin mining, however I'm a bit unsure of the performance and what troubles I would run into. Like I have a slight suspicion that a very weak CPU somehow be a bottleneck. Another issues that I foresee is that if I were to set up 4 Radeon 5830 would this and OpenCL use all 4 GPUs at the same time for a frame without crossfire? I'm guessing no and that I would have to just work on 4 frames at once with each frame having a dedicated GPU. Also is there a way I could see what my rig would be able to output using these 4 cards without actually setting it up?
Logged
KBeezie
Alien
***
Posts: 28


« Reply #18 on: January 31, 2014, 09:51:58 AM »

I can't seem to be able to get netrender to work from one Windows 7 machine to another Windows 7 machine, one being 192.168.1.6 and the other 192.168.1.16 they can ping each other, and windows firewall is turned off on both. when I try it via -nogui it just waits and waits and in the GUI if I click client enable it just waits til it says "No error" but the server side doesn't show any connected clients.
Logged
KBeezie
Alien
***
Posts: 28


« Reply #19 on: January 31, 2014, 10:21:10 AM »

Same issue trying to connect from Windows 8.1 to the Win7 Server.
Logged
Pages: 1 [2]   Go Down
  Print  
 
Jump to:  

Related Topics
Subject Started by Replies Views Last post
Mandelbulber 0.85 Mandelbulber Buddhi 6 2490 Last post July 25, 2010, 10:00:13 PM
by kram1032
Mandelbulber 0.93 Mandelbulber Buddhi 12 4109 Last post October 17, 2010, 03:19:15 PM
by Buddhi
mandelbulber Help & Support ramblerette 1 585 Last post October 18, 2010, 02:56:02 PM
by ramblerette
Mandelbulber 0.94 Mandelbulber « 1 2 » Buddhi 15 4868 Last post October 24, 2010, 09:36:01 AM
by Buddhi
Mandelbulber 1.12 Releases Buddhi 13 4440 Last post June 23, 2012, 08:44:32 PM
by visual.bermarte

Powered by MySQL Powered by PHP Powered by SMF 1.1.21 | SMF © 2015, Simple Machines

Valid XHTML 1.0! Valid CSS! Dilber MC Theme by HarzeM
Page created in 0.131 seconds with 29 queries. (Pretty URLs adds 0.008s, 2q)