Mantispa discolor

Mantispa discolor

Description

Attempt 2
Disabled the motion for the camera, and changed the scene.

Extra light, better background

I have run 1 frame myself as test. On frame 60 it took 81:44:30.38 hour to render

Message boards : Comments and discussion : 449

Author Message
Profile noderaser
Project donor
Avatar
Send message
Joined: 28 Mar 06
Posts: 506
Credit: 1,548,030
RAC: 75
Message 5899 - Posted: 26 May 2007, 0:32:39 UTC

This one doesn\'t seem to be going very well, eh?
____________

Profile PlainText s.r.o.
Send message
Joined: 11 Apr 07
Posts: 95
Credit: 3,505,542
RAC: 0
Message 5920 - Posted: 29 May 2007, 7:35:24 UTC

I think also so. But sometimes error hapens. For example to me, after a long computation time.

http://burp.boinc.dk/result.php?resultid=2235774

What exit code -10 means? (Exception caught: ProgressMonitor kills) that means, that boinc core gone off? (My notebook is running 6 days without any brake, and boinc also.)

PS: could it be caused by about 3 hours of playing warcraft?

Profile Janus
Volunteer moderator
Project administrator
Avatar
Send message
Joined: 16 Jun 04
Posts: 4461
Credit: 2,094,806
RAC: 0
Message 5921 - Posted: 29 May 2007, 9:07:13 UTC - in response to Message 5920.

PS: could it be caused by about 3 hours of playing warcraft?

Yup, the monitor currently has a known error that makes it kill based on wall time and not CPU-time. This means that if the workunit gets suppressed for more than an hour (by whatever other highly demanding process) it is very likely to fail.
This becomes exceptionally visible on extreme workunits like the ones currently in the render queue.

mip2
Send message
Joined: 12 Jan 07
Posts: 26
Credit: 25,862
RAC: 0
Message 5922 - Posted: 29 May 2007, 12:50:48 UTC

i see in the blend file of 449 the motion blur setting is \"1\",that increase the render time with factor 10.
is it possible to please the users to work more carefully with motion blur in the alpha phase of burp?
the wu i recive renders now 8 hours and need another 6 hours to finish.
i have \"Leave applications in memory while suspended\" set to \"yes\" but if i must shutdown or reboot sometimes, the wu start again with 0.00%.
is this a problem with the linux client?

indigomonkey
Send message
Joined: 31 May 06
Posts: 44
Credit: 18,633
RAC: 0
Message 5923 - Posted: 29 May 2007, 13:05:32 UTC

Actually the motion blur factor does not affect the time to render - it is fact the number of samples (OSA) that controls how many interpolated frames are produced - so in this example, there are 16 (OSA) x 16 (MBlur) = 256 overall samples being produced. The MBlur factor just controls how these resultant blur samples are mixed together.

If you want to turn off the computer, hibernate it - this works for Windows, though I haven\'t tried it in Ubuntu. The program is left in (virtual) memory, and can be started from where it left off when you boot again.

Profile PlainText s.r.o.
Send message
Joined: 11 Apr 07
Posts: 95
Credit: 3,505,542
RAC: 0
Message 5930 - Posted: 31 May 2007, 0:39:40 UTC - in response to Message 5921.

PS: could it be caused by about 3 hours of playing warcraft?

Yup, the monitor currently has a known error that makes it kill based on wall time and not CPU-time. This means that if the workunit gets suppressed for more than an hour (by whatever other highly demanding process) it is very likely to fail.
This becomes exceptionally visible on extreme workunits like the ones currently in the render queue.


Grrrrr, after 50000 seconds of work... (warcraft :-/) what about doing something with checkpoints? Where is problem? (just not to render part, in wu, but divide on client that wu to 32 smaler parts...) No coding problem, just overhead of more blender startst. Am I wrong?

loki
Send message
Joined: 7 Feb 07
Posts: 97
Credit: 158,266
RAC: 0
Message 5951 - Posted: 1 Jun 2007, 4:13:42 UTC

Looks like this wu is not going to finish.

http://burp.boinc.dk/workunit.php?wuid=420063

too many error results, isn\'t it able to just reset the few that needed to finish still?

Profile noderaser
Project donor
Avatar
Send message
Joined: 28 Mar 06
Posts: 506
Credit: 1,548,030
RAC: 75
Message 5952 - Posted: 1 Jun 2007, 6:30:34 UTC - in response to Message 5930.

Grrrrr, after 50000 seconds of work... (warcraft :-/) what about doing something with checkpoints? Where is problem? (just not to render part, in wu, but divide on client that wu to 32 smaler parts...) No coding problem, just overhead of more blender startst. Am I wrong?


Before you start your game, suspend BOINC. This will allow you to play without experiencing any lag, etc. You also won\'t loose any progress on your workunits, as they stay in memory while suspended. When you\'re done playing, resume computation and you should be all set.
____________

Profile PlainText s.r.o.
Send message
Joined: 11 Apr 07
Posts: 95
Credit: 3,505,542
RAC: 0
Message 5953 - Posted: 1 Jun 2007, 10:43:25 UTC - in response to Message 5952.

Grrrrr, after 50000 seconds of work... (warcraft :-/) what about doing something with checkpoints? Where is problem? (just not to render part, in wu, but divide on client that wu to 32 smaler parts...) No coding problem, just overhead of more blender startst. Am I wrong?


Before you start your game, suspend BOINC. This will allow you to play without experiencing any lag, etc. You also won\'t loose any progress on your workunits, as they stay in memory while suspended. When you\'re done playing, resume computation and you should be all set.


Yes, I hope I do not forgot to pause boinc. I should probably buy some more cores :-) and set the afinity for war3.exe only to one :-)

PS: now my boinc work on wu\'s from 450 session for 23 hours, and only 60% done. wow. I hope it will be nice session, after end.

Profile PlainText s.r.o.
Send message
Joined: 11 Apr 07
Posts: 95
Credit: 3,505,542
RAC: 0
Message 5955 - Posted: 1 Jun 2007, 13:29:39 UTC
Last modified: 1 Jun 2007, 13:31:25 UTC

http://burp.boinc.dk/result.php?resultid=2173988

wtf?!!! I had my pc working on that more than 24 hours and error? And nothing else, than boinc wasnt on cpu. (dont count firefox)

I chcecked boinc, that was working, and I go out. (for lunch etc.) after arriving, boinc was working on other wu from burp. I was happy, tahat the long wu is done, buth after look at the statistic...

Just one question: why?

EDIT: on my other pc no error. just on my notebook.

Profile noderaser
Project donor
Avatar
Send message
Joined: 28 Mar 06
Posts: 506
Credit: 1,548,030
RAC: 75
Message 5957 - Posted: 2 Jun 2007, 0:24:26 UTC

Looks like it\'s finally done, but now 450 is having problems as well. :(
____________

Profile Janus
Volunteer moderator
Project administrator
Avatar
Send message
Joined: 16 Jun 04
Posts: 4461
Credit: 2,094,806
RAC: 0
Message 5961 - Posted: 2 Jun 2007, 9:52:51 UTC

Sessions like these at least give a lot of nice debug info to aid the bug-hunts.

Profile PlainText s.r.o.
Send message
Joined: 11 Apr 07
Posts: 95
Credit: 3,505,542
RAC: 0
Message 5964 - Posted: 2 Jun 2007, 21:27:32 UTC - in response to Message 5961.

Sessions like these at least give a lot of nice debug info to aid the bug-hunts.


I doesnt looked at it from this point of view. Oki, I won\'t be angry nextime on error :-) (no irony)

PS: I think, that number of workers somehow fall down. But imho now the burp renders nice video. that viper. (I have looked at animation in blender wireframe view, and it looks very nice)


Post to thread

Message boards : Comments and discussion : 449