[RV480.1447] Thea for SketchUp 2017 and older
Moderator: Tomasz
- Anthony
- Licensed User v1.5
- Posts: 2138
- Joined: Wed Jan 08, 2014 9:33 am
Is it a constantly reproducible scenario with a specific scene? If yes and we can follow your instructions to reproduce the problem with a specific scene then we can of course fix this.
“ No man ever made a great discovery without the exercise of the imagination. ”
— George Henry Lewes
— George Henry Lewes
-
- Licensed User v1.5
- Posts: 612
- Joined: Sat Sep 13, 2014 2:10 pm
I'll try different scenes on my home 1070 rig and check if this is reproducible.
With the same file on my office Titan X rig (same thea4SU version, same nvidia driver), it works flawlessly (presto MC bucketed network render).
This is very weird.
EDIT: I've got the same error on my Titan X rig so I guess is a file or scene issue.
I'll dig it.
With the same file on my office Titan X rig (same thea4SU version, same nvidia driver), it works flawlessly (presto MC bucketed network render).
This is very weird.
EDIT: I've got the same error on my Titan X rig so I guess is a file or scene issue.
I'll dig it.
- Tomasz
- Thea Developer
- Posts: 8919
- Joined: Fri Jul 18, 2008 5:09 pm
- Contact:
You may have stored rendering settings in a scene. When you select it, they are getting loaded, hence bucket may get disabled. Isn't it no the case?
-
- Licensed User v1.5
- Posts: 612
- Joined: Sat Sep 13, 2014 2:10 pm
Nope. The error was due to a corrupted component. I'm sorry. Nothing to do with the version of the plugin.Tomasz wrote:You may have stored rendering settings in a scene. When you select it, they are getting loaded, hence bucket may get disabled. Isn't it no the case?
Nevertheless it would be easier if Thea console log could precise where the error comes from...

- Anthony
- Licensed User v1.5
- Posts: 2138
- Joined: Wed Jan 08, 2014 9:33 am
Well even if it is a corrupt component it would be good to know what the corrupt component was so that we can take this into account. The defacto rule is always render no matter what happens. So if we need to add a contingency scenario here it would be useful to know just what that issue was.
“ No man ever made a great discovery without the exercise of the imagination. ”
— George Henry Lewes
— George Henry Lewes
-
- Licensed User v1.5
- Posts: 612
- Joined: Sat Sep 13, 2014 2:10 pm
I successfully cornered and nailed the culprit component/proxy.Anthony wrote:Well even if it is a corrupt component it would be good to know what the corrupt component was so that we can take this into account. The defacto rule is always render no matter what happens. So if we need to add a contingency scenario here it would be useful to know just what that issue was.
The error is totally 100% reproducible. It's even funny because the first time I launch the bucket Presto render it detects the VRAM correctly (but without using the GPU for rendering), and the second time it says not enough memory (component=3Mb vs Titan X VRAM=12Gb) lololol
Do you want me to send you the file?

Last edited by quickit on Tue Nov 29, 2016 6:47 pm, edited 1 time in total.
- Tomasz
- Thea Developer
- Posts: 8919
- Joined: Fri Jul 18, 2008 5:09 pm
- Contact:
Yes, the proxy uses Fresnel ramp in an emitter. This is not supported yet and there wasn't a "safety check" for this, so it results in a crash.quickit wrote:Did you find anything?