Posts by Janus

1) Message boards : Comments and discussion : 3208 (Message 15567)
Posted 12 days ago by Profile Janus
Post:
No news on this one, the clients kinda work but there are very few. Not sure this is worth pursuing as the platform makes it difficult to debug.

Will probably focus more on OpenCL if the first test hadn't been a complete disaster.
2) Message boards : Number crunching : @ Janus - Volunteer moderator Project administrator (Message 15562)
Posted 13 days ago by Profile Janus
Post:
Time-left-estimation is broken with Cycles renders like this one. Just ignore it.
3) Message boards : Comments and discussion : 3499 (Message 15560)
Posted 14 days ago by Profile Janus
Post:
There only seems to be one RGBA layer in the multilayer EXR output (the rendered scene), is that intentional?


My bad, it does indeed seem to contain all the information. That's pretty cool, I didn't know you could store the entire compositor state like that.

It prompts the question, what on earth should the preview show in a situation like this? The server doesn't know which layers are the RGBA layers for the output (or if there are any colour layers), it just gets a whole bunch of individual colour channels, some of which are part of the ambient occlusion, some part of the mist, some part of UV etc... Maybe a stitch of all the layers next to each other in greyscale? Hm...
4) Message boards : Comments and discussion : 3499 (Message 15552)
Posted 14 days ago by Profile Janus
Post:
Multilayer EXR is still experimental, not sure what we will get out of this - the preview may be black
5) Message boards : Comments and discussion : 3499 (Message 15551)
Posted 18 days ago by Profile Janus
Post:
Sorry for the delay, this one is a bit more complicated than usual.

I'm getting a warning:
LIB: material 001-Rough-Wood-Bois missing from addons/materials_library_vx/Wood.blend
All used textures and materials must be contained in the .blend-file itself since we don't have any addons on the farm.
That said, it seems to be rendering just fine, so maybe we will try it anyways.

There only seems to be one RGBA layer in the multilayer EXR output (the rendered scene), is that intentional?
6) Message boards : Comments and discussion : 3490 (Message 15540)
Posted 22 May 2019 by Profile Janus
Post:
Validation for this session is not working since the render is not deterministic. Switching to best-effort rendering mode.

Some units will be marked as invalid initially but will be properly credited in the next PayDay run.
7) Message boards : Tutorials : Using Filmic Blender (Message 15533)
Posted 1 Apr 2019 by Profile Janus
Post:
False colour, log, raw, film, rrt, filmic and default seem to be the colour profiles available with the current version available on the farm.

You may have to request the use of EXR in the upload description if you need that, since that's still an experimental feature.
8) Message boards : Comments and discussion : 3487 (Message 15532)
Posted 1 Apr 2019 by Profile Janus
Post:
Blender may not be storing the simulation data in the blend-file but rather in a random directory on your machine. Let's see where this goes
9) Message boards : Comments and discussion : 3485 (Message 15526)
Posted 28 Feb 2019 by Profile Janus
Post:
There may be an issue with the physics simulation for this session
10) Message boards : Problems and Help : Submit for GPU (Message 15521)
Posted 10 Feb 2019 by Profile Janus
Post:
At the moment we try to run only enough GPU sessions in order to test that the code is working - there is an issue calculating how much credit to provide for GPU workunits and it is causing them to be particularly cumbersome since credit has to be granted or corrected manually.
11) Message boards : Comments and discussion : 3478 (Message 15518)
Posted 22 Jan 2019 by Profile Janus
Post:
This session was put on hold as we do not yet support the yet-to-be-released Blender 2.80 version.

If you like to I can reject this one and you can try uploading the scene again with a .blend-file from Blender 2.79b?
12) Message boards : Comments and discussion : 3476 (Message 15517)
Posted 22 Jan 2019 by Profile Janus
Post:
Sorry about the delay, the project ran out of memory based on the limit from the upload form and I missed the notice.

- Changed required memory from 1GB to 3GB
- Changed resolution to match .blend-file
- Changed split to 4 from 128 which seems a bit excessive.
13) Message boards : Server backend and mirrors : Time Machine (Message 15510)
Posted 17 Oct 2018 by Profile Janus
Post:
The files on the mirror suggest it was last updated 2009-10-04
That's almost a decade ago!

Thanks for being part of the mirror network back then - bandwidth was a real issue both for SETI@home and BURP for several years
14) Message boards : Comments and discussion : 3468 (Message 15503)
Posted 24 Sep 2018 by Profile Janus
Post:
Putting this session on hold until the new validator/stitcher code has been deployed
15) Message boards : Comments and discussion : 3470 (Message 15502)
Posted 24 Sep 2018 by Profile Janus
Post:
This session does not render consistent results. Switching to best effort validation. This will mark some results as "invalid" right now but they will also be credited in the next Payday run.
16) Message boards : Comments and discussion : 3468 (Message 15493)
Posted 20 Sep 2018 by Profile Janus
Post:
Yes, you are perfectly right, stitching of this session has been manually stopped. Normally it happens as part of the last workunit to return to the server but in this case it will take quite some time before we can proceed.
17) Message boards : Comments and discussion : 3464 (Message 15491)
Posted 19 Sep 2018 by Profile Janus
Post:
Updated the guide, thanks!
18) Message boards : Comments and discussion : 3468 (Message 15490)
Posted 19 Sep 2018 by Profile Janus
Post:
Streaming the images in a tile-based fashion is actually the way we do it normally for this exact reason.

OpenEXR and OpenEXR multilayer support is still a bit experimental since it relies on a 3rd party piece of software to load and manipulate the files rather than the native support we have for PNG. Unfortunately it has to load an entire colour layer at once in order to allow us to inspect it - at least for now - and that just takes up a whole lot of memory.

There's something wrong with how our current code disposes of the colour layers once they have been used, there is some memory that does not get released immediately. It is not a memory leak as such, but rather a question of needing to get rid of the previous layer before loading the next.
The stitcher definitely has the same issue.

Will have to dig a bit deeper.

I feel kinda bad with my scene causing so much trouble


No need to!

This is an interesting test since the validator system is currently being rewritten to better handle Cycles (and similar physical-based renderers with noisy results) and it would be nice to be able to handle large frames too - even when they are EXR.

Fun fact of the day: If you allocate enough RAM then "top", the Linux task manager, will show resource allocation in terrabytes... never seen that before.
19) Message boards : Comments and discussion : 3468 (Message 15485)
Posted 18 Sep 2018 by Profile Janus
Post:
Yeah this one is definitely a bit of a mouthful, each part is using almost 12GB of RAM while being checked. Previously we have been running multiple validators simultaneously but in this case that is a really bad idea because the server simply runs out of memory.

It is probably going to be even worse when it has to stitch it all together.

Switching the session to "best effort" mode which means some results will be marked "invalid" now but they will be fixed and credited later.

I'll have another look at this tomorrow evening.
20) Message boards : Number crunching : I want to be part of this project (Message 15477)
Posted 17 Sep 2018 by Profile Janus
Post:
That commandline looks a bit odd - did it get eaten by the forum? (also doesn't BOINC already add the threadcount to the commandline automatically with the current plan class?)

If you post a link to a completed result you can check towards the very top how many cores it was launched with. Check the number following the small "-t" parameter. That is the thread-count (in this case 32):
Executing blender.exe -noaudio --factory-startup -y -b in -P clirender.py -- -F PNG -T 16 -t 32 -f 18 0.0 0.0 1.0 1.0


Just assuming that you want to set this specifically for one app, otherwise it is much easier to set the overall CPU limit for BOINC.


Next 20