Coolio. Take care.
Mike
Hello,
We have found an example of a quick local render/slow Deadline render. Please see the attached log file.
This single-frame job batch renders in 4m30s on my local machine, but renders in 27m04s on a farm blade. Quite a difference!
At the start of the job, there are a lot of errors being thrown concerning “frame counter”. This is most likely an expression error caused by our internal frame counter and we are investigating ways of removing it before render-time. On the log, it looks like the errors take just a few seconds to go away, so I’m not sure how detrimental it is to the final render time.
Another weird spot is during the texture loading phase: there are a few textures which take a few minutes for the renderer to load. I loaded a few of these problematic textures on my machine, noting the speed of loading. They loaded quite quickly; within 3-seconds. Perhaps there is more to it than just loading the data, maybe there is some converting going on under the hood.
Looking at the log again, it really seems that there is a bottleneck happening during the texture loading phase. It takes about 12-minutes to load these textures, much too long for what our infrastructure can provide. The actual rendering of the image takes about 13-minutes. The total is a far cry from my local rendering of 4.5 minutes…
Thinkbox wizards: is there anything you can glean from the log? Thanks!
MayaJobLog.txt (194 KB)
Probably best to create a new forum thread to not confuse others
Observations:
- This needs fixing -
-
12 x mins of loading bitmaps. - when multiple slaves on the farm try to run the Maya job at the same time, is your network file server struggling to deliver the performance? If you run the job again BUT with only 1 Slave on the farm rendering this job, do the load bitmap times become reasonable? Bear in mind, if you already have the Maya scene file open, then the bitmaps could be already loaded…
-
12 x mins of V-Ray rendering - not much we can do here…that seems to be how long V-Ray just takes…
-
Something that looks odd is the final output filename with a “space” character at the beginning of each fileName, like this:
" multimatte_a.0251.png" <–whitespace character at beginning of fileName
" multimatte_b.0251.png" <–whitespace character at beginning of fileName
" multimatte_c.0251.png" <–whitespace character at beginning of fileName
" multimatte_d.0251.png" <–whitespace character at beginning of fileName
" rgba.0251.png" <–whitespace character at beginning of fileName
What exact version of Deadline are you using? What does you V-Ray file naming convention look like in Maya’s Render Settings dialog? Which, if any, V-Ray file naming $TOKENS are you using?
hello dear …
would you please specify how you removed ot exectly … as im face a slow nuke render on deadline … but i didnt understand where to find that “savebackup” script you mentioned and how to remove it from the menues as you mentioned ?
thank you
dear …
regarding point number 3.
how can i find and remove that script ?
thanks
Hello @Mohammed_Sami_Idrei
Thanks for reaching out. The script you are referring to is a custom script made by the customer themselves for their workflow. It may not apply to your case. Can you explain more about your issue of slow renders?
I need to look at the verbose Worker logs (Check off Worker Verbose Logging):
Then restart the Worker, reproduce the issue and share Worker logs from application logs folder from the render node which has rendered slow. The application logs sit here: Logs — Deadline 10.2.1.1 documentation
Hi, jumping in the topic. I have not read the entire convo, did a quick search for “Houdini” and it didn’t seem to be mentioned anywhere.
I’m having the same issue with Houdini. Is there a way to force that “Batch render” feature with Houdini? Renders are about 3 to 4 times slower through deadline than when rendering directly with Houdini / Redshift.
I checked everything i could think off, all assets are found ($JOB), licenses check is all good. I thought “submit houdini scene” would do the trick along with job is frame dependent… would make sense that it does not reload the scene in between tasks. But nothing does it. Still seems to reload everything in between frames. Amount of frames perk taks is higher than the total amount of frames (rendering with a single machine here).
Any idea? Thanks in advance.
Cheers.
EDIT: My bad, it is going way faster since i checked ‘SUbmit scene / Frame dependent’. Apologies. CHeers.