I am currently setting up a fluid pipeline at my studio and i have run into something I find strange. I take a fluid simulation and cache it with segmented cache. That part works great. But when I send it to the farm to render things get wierd. The renders turn out fine, but the fluids do a runup to the frame that is being rendered. I have tested this by outputing currentTime, so i am certain it is running through all the frames. Since there is no motion blur I would think the only frames that need to be loaded would be the frame the file was saved on and then the frame that is to be rendered. Is there a solution to this? it eats up a lot of time with long animations.
Thanks
your right it doesn't need to do the run up. I havn't had this particular issue tho so other than that I can't help. Try cache every frame indevidually see if that helps, and make sure that the file is really seeing the cache. Can you be sure the cache is being read? Seems to me like it's not.
ok, so i just logged on to a farm node and opened the file. it does see the cache. I rendered a farme mid-animation, no problem. also all the cache is single frame cache. This is very odd, what clued me into this issue was that the later the file was in the animation the longer it took to render. I am very confused
thanks
hmmm very odd, I wonder if it see the cache one when you open the file, not when it bactches over the farm... but that' snot relly helping you. I set my fluid cache to go to another cache server and set my projects accordingly, not sure that helps but maybe it might give you another idea to try, good luck.
ok i am even more confused. It looks like it does a runup from what ever frame it was save on. not the start frame.
i ran a bunch of tests last night and from the renders i can tell that the farm is not reading the cache. Any ideas on why this would happen or point me in a research direction.
thanks
ok, so after a whole day of working on this i still can figure out why the cache isnt being seen during the batch render. strangely enough the initial state is being seen. Also i have tried both maya sw and mentalRay renders. anyone know of some common problems that might cause this? start with the obvious 
thanks
the path to the cache is differnt on different machines? it being a networkk drive can cause problems too.
the path is relative to the project. the data folder. the strange thing is, no matter what machine i open the files on, in the GUI, the cache is connected. it is only when running the render command, that it doesnt find the cache files. And yes all the cache is located on a mapped network drive. all render nodes have the same drives mapped. Are there known problems of cache being located on network drives? if so how do you get around that? again thanks for every one helping out.
Hi Guys. I know it's been a year, but I was just hitting the same wall with my head.
Basically, the same situation as you describe. Cached fluids, a renderfarm, and runup hell.
Ok so here we go...
- The solution is actually a hack.
- It works only for mental ray.
Fixing the runup problem involves editing the renderDesc XML file for mental ray - you can find it in your Maya installation folder/bin/renderDesc
The file name is mrRenderer.xml
ATTENTION! You MUST make a back-up copy of the original file before editing of course so you can revert if anything goes wrong.
There is a line that goes something like this (it was line no. 31 in my version)
meltrailer s='dynRunupForBatchRender("mentalrayGlobals"); miSetImageDir($imgDir); ...........
Basically, what you want to do is disable the dynRunup command. You can comment-it-out like this:
meltrailer s='/* dynRunupForBatchRender("mentalrayGlobals"); */ miSetImageDir($imgDir); ...........
From now on you can enjoy rendering your cached fluids without runups.