particle disckCache, wtf?!

[mattswilson]

any cached data usually prefers to be in the project path. Particles and Fluids, in my experience, are notorious for breaking their links quickly.
If there is a manual override for the working project in Smedge, try that. It should enter the -proj command into render string. Also, make sure that the working directory was created with the workspace.mel file. That will have the exact string that you will want to point your rendered to. As long as that flag is set, and it is looking in the right place, your render should work. I have had a few problems with renderman for maya still not being able to render them if you specify an alternate name for the particle caches when you initially create them.

[azshall]

any cached data usually prefers to be in the project path. Particles and Fluids, in my experience, are notorious for breaking their links quickly.
If there is a manual override for the working project in Smedge, try that. It should enter the -proj command into render string. Also, make sure that the working directory was created with the workspace.mel file. That will have the exact string that you will want to point your rendered to. As long as that flag is set, and it is looking in the right place, your render should work. I have had a few problems with renderman for maya still not being able to render them if you specify an alternate name for the particle caches when you initially create them.

Yeah, I've hardcoded the command line project through smedge. The problem is this ..

I go to the network box ... say '\\networkBox\path\project\scenes\file.ma' .. right? ... I load this up, its no good.. particles are not found.

Now, say I VNC or Remote Desktop into "networkBox" .. and COPY that '\\networkBox\path\project' .. to tsomewhere on the local drive, say C:\temp\ .. whatever .. then load teh scene? .. It finds the cache perfectly.

For some reason, it just will not, find my cache if its loaded on the network. :(

az

[mattswilson]

are you using shared storage via a mounted drive?

[azshall]

I'll have to ask, I'll find out and post a reply ...

[azshall]

I can't find any rhyme or reason as to why the cache works on some machines and not others, over the network .. and not over the network.

Maya's PDC is a piece of shit, i'm beginning to think.

[mattswilson]

you know I have had that problem for a really long time. The only way that I have ever gotten around it is being absurdly cautious with drive structures, projects and that workspace.mel file.

try creating a shared directory that all of the machines can see. Map that as a network drive and make a folder in there called Data. Then hardcode that path as your PDC directory in that workspace.mel file and try rerendering. I would say just copy all the previously created caches from every directory you can find into the new one.

[azshall]

Good suggestion, I'll try that out as well ... Just so frustrating that a render that I launched at midnight, didn't even render properly because Maya's inability to keep track of a particle cache that exists properly. How annoying.

Its wierd, I would think there would be a node that ties that specific cache to the particles, like a fluid cache is connected to the fluid container itself.. With particles, how do you know what cache goes to what? Its reading it SOMEHOW and SOMEWHERE, just can't find the connection ..

az

[mattswilson]

i agree, its a serious pain! I think path assignment like fluids would be a resonable request for Maya v20. Maybe well get it sooner:)

[azshall]

the more I poked around here.. I found an interesting thread going over this exact topic ..

http://forums.cgsociety.org/showthread.php?t=367871&highlight=cache

I tried one of the suggestions which was to change in my workspace.mel .. the command that says ....

2)
open your workspace.mel and change the line:
workspace -rt "particles" "particles";
to
workspace -fr "particles" "particles";

Seems to have actually done some good, I'll be quick to try the seed option as well if things start to act funky after this .. but so far it seems to work... Need to try it on the farm though ..

az

[alexx]

i dont even dare to check if that bug is fixed in 8.0 :)

i really hope they have rewritten the complete cache sections for particles/hair/fluids because the were more bug than feature

good luck

alexx

[azshall]

Sheesh, someone just needs to be a nice lil developer :) .. and write a new particle cache engine... That actually will designate a "cache" to individual particles so you can actually cache multiple particle objects independantly and have them read... *sigh* .. I actually outlined a project like this that would work awesome. Unfortunately for me though, I only know MEL, not C++/C# :( ... Oh well ...

az

[pascal]

Yes, a new cache system would be cool.

Anyone ever checked that file format: http://hdf.ncsa.uiuc.edu/HDF5/ ?
From what I understand, it is pretty efficient for this kind of data.

I think it could be "somewhat" easy for an experienced programmer to just create a new emitter type that would read such file, and a command to create it...

If anyone come accross such thing, tell me ;)

Pascal.

[CGTalk Moderation]

This thread has been automatically closed as it remained inactive for 12 months. If you wish to continue the discussion, please create a new thread in the appropriate forum.