[deadalvs]

I had it working using six computers a few of times, but to set it up each time was no easy task, far to complicated when it does bot really need to be

could You tell me where to put the texture files so the slave can find them? this would be great!

* * *

deadalvs

[ajlynn]

Try this:

Make a directory on your main computer that's shared on the network. Use "map network drive" to assign the share the same directory on all computers (including the one it's on). E.g. I have a main computer called spuybroek, and the directory c:\netrender is shared as \\spuybroek\netrender, and is mapped to q:\ on all computers including spuybroek.

Use this mapped drive as your working directory, and use the drive letter as the reference - so if I have a subdirecory in there called maps, when I set up my materials I work out of q:\maps (instead of c:\netrender\maps - on spuybroek they mean the same thing but on the other computers they don't). I do the sam thing for saving scenes. (I haven't done this with Maxwell recently so I don't remember where you have to set the output image file, but my suspicion is that it is the only file that does not go in q:\.)

This is what I do with net renders in general and it generally works. Also works well with things like Autocad XREFs, Indesign file links, making it so you can easily switch between computers when working on Max files - any time you want one space on one hard drive to look exactly the same to more than one computer.

[aitraaz]

Can you imagine getting close to a deadline, and as your 4-day NET render finally draws to a close it tells you "render failed?"

Umm, yes, actually I can...

[deadalvs]

i could fix the texture path with the temp folder...

now, i «only» have two more problems.

1) how should the imagepath be called in maya? because the image name is just the file PREFIX, but the rest of the path can't be edited...

2) how can i be sure that the computers work together on the one frame?

i made a short test and one said the rendering is finished, the other one said when it was finished, the image could not be saved because of the path (see 1) ). and it was saved anyway... i don't get how this works.

the strangest thing is that the SL reached (5 min render) was even lower with 2.73 with both computers together comparing to SL 3 on just one computer... (they're about equally strong !)

any hints?

* * *

deadalvs

[deadalvs]

and why, after finishing, does it restart to render on the slave?
because it couldn't save the frame?

* * *

deadalvs

[deadalvs]

yeah now i'm sure they just render the same image twice...

* * *

they make exactly the same SL steps... with one or two computers and about the same times...

how can they be teamed up? (no, it's not animation mode!)
has this to do with hyperthreading? i think one of my machines can't handle hyperthreading...

*help*

* * *

deadalvs