Split irradiance map between multiple machines script?

[visualride]

I'm doing a 5000 frame animation of an exterior fly by a neighborhood. To save time I want to render every fifth or tenth frame of the irradiance map first with "incremental add to current map". This will still create a massive file (200-400kb) that I will be too memory intensive later. If I want to split up the irradiance map between multiple machines, such as give each machine 500 frames then I'm wondering if anyone has come up with a script to automate breaking up this job and sending them out to Backburner.

I'm not an expert with Backburner yet, so assigning a range of frames to specific render nodes isn't much fun. Anyone have much experience with this workflow?

[natty]

Hi

What we normaly do is split the frames up per pc... and send it over backburner for incemental add. so if you have say 5 pc's each comp will calc its own irr map and save it down and then merge the maps. then you can render from file over backburner the complete animation.

[visualride]

I just did a little test with backburner, and now I'm fine with assigning ranges of frames to individual machines.

Previously I've just assigned the same range of frames to the ir map that was created with that range before, and that seemed to work fine.

Natty: How do you merge your IR maps? When I've tried to merge irradiance maps over 100mb imapviewer crashes.

[natty]

To be honest my maps dont tend to get that large. we've just done a 8000 frame ani and the total irr map size was 80megs. origanaly split by 7 cams.

[visualride]

Thanks Natty, I think I've got a handle on it now. I guess that I could use a few more cameras, but the client just wants one slow pan.

[natty]

well if its a slow Pan try 20nth frame rather than 5th ... that should reduce the irr map size... ... also if you arnt already, use Lightcache in flythrough mode to speed up the irr map..

[visualride]

I'll try cutting them down to every 10th frame. I've got a 3 day weekend and a whole office (30-40 pc's) of render computers.

I've been using QMC for my secondary bounces. So far I've had good results at livable speed. Going by what I saw on Chris NICHOLS DVD, I'm waiting to do some interior shots before diving into using light cache.
Do you use light cache for exterior shots? Does the end result differ?

Thanks for your input Natty.