Hello Marc,
On 22/05/2012 10:44, Marc Gutowski wrote:
okay the rotos are big in textsize, but nuke seems to "evaluate" or cache
every roto/rotopaint on script load. doing this on render time would
save a lot of
time and ram when loading heavy scenes files - but this is only a guess.
The Roto & RotoPaint nodes keep a full text copy of their to_script to
cache for fast saving. This has the downside of taking up a lot of
memory if you have a big script.
We're working on making this much smaller for a future release.
Aside from that it also has to allocate the entire data structure for
each Roto/RotoPaint node. It could be anywhere from 3 to 10 times as
much extra memory.
So, if you have one Roto node taking up 5MB in script, it will take,
let's say 5 + 10 * 5 = 55 MB in memory. Multiply this by the 500 Roto
nodes you have, that's 27500 or 26GB worst case.
Of course this highly depends on how much animation you have in your
nodes and a variety of other variables.
Again, we're working on making the Roto & RotoPaint node more memory
efficient as well as quicker for a future release.
HTH,
Wouter
--
Wouter Klouwen, Software Engineer
The Foundry, 6th Floor, Comms Building, 48 Leicester Sq, London WC2H LT
Tel: +442079686828 • Fax: +4420 79308906 • thefoundry.co.uk
The Foundry Visionmongers Ltd • Reg.d in England and Wales No: 4642027
_______________________________________________
Nuke-python mailing list
[email protected], http://forums.thefoundry.co.uk/
http://support.thefoundry.co.uk/cgi-bin/mailman/listinfo/nuke-python
