attachment points and graphic gain

sister

Co-founder TFF Italy
I have a doubt : take an object, say a station, which, p.e., has a lot of benches or clocks. how is the gain for the graphic motor if I attach all of them using attachment points calling a single mesh instead of put them one by one in the 3D?
I remember in the old versions there was a big difference doing that. What about now? is it still valid?
thanks in advance for your kind help.
 
I'm not exactly certain what you are asking but each separate mesh has a 300 poly equivalent overhead when loading, and there is a 200 poly equivalent for each texture. I think this is true whether or not you use an attachment.

If you have the sources building the entire station or most of it as one asset would reduce the load on the game.

Cheerio John
 
I mean in a station instead of to have 40 benches in the station 3D you have one bench "called" 40 times by the station through the points. I'd be not so sure about the weight of the whole station. I think the RAM take in charge just one bench in that way even if it will be "called" 100 times. In the "normal" way (whole 3D with all the 40 benches) the total count of the poly should be greater. or not?
 
Not sure - there are still the same number of polys to render. I guess you make a (very) small saving in memory in terms of total mesh filesize, but I suspect that that will be more than cancelled out by the extra processing power required to load a second mesh file. If the material isn't shared between the main mesh and the bench, then you also have an additional material to load.

Unless you expect the bench to be used as an attachment, or stand-alone, in several places, I'd say it's better to make it all part of the same mesh. The only exception to this is if the mesh is getting too big for GMax to export reliably.

Paul
 
Back
Top