LucasForums (
-   JO Mapping (
-   -   What lags more # of brushes or size of brushes? (

MightyM 08-19-2003 08:49 PM

What lags more # of brushes or size of brushes?
What lags more # of brushes or size of brushes?

I have a floor but most of the floor is covered by a second floor. Would it be more performance friendly to leave it as is 2 brushes 2 floors, or cut the area of the first floor that won't be seen? If i cut it i'll end up with 5 brushes 4 for the first floor and one for the second, but there won't be anything that's being drawn that doesn't need to be drawn... so which option will lag more? This is just an example as i've had to make this decision many times... e.g. on a railing: Do i make one big crossbeam that goes through the other beams or to i make lots of little cross beams to fit 'between' the other beams...? Anyone know which method is better? Thanks.


Lil Killa 08-19-2003 10:05 PM

As long as the end result will not lead to the brushes overlaping and making a z-fighting or "sparklies" effect just make them out of one brush like i do:D

wedge2211 08-19-2003 10:06 PM

Performance sag, I have learned, in direct proportion to the NUMBER of triangles the engine has to draw--not the size of those triangles. A rectangular brush face will consist of 2 triangles. So you're asking if it's better to have 10 tris, or only 2 tris but you won't be able to see parts of those tris. What I'd do is just keave the floor as one brush.

However, if you're using q3map2 to compile, the TJunctions process and MergeMetaTriangles process will pretty much ensure that the same thing will happen either way--I think.

In my opinion, you can be lazy here and leave it as one brush. Although performance won't really be affected one way or the other with just a few tris.

MightyM 08-20-2003 05:32 AM

sounds reasonable, but say we're talking like 64 or 100 triangles extra, would it be worth it to make say 10 brushes into 1 or 2 where possible?

Eldritch 08-21-2003 03:08 AM

Not really, unless you're pushing the tri limit.

I think that a good map shouldn't push more than 10k tris at most at any given time (someone correct me if I'm wrong here).

So in the grand scheme of things, even 100 extra tris isn't going to make a whole heck of a difference.

Emon 08-21-2003 03:11 AM

Back for Quake III, they said 30K tris for singleplayer, and 15K tris for multiplayer.

However, systems running JO and soon JA can handle much more. Around 40K or more tris is acceptable, and mutliplayer could hover around 20K to 30K.

Eldritch 08-21-2003 03:14 AM

Yeah, last I heard about tri count was back in the heady days of Q3, so...

Thanks for the update.

Emon 08-21-2003 03:24 AM


Your avatar is great. All you need to do is get a funny animation like mine, and it would be one of the best avatars ever.

Eldritch 08-21-2003 03:26 AM

Thanks. I love that show.

I suppose I could make an animated avatar, but my options are limited by the small 20k file size. Perhaps I'll make a small animated image sig of some kind.

All times are GMT -4. The time now is 08:54 PM.

Powered by vBulletin®
Copyright ©2000 - 2016, Jelsoft Enterprises Ltd.
LFNetwork, LLC ©2002-2015 - All rights reserved.