PDA

View Full Version : LWBrush speed



Dennik
04-20-2017, 01:02 PM
I have a question about LW Brush from 3rd Powers.
Does it's speed diminish depending on how much total geometry is loaded in Modeler?
And by total I mean on other layers or objects.
Can anyone do a little test and let me know?

Norka
04-20-2017, 02:47 PM
Been a little while since I used LWBrush, because I have been swamped with stuff that doesn't require those tools... but I just did a quick test for you, since I don't think I ever used LWBrush on anything higher than a couple thousand polys. So I took a box and subdivided it to 24k or so polys, did some Push/Pull, Slide and Grab on it, and it just seems a little sluggish. Then I subdivided it again to over 90k and it was very slow (on my i7 5820k w/ 32GB). So for higher-poly stuff, it is a snail compared to ZB, and even SculptGL (webgl app). But it rocks for normal low-medium meshes, as does all the 3rdPwrs plugs. Hope this helps.

Sensei
04-20-2017, 03:48 PM
Does it's speed diminish depending on how much total geometry is loaded in Modeler?

It's what you would expect from any kind of tool which is operating on the all polygons/points.
The only question is whether it's linear or geometric slowdown.

Not to even mention that Modeler app has to upload new modified geometry to gfx card, every time you move mouse.

ZBrush has very cheap architecture in comparison to LW. They can do a lot of cheats and shortcuts.

Ryan Roye
04-20-2017, 05:27 PM
LWBrush performance depends on various factors:

- Total geometry of the mesh
- How much geometry LWBrush is operating on (smaller the brush, the faster it'll go up to a point)
- Maintain Surface Detail Toggle (Disabled is significantly faster)
- Fix points out of range (This is tied to the maintain surface detail option. Enabled isolates the surface area to only that which your brush covers instead of taking the whole model into account)

Assuming maintain surface detail is off (or fix points out of range is enabled), performance is generally very good up to around 50,000 polygons, after which things will get choppy and eventually you'll need to split up the model into parts to gain performance.

Norka
04-20-2017, 05:28 PM
Well, if GPUs are a factor, I have three 980Ti and the display one is at 8X, which is none too shabby.. so this corroborates LW as the culprit, since it can't be my display GPU. By no means is this an excuse to not get 3rdPwrs plugs. They shake their damn money-makers every time I am able to accomplish crap in LW and forgo jumping to ZB etc.