float vs fixed

by: nephi_johnson
Status: Offline
Joined: 2012-07-04

I've recently switched all of my matrices and vectors in my game engine to use the float versions instead of the fixed versions (CIwFMat vs CIwMat, etc). I did this because I've started using the Bullet physics engine where everything is a float and I didn't want to have to do IW_FIXED conversions all over the place.

Is there a performance boost to only using fixed matrices/vectors in IwGX, as opposed to using floating point matrices and vectors? I know I'd have to cast any floats coming from the physics engine using IW_FIXED (and convert fixeds to floats using the other #define - forget the name of that one right now though).

Is there a general practice for this?

Thanks

by: roo
Status: Offline
Joined: 2012-03-23
At the final step all

At the final step all vertices would be converted to int16 even if you use float point.

Marmalade 6.1 will have GX2 with float point
Right now the only solution is to use IwGL.

Otherwise you will have problems like these
http://developer.roolez.com/2012/07/03/marmalade-gx-artifacts/

Latest Posts

Latest Comments

Top Contributors

  • amanda
  • riaan
  • alexand3
  • alexand4
  • alexand7
  • alexand889
  • alexand12191
  • remote9991
  • remote9992
  • test