Jump to content

  • Log In with Google      Sign In   
  • Create Account


Is GL_NORMALIZE still slow?


Old topic!
Guest, the last post of this topic is over 60 days old and at this point you may not reply in this topic. If you wish to continue this conversation start a new topic.

  • You cannot reply to this topic
3 replies to this topic

#1 SillyCow   Members   -  Reputation: 845

Like
0Likes
Like

Posted 03 February 2013 - 06:05 PM

Is glEnable(GL_NORMALIZE) slow on modern hardware?

I read the opengl pitfalls article which suggests that enabling this option slows down vertex processing.

The article is from 2000.

In 2000 there were still some low-end cards with S/W only vertex processing...

So, is this still a problem?

 


My new android game : Enemies of the Crown

My previous android game : Killer Bees


Sponsor:

#2 claesson92   Members   -  Reputation: 406

Like
1Likes
Like

Posted 03 February 2013 - 07:21 PM

GL_NORMALIZE is the slowest normal transformation, but i wouldn't say it's a problem.



#3 shwasasin   Members   -  Reputation: 482

Like
2Likes
Like

Posted 03 February 2013 - 08:41 PM

One thing to note is if you're targeting OpenGL (desktop) that the GL_NORMALIZE target (for glEnable/glDisable) was deprecated in OpenGL 3.2 Core (July 2009). It won't be available any longer in Core profiles in 3.2 or later versions. Just an FYI. :)



#4 mhagain   Crossbones+   -  Reputation: 7453

Like
2Likes
Like

Posted 04 February 2013 - 02:50 AM

First be aware of what enabling GL_NORMALIZE actually does and what circumstances you may need it in.  If you never scale your matrix, you don't need it.  If you do scale but it's uniform (i.e. x, y and z are scaled by the same amount), you don't need it - use GL_RESCALE_NORMAL instead.

 

So having determined that you do need it, is it slow?  Well, it's certainly going to be slower than not using it, but that may not necessarily translate into "slow" - in other words, using GL_NORMALIZE may not suddenly throw you down to single digit framerates.  If you can still hit your target performance when using it, then you can hardly call it "slow", can you?

 

And finally, you've tagged your question "OpenGL ES" but you're not saying which version.  In 2013 you may well be lighting per-fragment, and renormalizing in your fragment shader anyway, in which case none of this discussion is relevant.


It appears that the gentleman thought C++ was extremely difficult and he was overjoyed that the machine was absorbing it; he understood that good C++ is difficult but the best C++ is well-nigh unintelligible.





Old topic!
Guest, the last post of this topic is over 60 days old and at this point you may not reply in this topic. If you wish to continue this conversation start a new topic.



PARTNERS