# Is there a limit?

Started by Oct 10 2001 12:01 PM

,
9 replies to this topic

###
#2
Members - Reputation: **1087**

Posted 10 October 2001 - 12:44 PM

No, there shouldn''t be in any complient set of drivers. Polygon''s are ''bad'' though, they''re slower for most situations. Stick to triangles, quads, and their derived forms.

[Resist Windows XP''s Invasive Production Activation Technology!]

[Resist Windows XP''s Invasive Production Activation Technology!]

###
#4
Members - Reputation: **122**

Posted 10 October 2001 - 01:58 PM

(Null and Void), I''m hoping you don''t mean to say triangles and quads aren''t polygons? I guess that instead you can say.... the use of N-gons serves a role in determining dynamic (and perhaps static) efficiency, where smaller values for N correspond to faster rendering times. Recall though, that as N goes to infinity, you''re simply plotting a circle........which in most cases is just as efficient as an N-sided polygon with a low-valued N. This seems paradoxal, but remember that when calculating the plot for a circle, the GPU takes into consideration the formula for drawing out such shapes, which demonstrates the use of (mostly) preprocessed primitives. This contrasts with drawing a shape with a finite number of sides.

Since this deviates greatly from the question at hand, I''ll shut up now

Btw, such limits depend on available memory (allocated space), but for all practical purposes there is no limit.

Since this deviates greatly from the question at hand, I''ll shut up now

Btw, such limits depend on available memory (allocated space), but for all practical purposes there is no limit.

###
#6
Members - Reputation: **2076**

Posted 10 October 2001 - 02:08 PM

quote:Original post by HyprLogik

Recall though, that as N goes to infinity, you''re simply plotting a circle...

Just niggling, but you''re actually plotting an ellipse, of which a circle is a special case (where all interior angles are equal).

###
#7
Members - Reputation: **1087**

Posted 10 October 2001 - 02:35 PM

I was talking about the GL_POLYGON rendering modes in general.

[Resist Windows XP''s Invasive Production Activation Technology!]

[Resist Windows XP''s Invasive Production Activation Technology!]

###
#8
Members - Reputation: **122**

Posted 10 October 2001 - 02:56 PM

quote:Original post by Oluseyi

[quote]Original post by HyprLogik

Recall though, that as N goes to infinity, you''re simply plotting a circle...

Just niggling, but you''re actually plotting an ellipse, of which a circle is a special case (where all interior angles are equal).

Damn, Don''t remind me of calculus ! As my head approaches its limit, my brain wants to explode.

###
#9
Members - Reputation: **100**

Posted 10 October 2001 - 04:53 PM

Another problem with polygons which are not triangles is that after some transformations, the N points may be not set on the same plane anymore (due to floating points inacuracy) which may result in strange rendering effects.

Triangles, however, always defines one plane (unless the three points have the same coords).

Triangles, however, always defines one plane (unless the three points have the same coords).

###
#10
Members - Reputation: **122**

Posted 10 October 2001 - 08:51 PM

quote:Original post by Oluseyi

[quote]Original post by HyprLogik

Recall though, that as N goes to infinity, you''re simply plotting a circle...

Just niggling, but you''re actually plotting an ellipse, of which a circle is a special case (where all interior angles are equal).

Nice note, but such semantics are of little relevance to this situation. To affirm the specific classification of geometric figures in terms of categorical subsets doesn''t lend significance to the matter. But here I go again, and I''m being a hypocrite

By the way, my life is like a polygon: there are many sides to how I present myself; my days are rendered at the waking hour, my position changes throughout my circadian trek, I am constantly culled at inconvenient times, and at the end of the day I am clipped, overwritten, erased from memory, and soon forgotten... (or not)