best of graphics cards and test methods for game engine

Game Engine Game Engine Support and Discussion
1

best of graphics cards and test methods for game engine

	hi!
	
	Witch kind of MANNERS there are to testing Blender Game Engine
	wint graphics cards and to confront with oter OS and hardware
	with other.
	
	for example an idea:
	
	A 'GLOBAL' gameEngineTest.blend file that
	everyone test on own computers and 
	switch on Game menu  --> Enable All Frames
				 Show frame rate and profile
	
	and then contemplate the parameters like frame
	rate,Physics,Logic etc. and write in the forum.
	
	for example parameters:
	
	OS (opreation System):
	video card types and parameters:
	memory:
	processor:
	and other....
	
	
	Swap:......(frame per second)
	Physics:......
	Logic:......
	and more
	
	
	In gameEngineTest.blend file let...?
		
	N number of polygon count...?
	N number of textures...?
	N number of LogicBricks...?
	
	any idea?
2

enable all frames shouldnt’ have a speed difference (by framerate) however the motions may move faster or slower depending on if your framerate is faster or slower than 50fps (enable all frames acts as if 1/50th of a second passes each frame, regardless of the actual amount)

wait, are you proposing the creation of just one file?
you mean like http://home.earthlink.net/~nwinters99/BlenderStuff/pluginPlay.blend ? (1.75Mb)

though it doesn’t require much physics, that one really requires a good graphics card, and working drivers

WE MUST DECIDE ON A PARTICULAR RESOULTION AND TESTING SITUATIONS TO TEST THIS STUFF IN
I was thinking 1024*768 fullscreened, graphics settings (for video card) at highest, except with antialiasing and anistropic filtering off.

(seems my ati radeon card is a bit faster with that file than my geforce 2 was)

3

its better to assess the speed of the graphics card using proper software like 3d Mark rather than blender.

4

blender has been used to benchmark machines before (using the draw timer, control+t)

also, many compainies optimize for 3d Mark, and the like, often to the point of cheating (lowering aa/af values only for 3d mark, running different shaders with re-ordered instructions, changing clip planes…)

they haven’t done the same for blender (especially ati)