GL Test

BlitzMax Forums/BlitzMax Programming/GL Test

Perturbatio

(Posted 2005) [#1]

I was beginning my foray into GL programming and ended up with this as my first test program:

Const SCREENWIDTH = 800
Const SCREENHEIGHT= 600
Const HALFSCREENWIDTH:Float = SCREENWIDTH / 2.0
Const HALFSCREENHEIGHT:Float = SCREENHEIGHT/ 2.0

bglCreateContext(SCREENWIDTH, SCREENHEIGHT, 32, 1000, BGL_FULLSCREEN|BGL_BACKBUFFER|BGL_DEPTHBUFFER)
bglSetSwapInterval 0
glShadeModel(GL_SMOOTH)


Global QuadList:TList = New TList

Type TQuad
	Field X : Float
	Field Y : Float
	Field Width : Float
	Field Height : Float
	Method Update()
		X = X + RndFloat()-RndFloat()
		Y = Y + RndFloat()-RndFloat()
	End Method
End Type

SeedRnd MilliSecs()

For Local t = 1 To 10000
	Local tempQuad:TQuad = New TQuad
		tempQuad.X = Rnd!(0.0, SCREENWIDTH)
		tempQuad.Y = Rnd!(0.0, SCREENHEIGHT)
		tempQuad.Width = Rnd!(16,64)
		tempQuad.Height = Rnd!(16,64)
		QuadList.AddLast(tempQuad)
Next

Local LastTime:Int
Local TimeDiff : Int
Local StartTime:Int
Local timecount:Int = 0
Local counter :Int =0
HideMouse()



While Not KeyDown(KEY_ESCAPE)
StartTime = MilliSecs()

	
	glClear(GL_COLOR_BUFFER_BIT | GL_DEPTH_BUFFER_BIT)

	For Local a:TQuad = EachIn QuadList
		a.Update()
		DrawQuad(a.X, a.Y, a.Width, a.Height)
	Next

	bglSwapBuffers()


TimeDiff = MilliSecs()-StartTime
TimeCount :+TimeDiff
counter :+ 1

Wend



ShowMouse()


Print timecount
Print counter
Print TimeCount/counter +"ms/frame"
Print "FPS: "+1000/(TimeCount/counter)

bglDeleteContext()

QuadList.Clear()
QuadList = Null


Function DrawQuad( qX:Float, qY:Float, qWidth:Float, qHeight:Float)
	qX = (qx-HALFSCREENWIDTH)/HALFSCREENWIDTH
	qY = -((qY-HALFSCREENHEIGHT)/HALFSCREENHEIGHT)
	qWidth = qWidth/HALFSCREENWIDTH
	qHeight = qHeight/HALFSCREENHEIGHT

	glBegin(GL_QUADS)
		glColor3f(1.0, 0.0, 0.0)
		glVertex3f(qX, qY - qHeight, -1.0)
		glVertex3f(qX + qWidth , qY - qHeight, -1.0)
		glColor3f(1.0, 1.0, 0.0)
		glVertex3f(qX + qWidth, qY, -1.0)
		glVertex3f(qX, qY, -1.0)
	glEnd()
End Function

WARNING!: Bright colours contained within, may well nauseate.

I get a constant 10ms/frame (100fps) regardless of whether I use 16bit mode or 32 bit mode.
Tested the same on my Dad's PC while I was at his house, he has an Athlon64 3200 and a Radeon Atlantis (9800 128MB I think).
He gets 8ms in 16bit mode and 16 in 32bit mode.

Could some of you guys test this and tell me what your specs are and what you get?

Floyd

(Posted 2005) [#2]

I get

5124
198
25ms/frame
FPS: 40

which seems about right, one third the speed of a recent model PC.

N	(Posted 2005) [#3]

After you call bglCreateContext you don't need to set the swap interval. It does that on its own.

4924
335
14ms/frame
FPS: 71

AMD 64 3000+ 2ghz, 1gb ram, ATI X800 XT 256mb.

rogue

(Posted 2005) [#4]

On mine:

13421
727
18ms/frame
FPS: 55

Pentium 4, 2.66GHz, 1GB RAM, ATI Radeon 9000 64MB, ATI VER008.004.010.000

JoJo	(Posted 2005) [#5]

21018
617
34ms/frame
FPS: 29

PII 400Mhz 256 ram geforce 4mx 440

Garred

(Posted 2005) [#6]

In not debug mode I had:
2613
231
11ms/frame
FPS: 90

In debug mode:
3969
246
16ms/frame
FPS: 62

On my AMD 1600, 512 Mb Ram, NVIDIA GeForce FX 5600 Ultra 128 Mb, Windows 2000

Perturbatio

(Posted 2005) [#7]

Thanks guys :)

taumel

(Posted 2005) [#8]

AndyBoy_UK

(Posted 2005) [#9]

15889
49
324ms/frame
FPS: 3

on a Mobile Pentium4-1.8ghz, 1gb ram, NVIDIA GeForce2GO 32mb (oh dear)

daqx	(Posted 2005) [#10]

Pentium 4 3,0 GHz HT
1 GB DDR RAM
Radeon 9800 Pro (128 MB)

11519
1418
8ms/frame
FPS: 125

StuC	(Posted 2005) [#11]

@Perturbatio

Neat little test, but one point for consistency - you should really seed the random number generator with a fixed value (e.g. 1), rather than MilliSecs(), to ensure the same 10,000 squares are generated each time. I know you intended this as just a quick test, but thought I'd point that out.

Cheers,

Stu

Perturbatio

(Posted 2005) [#12]

yeah I noticed that, but couldn't be bothered to change it, the results seemed consistent enough on my system to suggest that they would be good enough on other's too.