Re: offtopic: quake3 source

[Boris P. <bo...@za...>]

Aapo Tahkola wrote:
> On Sat, 17 Sep 2005 14:05:11 +0200
> Rune Petersen <ru...@me...> wrote:
> 
> 
>>Bernardo Innocenti wrote:
>>
>>>Roland Scheidegger wrote:
>>>
>>>
>>>
>>>>>I've just noticed I have page-flipping enabled in my xorg.conf.
>>>>>I can't afford to crash my X session at this time, but I'll
>>>>>try disabling it later.
>>>>
>>>>Works just fine for me (x86, latest drm, latest Mesa, rv250,
>>>>pageflipping, color tiling, hyperz). My xorg is a bit "outdated" though
>>>>(one week old).
>>>
>>>
>>>Hmmm... I'm on x86_64.  Perhaps it's a bug in the
>>>DRM ioctl32 code?
>>>
>>>Neverball for x86_64 works fine btw... I shall try with
>>>a 32bit version too.
>>>
>>
>>You may also want to try this patch, which makes quake3 more stable on 
>>my R420 card.
> 
> 
> Really?
> Could you try if demos/lodbias works for you?
> 
> 
>>or you can try setting r_textureMode to "GL_NEAREST_MIPMAP_NEAREST" and 
>>try using 16 bit textures (r_texturebits).
> 
> 
> Fixes in this patch:
> -Invalid mipmap filters chosen
> -Fix and enable cube maps
> -Fix command buffer corruption caused by r300EmitBlit
> -Fix bad temp count for vsf(fixes lots of little problems with it)
> 
This is a really nice patch. I've been playing with Ogre3D and thought 
of trying all it's samples and sending the result and frame rates to 
dri-devel. After I saw this, I re-ran all the samples with the patch 
applied.
I ran all the tests on an athlon 2.5GHZ with radeon 9800 pro, after 
running fglrx first. I ran the tests in an 800x600 window in a 
1024x768x32 resolution.
Here's what I got:
Order is: test, opinion, average, maximum, minimum frame rates
If I don't write something, it means there was no change.

Bezier: OK, 97.9, 547.9, 6.4
After patch: 308, 541, 19

CameraTrack: OK, 204, 210.1, 70.6

CelShading: pretty bad- the cell is full of squares, 79, 79.8, 42.8
After patch: almost all white, only edges look OK

CubeMapping: only possible after patch: OK, 8.64, 10.86, 0.527

DeferredShading: It says "Your card does not support at least two 
simulataneous render targets, so cannot run this demo. Sorry!."
With fglrx this sample runs, but I didn't see anything except the fps

Dot3Bump: Refused running
After patch: Runs...detail by mapping:
	Specular: Can see black squares interchanged with the material
	totally Basic: all ok
	NormalMapped: a bit broken into triangles
	NormalMappedSpecular: white and red squares under any lighting
	bump mapping-multi light specular: all breakages possible in one
	bump mapping-single light: less breakage, some triangles,
				a little tv-like interpolation
	bump mapping-multi light: same as single light
	55, 340, 3	

DynTex: OK, 61.3, 62.1, 17.9

EnvMapping: OK, 251, 418, 74
After patch: 280, 580, 47

Fresnel: gives fragment shader errors, unknown fpi->Opcode 30,20,21,2

Grass: the cell looks pretty bad-red green and black squares with 
strange color in between, 44, 137, 21

Gui: OK, 4.88, 11.09, 0.78

Lighting: OK, 234, 529, 1

ParticleFX: OK, 164, 264, 1

RenderToTexture: OK, 4.8, 4.9, 0.7
Got a warning here:
*********************************WARN_ONCE*********************************
File r300_tex.c function r300TexEnv line 819
I am broken - Fixme !

SkeletalAnimation: OK, 164, 503, 50

Transparency: OK, 111, 238, 1.5

Water: broken textures-first texture is pure white instead of more water 
like transparent, 26, 33, 1
After patch: Much better, first texture fixed, water reflection seen, 
still some texture breaking into triangles.

I also got some warnings about using elt buffers.
There are 2-3 more samples, but they all looked fine and they're speed 
didn't really change after the patch.


Boris Peterbarg