Benchmarking quake4

chump

I did some benchmarking with quake4... you have to run with +set com_allowConsole 1 command

System is as here: http://www.boards.ie/vbulletin/showthread.php?t=316401

playNetTimeDemo "ggl-timedemo.netdemo"
As referenced from http://www.ggl.com/view_forum_topic.php?TopicId=769

Standard: High quality, all settings max (or ON), except AA off.
FPS results below.

640x480 - 46.44
800x600 - 44.5
1024x768 - 44
1280x1024 - 37.01


Downloaded new drivers that apparently add extra support for dual processors...
81.87_forceware_winxp2k_english.exe

800x600 - 48.97
1280x1024 - 44.28

A pretty decent increase as you can see...
Other settings tried:
1280x1024 - low settings, vsync off, high qual off - 47.98
1280x1024 - ultra quality, all on - 37.94

Likely to use settings playing quake 4:
800.600, medium quality, with autoCFG - got 55.87 fps...

I would point out that this demo is a bit crappy, and I actually noticed a huge in-game (single player) increase in my actual fps... so these drivers definetely are worth the punt! I used the above (in bold) settings in single player... pretty much constand 60fps playing


autoexec.cfg contains

set com_showFPS "1"
set com_fixedTic "-1"


// G R A P H I C S
set r_swapInterval "1" //vsync on/off - keep it off
set r_fullscreen "1" //fullscreen or window mode (can alt-enter for window mode toggle)
set r_brightness "1.35" //default 1.2
set r_gamma "1" //default 1
set g_fov "110" //Field of View, default 90
set r_finish "0" //0 is faster
set r_ignoreGLErrors "1" //keep at 1
set r_useOptimizedShadows "1" //faster shadows
set r_useTurboShadow "1" //faster shadows
set r_useStateCaching "1" //1 is faster
set r_alphaToCoverage "1" //unknown
set r_useDepthBoundsTest "1" //NVIDIA CARDS ONLY 1 is faster
set image_downSizeSpecular "1" //enables Specular downsizing setting
set image_downSizeBump "1" //enables Bump mapping downsizing setting
set image_downSize "1" //use to enable image downsizing setting
set image_downSizeSpecularLimit "256" //set same as graphics memory for optimal
set image_downSizeBumpLimit "1024" //1024 for 256mb gfx. 512 for 128mb gfx. 256 for 64mb gfx lower is faster
set image_downSizeLimit "256" //set same as graphics memory for optimal
set image_lodbias "-1" //appears broken somewhere, use -1
set g_decals "0" //bullet holes on walls and other effects etc, on/off
set g_muzzleFlash "0" //default 1
set g_brassTime "0" //ejecting spent ammo, 0-1, 1 default
set g_projectileLights "0" //remove lights from projectiles, 0 is faster
set g_skipItemShadowsMP "1" //remove item shadows - MP only, default 0
set g_skipPlayerShadowsMP "1" //remove player shadows - MP only, default 0
set g_showPlayerShadow "0" //player shadows on/off, default 1

Illkillya

Those results sound very low... what system do you have?

chump

Did you actually run any benchmarking yourself?
It's a beast of a game ...
Results are actually not bad at all.
You'd see my specs if you read my original post properly.

chump wrote:

I did some benchmarking with quake4...

1280x1024 - 44.28

&

Likely to use settings playing quake 4:
800.600, medium quality, with autoCFG - got 55.87 fps...

are actually quite good...

You'd actually have to do the timedemo yourself to see really...

airetam_storm

chump wrote:

Standard: High quality, all settings max (or ON), except AA off.
FPS results below.

1024x768 - 44

Would they not be the optimal settings? 44frames is decent enough

Giblet

chump wrote:

Did you actually run any benchmarking yourself?
It's a beast of a game ...
Results are actually not bad at all.
You'd see my specs if you read my original post properly.



are actually quite good...

You'd actually have to do the timedemo yourself to see really...

That's pretty low for those specs. Killys machine hammers that easily.

And if these frames aren't constant it's a load of muck.

chump

Giblet, how can you say that having not run the test?

It's ridiculous that your commenting on timedemo fps results when you haven't even run a timedemo ... it's a multiplayer demo...

Do you not understand how it works?

A fair test means all people run the same test...

The results actually aren't that bad when compared to other people who've run the test, see link provided in oringinal post...

Why don't you run the test, and post your scores?

I posted up the results to see what other people get and compare ... maybe tune things up a little

Uploaded the demo on http://www.iol.ie/~yoyoyo/ggl-timedemo.zip to make it easier for yas to DL and show off yar fine results....

So if any1 is up for it... using this demo, standard autoexec.cfg file, all settings on High and all advanced options ticked except AA, what do yas get for the following resolutions...

640x480 -
800x600 -
1024x768 -
1280x1024 -

REM to set +set com_allowConsole 1 in shortcut to quake4, and for best results close all background apps...
Oh and you put the demo in a demos subdirectory in q4base directory

and then run
playNetTimeDemo "ggl-timedemo.netdemo"
from console in game

Giblet

Haha, I know more about it than you thanks, and those results are ****e. Who gives a crap about the quality in a mp game. I get 60fps constant on most maps in mp on a ****ing barton 2500 and a 9600pro with my config.

Did you run each timedemo twice to allow all the stuff to cache?

chump

ah giblet, muppé...

Why don't you run this test then big boy?
Run it twice and post your scores?

and tbh it doesn't matter a damn what you get with your config...
sure i used to get 4million fps with the d_mipcap d_mipscale settings in q1... you can make any game run faster if you wanna strip it to its bones

Giblet

OH WOW I BETTER POST MY SCORES BECAUSE THEN YOU'LL WIN!

You seem to be missing the point. Timedemos matter ****, it's just numbers. I don't play on those settings so what difference does it make unless you want to prove something with the PC you bought? Again, Killys machine is better so when he posts an OMG bigger number, what difference is it going to make?

You want to tweak to make it better? Then why run the timedemo at those settings and find a config elsewhere? Timedemos suck because they average frames out over a long time and while the high number might look nice it doesn't matter ****. Walk around the level yourself with showfps and adjust if it's annoying you in a game setting.

Also the link is being raped, so I can't dl it :v:

chump

giblet your the one who came on here and started shouting how sh!te the results were and now your saying I'm saying I want to win etc.. etc... I don't know what your on tbh...

Running a timedemo is the clearest way to see what kind of performance you're getting and how tweaking is improving that...

You take a timedemo, you run it, and you try to improve it by changing settings.
You can also see if your computer is performing up to scratch by comparing it to similar spec systems, and if something is amiss you look at the computer setup...

Sure I hardly expect my computer to come close to any top gaming rigs... it's a pentium D for god sake...
and I provided my own link for the demo ... http://www.iol.ie/~yoyoyo/ggl-timedemo.zip

and as is the case with any benchmarking thread, if you think the whole idea of putting up scores from timedemos is all about OMG WHO's got the BIGGER WILLY, and OMG bigger bnumber blah bljha blah why did you bother replying to it... shure that's what their all about

Giblet

All settings high. 35 fps @640x480
My config: 58 @640 x 480

Barton 2500+ @ 3200+
512mb twinmos ram
9600pro @ 500/350, 128mb ram.

chump

ya see tis a tough enough demo

anyway cheers for throwing up the scores, hopefully a few others will bother

Illkillya

All settings high: 86 fps @800x600
Thats with McAffee Antivirus hogware running, and mIRC and loads of stuff running, long uptime etc... might try it properly later.
I could overclock both my CPU and gfx card and get a huge boost in performance but cba because I never drop below 63fps, so why waste my time?

And chump, in fairness I don't think its a lot to ask you to repeat your system spec here. First I had to read through that really long post of yours, now you expect me to click the link and read through a whole load of long posts. I can't concentrate for that long.

The reason I said your fps seems low is because to me it does seem low, given that I get nearly double that, and I don't know the first thing about performance tweaking.

chump

Not too shabby a score...

Regarding my system, it was the 2nd line of my original post, and the 1st thing on that linked page so I dunno what yar talking tbh... and you didn't ask me to repeat it, you asked me what my system specs were... (just to be pedantic) :v: and ya didnt even list your own specs!

I'm quite happy with my performance, as in single player its maxed at 60 mostly with occasionaly dropping to 40 when it gets busy!

Prior Of Taize

Giblet wrote:

barton 2500

The barton is still a savage cpu though. thats a lot of the reason you can even run the game :v:

Fobia

Giblet: I'm running same gfx card, 512mb RAM, not as good a procesor as yours but surely I should still be able to get comparable fps in game? But I'm not, it's usually about 40, maybe it hits 60 while staring at a wall but it's rare, any ideas?

Rev Hellfire

I took a screenshot from giblets machine, he's reduced the graphics down as far as they can got, but it should give you an idea of what u need to set

Illkillya enters the arena

You are standing in a corridor, in the centre of the room slowly rotating is a rocket launcher.
The corridor runs north to south.

> Get rocket launcher
You pick up the launcher.

Illkillya exits to the south
>go south
You enter an open intersection with exist leading north, south, east west and up.
You see Illkillya.
Illkillya fires a railgun.
>evade
You evade the rail.
> fire rocket at illkilla
There is no one called illkilla at your current location
Illkillya fires a railgun
> fire rocket at Illkillya
Illkillya rails you for 12 points.
Illkillya evades your rocket.
> s

Giblet

My font is green!

Illkillya

amd 4000+ @2.4ghz, 1gig corsair 2-2-2-5 ram, geforce 7800gt

Mythago

Well.... just to give you all a boost... here's my score:

Medium setting @1024*768... a whopping 39 fps!?!?!?!

Need to take a serious look at my settings!

marta19

Shame....

Placebo

whats the optimal fps, what FPS should i be playing at ?

Placebo

marta19 wrote:

http://lolitas.eclub.lv

i clicked that in work ! :mad: :cool:

Illkillya

Placebo wrote:

whats the optimal fps, what FPS should i be playing at ?

The game is capped at 60fps - might show up as 63fps but you're still getting 60. So if you can get 60 fps constant then thats the best you can get.