how to timedemo in Doom3?

[DefRef]

Originally posted by: DefRef

Originally posted by: DefRef
The rig: Athlon 3200+/1GB/128MB 5900NU at stock/WinXP SP1/61.77 drivers

Doom settings: 1024x768/Ultra Quality/everything one but AA plus my shadow's on

The punchline: 26.9 fps

For kicks, I tried the Ultra setting though it's supposed to be too much for any mortal video card but it's actually playable and looks great.

Just an update after plunking a PNY 6800GT clocked 400/1000 into the above rig:

1024 rez/Ultra/No AA = 46.2 fps (72% faster!)

1024 rez/High/4x AA = 45.5 fps (69%)

1024 rez/Ultra/4x AA = 41.7 fps (55%)

1280 rez/High/4x AA = 33.4 fps (24%)

1280 rez/Ultra/4x AA = 31.7 fps (18%)

Still wondering how BFG is getting a zillion fps (cg_UseTransparentZenStuff 1?), I ran 1600x1200, no AA, high detail and got 40.8 fps. Huh? Then, I unticked Vsync and suddenly I leapt to 52 fps. Sheesh. OCing to 404/1.1, it improved to 53.1 - no biggie.

Lesson learned: Vsync kills performance DEAD!

 

[BFG10K]

Allowing that you appear to be running at "Untextured Polys" LOD,

Where did you get such a notion?

(cg_UseTransparentZenStuff 1?),

I've never touched that setting.

Lesson learned: Vsync kills performance DEAD!

I'd wager that you still have quite a few things to check if you didn't realise vsync had such a large performance impact.

 

[DefRef]

Originally posted by: BFG10K

Allowing that you appear to be running at "Untextured Polys" LOD,

Where did you get such a notion?

(cg_UseTransparentZenStuff 1?),

I've never touched that setting.

Lesson learned: Vsync kills performance DEAD!

I'd wager that you still have quite a few things to check if you didn't realise vsync had such a large performance impact.

You know, you've always had a tendency to be an arrogrant jerk, but are you really contributing anything when you don't include details like what texture detail level your benches were run with or whether you turned off all the eye candy or what? (Making sneering comments while brushing the Cheeto crumbs off your Menudo shirt may make you fell all smug and superior, but you're still not contributing anything, are you?)

I knew vsync could be a problem in the past as cards could rip frame rates beyond the refresh rate of the monitor and thus would be held back by the need to wait for the monitor to redraw and led to benches of 60, 75, etc. fps instead of the 170 fps the card was capable of. My monitor runs @ 75Hz at 1600x1200, 85Hz @ 1280x1024, 100Hz @ 1024x768 and 140Hz at 640x480, so I figured as long as the card didn't render more than 85-100 fps, it wouldn't be a problem - looks like I was mistaken.

So, I made a goof and I'll admit it (could you? - since I'm sure you've sneered at many people from your lofty perch of perfection :disgust: ) but unless you'd care to suggest other settings that the stupid people should look at to match your awesomeness score, it may be best if you went back to writing captions for your Ricky Martin Photobucket gallery and let people help people around here, mmmkay? It must pain you greatly to have to walk amongst mere mortals.

 

[1ManArmY]

how do you change the field of view for soom III? I beleive the default FOV =90

 

[Bumrush99]

FYI: All D3 settings like shadows and other advanced video options were left ON, except for AA and vsync.

1024*768 0AA, 8AF High Quality, All Nvidia Options set to high quality, optimizations off, 65.62 drivers:
92.7 FPS

Same settings as above, w/AF/Triliniar optimizations, Nvidia Control Panel set to Quality to enable optimizations:
97.3 FPS

1280*1024 8AF, 0AA, HQ, No optimizations: 82.6
1280*1024 8AF, 0AA, HQ, Optimizations ON: 86.6


I play in 1024*768 8AF, with the optimizations and the game runs flawlessly.

 

[DefRef]

OK, something's screwy because my scores are insanely low in comparison to similar rigs. With a gig of RAM, a 2500+ running at 3200+ speed (2.1GHz) and a 6800GT OCed to Ultra speeds, WTF am I running at HALF speed?

(I suspect Bumrush99 will be a bit more helpful than some other users here.)

 

[Bumrush99]

Doom 3 timedemo is very odd with different settings.
Here are my suggestions to ensure that the timedemo is running correctly.

Make sure that in the Nvidia Control panel monitor refresh rates are NOT forced. Manually forcing refresh rates gave me a 12fps drop in 1024*768.

Make sure to set vsync to OFF, not application preferance.
Leave AA, AF to application preferance.
If you have dual monitor setup DISABLE it, this is a known bug that significantly hampers performance. Also, your 3200+ can not come close to touching a 64 system, my 3200+ overclocked to 2310 is a beast in comparison. I think you should be around 75fps with your rig.

 

[DefRef]

Originally posted by: Bumrush99
Make sure that in the Nvidia Control panel monitor refresh rates are NOT forced. Manually forcing refresh rates gave me a 12fps drop in 1024*768.

Do you mean the overall global monitor refresh rate or the override where you can select the refresh rates for the various rezs? I do have manual refreshes set in both places, otherwise it's 60Hz and flicker city for me.

Make sure to set vsync to OFF, not application preferance.

I'm pretty sure this is set to app controlled.

Leave AA, AF to application preferance.

OK.

If you have dual monitor setup DISABLE it, this is a known bug that significantly hampers performance.

Only one monitor

Also, your 3200+ can not come close to touching a 64 system, my 3200+ overclocked to 2310 is a beast in comparison. I think you should be around 75fps with your rig.

This is the rig I built a year ago in prep for HL2 - one moment, SCREW YOU, GABE!!!! - but while it's not a dog, I'm sure it's pokier than the latest and greatest chips. (Heck, I haven't even followed the new stuff cuz I'm not planning on building anything for a year or so.) I figured CPU speeds could be a factor, but I believe BFG's rig is basically the same, though he was more interested in being a snot than being helpful in troubleshooting this situation.

Thanks a lot for your reply - this is quite helpful and when I get home, I'll test these suggestions and report back. Cheers.:beer:

 

[DefRef]

Originally posted by: DefRef

Originally posted by: DefRef

Originally posted by: DefRef
The rig: Athlon 3200+/1GB/128MB 5900NU at stock/WinXP SP1/61.77 drivers

Doom settings: 1024x768/Ultra Quality/everything one but AA plus my shadow's on

The punchline: 26.9 fps

For kicks, I tried the Ultra setting though it's supposed to be too much for any mortal video card but it's actually playable and looks great.

Just an update after plunking a PNY 6800GT clocked 400/1000 into the above rig:

1024 rez/Ultra/No AA = 46.2 fps (72% faster!)

1024 rez/High/4x AA = 45.5 fps (69%)

1024 rez/Ultra/4x AA = 41.7 fps (55%)

1280 rez/High/4x AA = 33.4 fps (24%)

1280 rez/Ultra/4x AA = 31.7 fps (18%)

I ran 1600x1200, no AA, high detail and got 40.8 fps. Huh? Then, I unticked Vsync and suddenly I leapt to 52 fps. Sheesh. OCing to 404/1.1, it improved to 53.1 - no biggie.

Lesson learned: Vsync kills performance DEAD!

Here's the latest:

In the Performance and Quality area of the Nvidia control panel, I have AA/AF set to App-Control, Image to Quality and Vertical sync turned Off as the Global Driver setting. In the Advanced part, I switched on Anisotropic optimization and Trillinear optis were already on. Everything in D3 is switched on except vsync and AA; texture detail at High.

1024 - 61.3 fps
1280 - 57 fps
1600 - 57.3 fps (huh?)

I then switched off the refresh override so that apps control it (you didn't answer if you meant that part) and got:

1024 - 59.8 fps
1280 - 60.9 fps
1600 - 56.5 fps

Weird. Any more suggestions?

 

[BFG10K]

You know, you've always had a tendency to be an arrogrant jerk,

Arrogant? You're the one that had a go at me like you're some kind of benchmarking guru and then I find out you didn't even disable vsync in your own tests. How about you at least learn the basics of how to tweak a system before having a go at someome because you can't understand how they got their results?

but are you really contributing anything when you don't include details like what texture detail level your benches were run with or whether you turned off all the eye candy or what?

High quality, everything on except shadows.

though he was more interested in being a snot than being helpful in troubleshooting this situation

Troubleshooting? Don't make me laugh. You weren't interested in troubleshooting, you were simply annoyed that I produced some scores which you couldn't understand. If you want help then just ask instead of digging at my methods.

 

[CrystalBay]

what driver rev. are you using defref ? the 66 series kicks it up a notch....

 

[DefRef]

Originally posted by: BFG10K

You know, you've always had a tendency to be an arrogrant jerk,

Arrogant? You're the one that had a go at me like you're some kind of benchmarking guru and then I find out you didn't even disable vsync in your own tests. How about you at least learn the basics of how to tweak a system before having a go at someome because you can't understand how they got their results?

Yeah, ARROGANT and you're still being a useless crapweasel. ah heck off because you STILL aren't contributing anything other than your inflated sense of superiority.

but are you really contributing anything when you don't include details like what texture detail level your benches were run with or whether you turned off all the eye candy or what?

High quality, everything on except shadows.

Gee, if EVERYTHING is on, wouldn't that include Vsync? It's an item that can be checked on.

though he was more interested in being a snot than being helpful in troubleshooting this situation

Troubleshooting? Don't make me laugh. You weren't interested in troubleshooting, you were simply annoyed that I produced some scores which you couldn't understand. If you want help then just ask instead of digging at my methods.

You're initial post didn't mention what LOD you benched at and you mistook my ironic (have someone look it up for you) comment as an attack on your methods. I realize irony and sarcasm doesn't tranlate well in print, but if you're an arrogant jerk, it doesn't really matter.

There were two ways a person who claims such superior skills could've handled this discussion:

1. You could've queried whether I had the vsync on or off and mentioned that it can negatively impact scores. (This is what Bumrush99 did.)

2. You could spank your arrogant monkey all over your keyboard and act like you're too smart to be helpful to anyone not as Jeff K. as you.

Most people come to forums to seek help and to help others. Why are you here if you're too good to have problems and too much a jerk to help anyone?:disgust:

CrystalBay: As noted, I'm running the 61.77s - I don't like fooling with beta drivers. Let those who can withstand downtime be the happy guinnea pigs.

What's odd is my 3DMark is 11266 using the same settings as before and according to another thread here, that's about 10% higher than expected for this rig's specs. Odd.

(Of course, BFG probably gets 59,399,202,502 3Dmarks with his card, simply because he's just that l337.)

 

[CrystalBay]

D3 @ 1280x1024 High settings/ No CP forced AA/AF- 57.8 FPS / FW 61.77
D3 1280x1024 High settings/ CP forced 4Xaa/8xaf 31.7 FPS

 

[Cheetah8799]

Just finished running a quick timedemo on my rig:

Mobile XP 2500+ @ 2.4ghz
1gb PC3200
2 x 80gb RAID 0
ATI Radeon Pro at STOCK speeds

1024x768 w/ Medium settings, AA/AF off
first run: 39.3 FPS, loading of textures into ram was very choppy.
second run: 47.3 FPS

 

[BFG10K]

Yeah, ARROGANT and you're still being a useless crapweasel.

At least I disable vsync before pointing fingers around at other people.

Gee, if EVERYTHING is on, wouldn't that include Vsync?

ROFL. No, it wouldn't.

You're initial post didn't mention what LOD

Do you know what LOD is? It's level of detail bias. What that has to do with anything here only you would know because nobody else does.

There were two ways a person who claims such superior skills could've handled this discussion:

Where did I claim my skills are superior? I merely claimed that your skills did not put you in a position to accuse me of doing something wrong.

I realize irony and sarcasm doesn't tranlate well in print, but if you're an arrogant jerk, it doesn't really matter.

The next time you accuse me of running textureless polygons or some such rubbish remember to check your own vsync settings first, okay?

As noted, I'm running the 61.77s - I don't like fooling with beta drivers.

That's your problem, not mine.

 

[DefRef]

Gawd, you're still hung up on the joke about textureless polys?!? Sounds like someone HAS been cheating and is getting all defensive and paranoid - imagining accusations where none exist. Lame.

Try less weed and more Cheetos, Bub.:roll:

 

[SnOop005]

Have anyone noticed or experienced a loss of 9-10 fps after you install full windows updates?This is kinda strange but here's my story. I benched my card (BFG 6800 GT) on a fresh install windows with 61.77 on a XP first verison I got 71 fps 1280-1024 on high quality mode(seems pretty normal). I went ahead and installed SP1A i got the same fps rate, then I updated my windows wit all the additional updates after sp1a. Now I only bench around 59-60 on the same mode. I have no idea what happend. Anyone know if there's certain updates that affects the game? My 3dmark seems normal at around 11500...or is it?

here's my set up:

Abit IC7-G max 3
P4 2.8C@3.4
250 gig sata
BFG 6800GT
1 gig of PC4200 running at 1:1

Thanks

 

[auto]

My Budget System:
Athlon XP 2800+ stock
ECS N2U400-A - AGP 8x aperture size set to 256 mb
BFG 6800 OC (350/700)
1 gb Kingston PC2100
Enermax 350W power supply
NEC LCD1920nx 1280 x 1024 60hertz

HIGH: Vsync OFF
1280x1024 - 44.2 fps 2x AA
1280x1024 - 48.2 fps no AA

used "timedemo demo1 1 "

Would switching to 1 gb PC3200 up my fps ?

Thanks

 

[auto]

^^