Doom 3 Benchmark Inconsistencies

[Mr.Tweak]

I'm confused (it doesn't take too much, heh)

I've been hearing stories that Doom 3 is capped at 60 FPS and this is evident throughout HardOCP's benchmarks with the highest-end systems. However, in the benchmarks that ID provide we see FPS over 70 in some cases.

Am I behind the times or is there something ID isn't telling us? I think we should be told how to disable the FPS cap!

[Yawgm0th]

Re: Doom 3 Benchmark Inconsistencies

http://www2.hardocp.com/article.html?art=NjQ0

Have you seen that? It's not just the benchmark, it's the entire guide. I'm not sure, but I think Nvidia, HardOCP, and maybe even Id may be up to something.

http://www2.hardocp.com/image.html?image=MTA5MDc4NzE0M1RPNjJBTU9FV1hfOV81X2wuZ2lm

http://www2.hardocp.com/image.html?image=MTA5MDc4NzE0M1RPNjJBTU9FV1hfMTZfNl9sLmdpZg==

http://www2.hardocp.com/image.html?image=MTA5MDc4NzE0M1RPNjJBTU9FV1hfMTdfNV9sLmdpZg==

Um, GPU limitation with an X800 XT-PE? I don't ****ing think so.

Also, there's no question about the cap; it's there.

http://www2.hardocp.com/image.html?image=MTA5MDc4NzE0M1RPNjJBTU9FV1hfMTdfNV9sLmdpZg==

http://www2.hardocp.com/image.html?image=MTA5MDc4NzE0M1RPNjJBTU9FV1hfMTdfMV9sLmdpZg==

I mean, the guide does show how nicely the game performs on lower systems, but this is utter BS.

[Mr.Tweak]

Re: Doom 3 Benchmark Inconsistencies

I'm deeply investigating and will get to the bottom of things shortly

[Mr.Tweak]

Re: Doom 3 Benchmark Inconsistencies

It's well in the works... :)

[Yawgm0th]

Re: Doom 3 Benchmark Inconsistencies

Nice...

Just a thought: It could have been the brand of cards they used. The X800s were VisionTek cards. I don't know from experience or tests whether or not VisionTek makes good cards, but www.newegg.com, possibly the largest U.S. hardware retailer, doesn't even carry VisionTek cards. For all I know, the VisionTek's use inferior hardware and even come with lower stock speeds.

[tuxfreak]

Re: Doom 3 Benchmark Inconsistencies

VisionTek used to make excent cards, really great stuff. However, the last one I bought was when they were still making Nvidia cards. VisionTek got pissed and droped nvidia when the FX cards came out and started building ATI cards. I havn't bought a VisionTek card since my fabulous Ti4200, so it's possible their quality has deteriated since then.

[amd_man2005]

Re: Doom 3 Benchmark Inconsistencies

AFAIK Visiontek still makes a quality card.

[Yawgm0th]

Re: Doom 3 Benchmark Inconsistencies

Then I guess either HardOCP or id (or both) are full of it, because those benchmarks are BS. Well, we'll have to wait for some tests from other sites (*cough* :) ).

[wayout44]

Re: Doom 3 Benchmark Inconsistencies

Try these then,
AnandTech Doom 3 Graphics Deathmatch
AnandTech Doom 3: CPU Battlegrounds

[Yawgm0th]

Re: Doom 3 Benchmark Inconsistencies

Awesome. The X800s don't do nearly as bad as before, but it seems as if Doom 3 were really designed to run on the 6800 and vice versa. There's no way the collaboration between id and nVidia (it's a TWIMTBP game) didn't involve optimization of both the game's code and the card's design. This isn't a good thing for the gaming industry.

On a side note, the CPU Battlegrounds article really shows just how bad Intel processors are for gaming.

[Mr.Tweak]

Re: Doom 3 Benchmark Inconsistencies

Quick FSB Overclocking Performance Investigation

It's almost ready!

[Yawgm0th]

Re: Doom 3 Benchmark Inconsistencies

Awesome, but there's no front side bus on an Athlon 64... That's why they're so much faster than other processors. Some BIOSes will still use FSB as the overclocking term, but it's not really FSB. Just being picky.

[Yawgm0th]

Re: Doom 3 Benchmark Inconsistencies

After reading that article (a week a ago was it? two minutes after you put it up there...) on Doom 3 benchmarking, I realized that some of the inconsistencies were explanable by the fact that time demos aren't as CPU-intensive as the actual game, so GPU limitation was seen in areas where it shouldn't have (even for such a GPU-intensive game, some stuff was rediculous).

That said, there are still some parts that make absolutely no sense. The Athlon XP 2500 system outperforms the Athlon 64 3000 system. The FX5700 Ultra on the 3000 has best playable settings of 800x600@Medium quality. The XP 2500 has best playable settings of 800x600@High Quality OR 1024x768@Medium quality. Now how does that make any sense? The Athlon 64 3000 beats the Athlon XP 3200 in every single benchmark I've ever seen (including Doom 3 at Anandtech), and this is a 2500, not a 3200. On benchmarks, or through specs the XP 2500 doesn't have any advantages over the Athlon 64 3000, so what the hell (no pun intended)?

There's also the fact that my Athlon XP 2200 system (haven't tried on my 2GHz P4 with 9800 Pro yet) with a Radeon 9600 and 512MB of PC2100 RAM outperforms their 2500 with 9600XT (and their Athlon 64 with FX5700). Even with all the tweaking I've done, that's still a bit odd...

[deza992450]

Re: Doom 3 Benchmark Inconsistencies

Test Ran with the Comp in my Sig" Below.
I Ran each test once only. Under normal load.
53 Processes Running, including Edonkey, Email,
Yahoo, MSN, Norton AntiVirus etc.....
I used the Timedemo demo1.demo

To run a Test, Run Doom 3, then bring down the console by pressing CTRL-ALT-~ (Tilde button)
then type in timedemo demo1.demo wait for test to complete and run again if you wish.


These things did not change.
Anisotropic Filtering Method........Quality
Texture Preference...................High Quality
Mipmap Detail Level..................High Quality
Wait for Vertical Sync................Application PreferenceApplication Preference

1600x1200
High Quality

6x AA 16x AF: 286.0 seconds @ 7.5 FPS

4x AA 8x AF: 194.3 seconds @ 11.1 FPS

2x AA 4x AF: 134.5 seconds @ 16.0 FPS

2x AA 2x AF: 135.4 seconds @ 15.9 FPS

AA (App Pref) AF (App Pref): 102.5 seconds @ 20.9


1280x1024
High Quality

6x AA 16x AF: 194.6 seconds @ 11 FPS

4x AA 8x AF: 135.4 seconds @ 15.9 FPS

2x AA 4x AF: 95.9 seconds @ 22.4 FPS

2x AA 2x AF: 95.9 seconds @ 24.2 FPS

AA (App Pref) AF (App Pref): 89.1 seconds @ 24.1


1024x768
High Quality

6x AA 16x AF: 133.8 seconds @ 16 FPS

4x AA 8x AF: 96.8 seconds @ 22.2 FPS

2x AA 4x AF: 68.0 seconds @ 31.6 FPS

2x AA 2x AF: 62.5 seconds @ 34.4 FPS

AA (App Pref) AF (App Pref): 63.1 seconds @ 34.0


The numbers seem low, but the game runs very very smooth on 1280x1024 and is very playable on 1600x1200.

[metallicat666]

Re: Doom 3 Benchmark Inconsistencies

53 processes is really nasty. you would get much better scores if you got rid of some of them. try running the windows xp tweak guides