Doom 3 Low-end/Obscure roundup

[klah]

http://techreport.com/etc/2004q3/doom3-midlow/index.x?pg=1

Parhelia, FX5200, 9000Pro and 5800Ultra; among others.

 

[lament]

GeForce FX 5900 XT 256MB

is there such a thing? my 5900 XT was 128MB and I thought all of them were.

 

[lament]

*sigh*

AT is having some problems.

 

[Don Vito Corleone]

Jeez - those numbers are a lot better than I get from my 9700 Pro, on a similar CPU/mobo.

 

[MemberSince97]

It's about time the general population gets a comparison....

 

[PG]

Be careful in comparing numbers, that's a different demo used for the benchmark. The game comes with demo1, those guys made their own.

 

[eplebnista]

Thanks for the link!

 

[RussianSensation]

Wow the performance difference between Ti4200 and 4600 is massive (from 25-36% difference)
Also 9500pro beats the 9600 pro card. The 5800Ultra beats 5900xt and 9700Pro at Medium Quality across the board...wow (rollo's 4 year old will be proud).

My lowly 8500 gets 38.6 frames per second at 680x480 - OUCH.

 

[KenSimone]

Good thing they used an Athlon 64 3200, because that's probably what most people with older and lower end cards have (roll eyes). I just wish ID would release the freakin' demo already so I could see how pathetically my ti4200 P4 2.4 does.

 

[CraigRT]

Originally posted by: KenSimone
Good thing they used an Athlon 64 3200, because that's probably what most people with older and lower end cards have (roll eyes). I just wish ID would release the freakin' demo already so I could see how pathetically my ti4200 P4 2.4 does.

dude, it says right in the article that they used the A64 3200+ to make sure there was no CPU bottlenecking.
I would have done the same thing.. maybe even a faster CPU.

 

[clicknext]

Originally posted by: MemberSince97
It's about time the general population gets a comparison....

yep...

 

[AnnoyedGrunt]

If you are concerned about their processor choice, here is where you can see how various processors compare:

http://www.anandtech.com/cpuchipsets/showdoc.aspx?i=2149&p=7

It looks to me like those with slower processors should choose a resolution that gives them around 50 FPS or so if possible to best match the proc. Going significantly higher than that may not make a huge difference in overall FPS.

-D'oh!

 

[KenSimone]

dude, it says right in the article that they used the A64 3200+ to make sure there was no CPU bottlenecking.

All these benchmarks show are what everyone already knows: The pecking order of the older cards. IMO it would've been far more useful to benchmark Doom's performance on an average system (~2ghz). Seriously, what useful information can be gleaned from the numbers he produced?

 

[RussianSensation]

Originally posted by: KenSimone

dude, it says right in the article that they used the A64 3200+ to make sure there was no CPU bottlenecking.

All these benchmarks show are what everyone already knows: The pecking order of the older cards. IMO it would've been far more useful to benchmark Doom's performance on an average system (~2ghz). Seriously, what useful information can be gleaned from the numbers he produced?

Well for one, if your videocard cannot run Doom3 smoothly with A64 3200+ at a certain setting, you already know it wont run smoothly with a P4 2ghz processor right (or a slow equivalent). Either way as long as your cpu is above 2500+ then most of the stress will be produced on a videocard.

Here is Anandtechs CPU Impact in Doom 3 Benchmarks

Considering that with a 2.4ghpz P4 6800Ultra produced ~60 frames per second@800x600 and ~73 frames per second at the same resolution with a 3.2ghz p4 you can see that increasing the cpu speed does produce some increase in performance. However this is insignificant considering that A) the cpu difference matters much more for the fastest videocards and almost wont make any difference for slower videocards like 9800Pros and below; B) When the videocard is fill-rate limited and memory bandwidth limited (as is the case with anything other than the new generation cards), having a faster cpu wont improve the situation.

So you can conclude that in Doom 3 with a slower videocard like 8500, 4200, 9600, 5900, 9800 the performance will be roughly within 10-15% slower with 2000+ performing cpu in comparison to A64 3200+ and above. On the other hand going to a 6800nu from 9800Pro produces 100% performance increase in Doom 3. So you can be sure that even a P4 2.0ghz/AXP 2000+ will be faster with a 6800 than A64 3800+ and 9800xt 256mb.

If you want to see previous benches I posted on how cpu and videocard combinations affect performance, please refer to this thread:

The videocard upgrade is a much much better option for you if you are looking to increase your framerates for gaming as your primary objective.

Frankly, the majority of computer users and users on this forum tend to exaggerate the importance of a cpu for gaming. But few tend to actually analyze the situation closely thinking that you need the absolute fastest components together or it wont fly. Sure a fast cpu and a fast videocard will be the optimal solution. Yet a decent cpu and a faster videocard will most surely always win. The irony is that most peoplel that care about gaming tend to get A64 3400+ and 6800nu instead of A64 2800+ and 6800GT which cost the same (both setups that is).

 

[LTC8K6]

Also 9500pro beats the 9600 pro card

Yeah, 8 pipes are better than 4!