Confused about Oblivion Benches

[inveterate]

Why does the x1900xt do better than the XTX

 

[LittleNemoNES]

What benches are you looking at?

 

[inveterate]

the one at anandtech. http://www.anandtech.com/video/showdoc.aspx?i=2746&p=4
Both the TOWN and the Dungeon,, even at the gate it shows higher average,... I'm just confused

 

[n7]

Originally posted by: inveterate
the one at anandtech. http://www.anandtech.com/video/showdoc.aspx?i=2746&p=4
Both the TOWN and the Dungeon,, even at the gate it shows higher average,... I'm just confused

It's likely a mistake (they should be vice versa).

That, or issues with XTX combined with reasonable fps margin for error resulted in those odd results.



Actually, upon further investigation, it appears it's not just the XT/XTX results that are weird.

The 7800GTX 512 MB SLI is beating the 7900GTX in SLI too :laugh:..
Wait, i see one graph with the single 7800GTX 512 MB ahead of 7900GTX SLI :Q

Methinks there was a lot of fps fluctuation that was likely caused by other factors.

Take those numbers as a general guide, not total accuracy.

 

[josh6079]

That's why you can never really go by benches anymore. There are a few around the web that I think are usually pretty good (Anandtech is generally discluded). I remember when the 7900GT came out Anandtech showed the 7800GT beating it on one of their previous benchmarks on older hardware and drivers.

 

[n7]

Originally posted by: josh6079
That's why you can never really go by benches anymore. There are a few around the web that I think are usually pretty good (Anandtech is generally discluded). I remember when the 7900GT came out Anandtech showed the 7800GT beating it on one of their previous benchmarks on older hardware and drivers.

Do you have any idea how difficult it is to do a massive review like that?

Unless you can run an identical timedemo etc, it's not uncommon for lower end cards to beat higher cards, especially when the difference is minimal in the first place.

I don't think it's really AT's fault in this case so much as it's hard to bench Oblivion, & the game is so insanely intensive.

 

[LittleNemoNES]

Most likely HEAVY CPU limitation. Also I think it's reasonably similar so that has more to do with the randomness of SpeedTree

 

[inveterate]

what is speedtree.. and if fx60 is cpu limited,, wtf am i gonna do..

 

[LittleNemoNES]

Originally posted by: inveterate
what is speedtree.. and if fx60 is cpu limited,, wtf am i gonna do..

Wait for conroe :-D

I'm in the same boat as you. All I need is a crossfire edition x1900 and I pretty much have one of the best consumer level PCs (within reason) and even then it will prob not be enough.

SpeedTree is the program that bethesda used for the vegetation. It's randomly generated.

 

[Bull Dog]

Originally posted by: inveterate
the one at anandtech. http://www.anandtech.com/video/showdoc.aspx?i=2746&p=4
Both the TOWN and the Dungeon,, even at the gate it shows higher average,... I'm just confused

Looking at the Town and Dungen benchmarks are WORTHLESS. All you really need to be concerned about are the Outdoor benchmarks. Those are the only benchmarks that aren't CPU limited.

When a test becomes CPU limited wierd things happen and a mid range card can wind up outperforming a ultra-high end card. But the outperformance is within the margin error so in reality its just CPU LIMITED.

 

[inveterate]

Yo Bull Dog,, Did you change ur x1900xt's voltage at all to get to 725,

 

[inveterate]

i have a vf900, so i want to try oc

 

[raystorm]

Nice to see my 6800GS at the bottom of every benchmark.

 

[TheRyuu]

I thought that alot of stuff in Oblvion was placed randomly (like grass and stuff) so the benchmarks could be a tad off?

 

[emilyek]

A 3700 is enough CPU for Oblivion. My opty 146 at 2.5 does fine, although I usually run it at 2.8.

Oblivion is all about GPU and video RAM.

The benches all over look slightly low. I read all the benches before I bought the game and was worried I wouldn't be able to play it.

I have a 6800 vanilla with all distant stuff on, Large textures, no Bloom, no HDR, other settings just on the low side of the 'medium' range and the game looks pretty good and is very playable-- most of the time, anyway-- but large outdoor areas and lots of creatures on screen dip me into the teens-twenties. At 1024 x 768 outdoor areas are in the 30-40's and regular dungeon crawling is in the pleasant 50-70 fps range.

 

[Howard]

Since we measured performance using FRAPS and not through a scripted timedemo sequence, the amount of variance between runs is higher than normal; differences in performance of 5% or less aren't significant and shouldn't be treated as such.

First page.

 

[LittleNemoNES]

Originally posted by: wizboy11
I thought that alot of stuff in Oblvion was placed randomly (like grass and stuff) so the benchmarks could be a tad off?

 

[Powermoloch]

oblivion ran fine w/ my sempr0n 3100+ (@ 2.069ghz) / mushkin xp4000 2 x 1gb / x850xtagp stock



I play on 1280x1024 w/ med-high settings.