GTX 280 vs. GTX 260 vs. 8800 GTX review

[-n7-]

Review


System Specs
QX6850 @ 3.6 GHz (9x400)
Asus P5Q Deluxe (Ket's mBIOS 1306)
8 GB Mushkin 996580 @ DDR2-960 5-5-15-15
Raptor 150 GB & Seagate 1 TB (games installed on those)
Corsair HX1000 1000w
Vista Ultimate x64 SP1

BFG 8800 GTX OC clocks: 600/900/1404
eVGA GTX 260 SC clocks: 621/1026/1296
BFG GTX 280 OC clocks: 615/1107/1350

Latest official drivers for all cards were used.
BFG 8800 GTX OC Forceware 175.19 configuration
eVGA GTX 260 SC Forceware 177.41 configuration
BFG GTX 280 OC Forceware 177.41 configuration
3D settings were left at defaults other than quality, which was set to "High Quality", & vsync, which was forced "Off", or when i needed to apply AA/AF to a game that didn't have in-game options for that.


For all benchs, i ran tests at least 3 times or more (excepting: CS:Source - twice; Plasma Pong - once).
Results posted are averaged from the multiple runs.



UT3 botmatch @ 2560x1600 DX10 maxed

Warfare: Floodgate
8800 GTX
min - 33 fps
ave - 42.02 fps
max - 53.33 fps
GTX 260
min - 41 fps
ave - 56.39 fps
max - 72.33 fps
GTX 280
min - 51.67 fps
ave - 66.04 fps
max - 88.67 fps

Deathmatch: Defiance
8800 GTX
min - 37.66 fps
ave - 50.49 fps
max - 68.67 fps
GTX 260
min - 49 fps
ave - 66.1 fps
max - 91 fps
GTX 280
min - 60.67 fps
ave - 82.72 fps
max - 124.67 fps


UT3 botmatch @ 2560x1600 DX10 maxed 4xMSAA 16xAF

Warfare: Floodgate
8800 GTX
min - 15.33 fps
ave - 21.07 fps
max - 27 fps
GTX 260
min - 20.33 fps
ave - 30.07 fps
max - 37.67 fps
GTX 280
min - 28.67 fps
ave - 36.85 fps
max - 44.33 fps

Deathmatch: Defiance
8800 GTX
min - 5.67 fps
ave - 23.72 fps
max - 36.33 fps
GTX 260
min - 26.33 fps
ave - 36.96 fps
max - 51 fps
GTX 280
min - 34 fps
ave - 42.96 fps
max - 56.67 fps


Bioshock Fraps run @ 2560x1600 DX10 maxed
8800 GTX
min - 34.67 fps
ave - 41.87 fps
max - 49 fps
GTX 260
min - 47.33 fps
ave - 56.13 fps
max - 67.67 fps
GTX 280
min - 56.33 fps
ave - 69.63 fps
max - 84 fps


Bioshock Fraps run @ 2560x1600 DX10 maxed 4xMSAA 16xAF
8800 GTX
min - 16.67 fps
ave - 20.66 fps
max - 27.33 fps
GTX 260
min - 23.33 fps
ave - 27.37 fps
max - 33.67 fps
GTX 280
min - 28 fps
ave - 34.42 fps
max - 43.33 fps


GRID Fraps run on Le Mans @ 2560x1600 maxed 4xMSAA
8800 GTX
min - 22.33 fps
ave - 29.16 fps
max - 34.67 fps
GTX 260
min - 30 fps
ave - 39.38 fps
max - 46.67 fps
GTX 280
min - 35.67 fps
ave - 46.49 fps
max - 56.33 fps


UT2004 botmatch @ 2560x1600 8xMSAA 16xAF
Onslaught: Ascendancy
8800 GTX
min - 49 fps
ave - 70.18 fps
max - 86.67 fps
GTX 260
min - 59 fps
ave - 79.33 fps
max - 101.67 fps
GTX 280
min - 63.33 fps
ave - 85.79 fps
max - 111.67 fps


Call of Duty 4 Fraps run @ 2560x1600 maxed 4xMSAA
8800 GTX
min - 19.33 fps
ave - 31.07 fps
max - 51.33 fps
GTX 260
min - 31.33 fps
ave - 46.07 fps
max - 76.67 fps
GTX 280
min - 32.67 fps
ave - 53.12 fps
max - 84.67 fps


HL2: Episode 2 Fraps run @ 2560x1600 maxed 8xMSAA 16xAF
8800 GTX
min - 15 fps
ave - 26.86 fps
max - 33.67 fps
GTX 260
min - 26.33 fps
ave - 51.18 fps
max - 67 fps
GTX 280
min - 27.67 fps
ave - 53.62 fps
max - 69.33 fps


Clive Barker's Jericho Fraps run @ 2560x1600 maxed 4x"Smoothing" (in-game AA)
8800 GTX
min - 12 fps
ave - 16.66 fps
max - 22 fps
GTX 260
min - 20.67 fps
ave - 26.47 fps
max - 33.67 fps
GTX 280
min - 24.33 fps
ave - 31.74 fps
max - 39.67 fps


Mass Effect Fraps run @ 2560x1600 maxed
8800 GTX
min - 23.67 fps
ave - 35.67 fps
max - 42.67 fps
GTX 260
min - 32.67 fps
ave - 46.42 fps
max - 55 fps
GTX 280
min - 39.33 fps
ave - 55.22 fps
max - 63.67 fps


Mass Effect Fraps run @ 2560x1600 maxed 4xMSAA
8800 GTX
min - 11 fps
ave - 15.83 fps
max - 19 fps
GTX 260
min - 17.33 fps
ave - 23.59 fps
max - 27.67 fps
GTX 280
min - 20.33 fps
ave - 28.13 fps
max - 33 fps


Devil May Cry 4 Performance Test @ 2560x1600 DX10 maxed 8xMSAA
8800 GTX
min - 20 fps
ave - 25.73 fps
max - 32 fps
GTX 260
min - 30.67 fps
ave - 37.96 fps
max - 45.67 fps
GTX 280
min - 32 fps
ave - 43.78 fps
max - 53 fps
Note on the GTX 280: Saw weird big fps drops in some runs. I discarded those results, but wanted to mention the oddity.


F.E.A.R. benchmark @ 2560x1600 maxed 4xMSAA 16xAF
8800 GTX
min - 23.33 fps
ave - 42.67 fps
max - 96 fps
GTX 260
min - 34.67 fps
ave - 66 fps
max - 137.67 fps
GTX 280
min - 39.67 fps
ave - 76 fps
max - 165.67 fps


CS:Source Stresstest @ 2560x1600 maxed 8xMSAA 16xAF
8800 GTX
ave - 82.7 fps
GTX 260
ave - 158.28 fps
GTX 280
ave - 170.6 fps


Crysis Fraps run @ 1920x1200 all high 64-bit DX9
8800 GTX
min - 19.33 fps
ave - 24.51 fps
max - 28.67 fps
GTX 280
min - 34.67 fps
ave - 37.73 fps
max - 41.67 fps


Crysis Fraps run @ 1920x1200 all high 64-bit DX9 4xMSAA
8800 GTX
min - 15.67 fps
ave - 17.56 fps
max - 21 fps
GTX 280
min - 19 fps
ave - 21.84 fps
max - 29.67 fps


And most importantly...
Plasma Pong Fraps run @ 2560x1600 maxed 4xMSAA
8800 GTX
min - 60 fps
ave - 67.28 fps
max - 70 fps
GTX 280
min - 58 fps
ave - 64.88 fps
max - 68 fps



Impressions & Pics





I'd like to say i'm very happy with my new GTX 280, but that wouldn't really be the whole story.

My old 8800 GTX was a very good card, & held its own for a long time.
My new GTX 280 is certainly a pretty huge improvement, & the GTX 260's performance sits in between the two quite nicely.

This review was supposed to happen a long time ago, in July.

I ordered a BFG GTX 280 back on July 10th, nearly two months ago.
But it turned out to be one of the defective ones that overheat, so back to BFG it went.
They helpfully shipped me back an overheating defective replacement, so again, i had to fight with them.
Finally after what can really only be described as a horrible nightmare, i received what seems to be a functional third GTX 280.

I say seems, because i have zero faith in the quality control that nVidia & BFG are performing (or aren't) on these cards.
As far as i am concerned, only time will tell if this card is actually okay.

Thus far though, at least it doesn't overheat.

In the interim, i purchased an eVGA GTX 260 to play with, & it has done well for me.

Aside from the issues getting a good one, the GTX 280 performs well & scales consistantly, & provides a much-needed boost over my 8800 GTX for newer games @ my 2560x1600 resolution.
I will add more impression of actually playing games (other than for Fraps scores) if i get some time :p

Noise is not perceptible at idle, however, it does become quite loud at load, louder than the 8800 GTX.

It seems the cooling is certainly not overkill, as usually the fan hits 100% or very close to during games, which concerns me alot for future more intensive games that will no doubt run even hotter.
AMD's dual slot coolers are usually even louder at load, but at least there's always lots of headroom, as you never see 100% fan unless you manually set it there.

Thanx for reading.

[[XC] mysticmerlin]

Very nice review

[RADCOM]

Thanks, good review and a wide range of games. I really didn't utilise my 8800GTX enough and then I got another one just before the prices plummeted. I think they are okay in SLI and I will build a new home for them around my recently replaced Striker Extreme. Meanwhile I am happy with my 4870 x2. Next year I will be on a hardware diet lol

[ahmad]

Thanks for sharing your results!

[Computerville]

Very very very interesting review!

This took some time to make, right? :P

Keep it up!

[Martijn]

I like this review! Could you test Crysis with the same settings at DX10?

[BigJoe]

Very good review n7

[PaRn ssR]

nice review

[The0men]

I have one question, Can you please test the GTX260 on the couple of tests where it was missed at the end, like crysis? That would be great.

Awsome review thank you, Exactly what I have been hoping to see, as I was thinking of getting a GTX 260 or 280 to replace my current 8800GTX, was just wondering whether it was worth splashing out on a 280, but I think a 260GTX would do me well, or a 4870 maybe even.

[kiikkuja]

Why didn't you use the same driver for all the tests? All the new betas or whatever they are have worked for my 8800gt up to 177.92.

The performance difference would propably be there with the same driver used but then you could eliminate the driver bottlenecking your performance.

[MomijiTMO]

Nice review. Of course the oc limit of the GTX260 really throws a spanner into the works O___o

[Janno]

if the 8800gtx in sli can keep up with the gtx280, i'm not rushing to upgrade just yet

great review, nice to see 'em new cards compared to the good ol' g80

[Bo_Fox]

Wow! Nice review! (Holy thread resurrection, but hey.. I just came across this thread so had to comment!)

[Dan_c]

great review!!! awsome read...

[systemviper]

thnks for the effort, interesting, love real world stats