Looks like poor Quad SLI scaling in COD4.
1 GX2 at 2560x1600= 85FPS no AA or AF
2 GX2's at 2560x1600=80FPS 4xAA and 16xAF
If I remember right my 3 GTX's do better than that, seems they stayed over 100 FPS at all times.
Printable View
Looks like poor Quad SLI scaling in COD4.
1 GX2 at 2560x1600= 85FPS no AA or AF
2 GX2's at 2560x1600=80FPS 4xAA and 16xAF
If I remember right my 3 GTX's do better than that, seems they stayed over 100 FPS at all times.
Would two of these (quad SLI) work in 16x/8x board or do they require 16x/16x slots?
then get a 680i.. haha
GeForce 9800 GX2 with ForceWare 174.51
http://www.tweaktown.com/articles/13...ion/index.html
Better now but still very close to HD 3870X2 with such diference in price and less software and hardware features.
You could say the same thing about the 8800GT against the 3870X2. Look at the benchmarks again, the 8800GT ain't that far off, in fact it even beats the 3870X2 in a few games.
It does appear to me that these dual GPU cards (from both ATI and nVidia) are overpriced and power sucking monstrosities that only show good scaling in certain games. :down:
Witch games 8800GT wins??
PT boats - HD 3870X2 wins by far far magin in avg and same minimun.
HL2- HD 3870X2 wins with almost 100% plus performance.
WIC- HD 3870X2 wins again by 10fps avg.
Enemy Territory- Quake Wars- Both equal.
Crysis- at 1920 resolution the HD 3870X2 don´t have bar.
Unreal Tournament 3- HD 3870X2 wins with almost 100% plus performance.
Ofcourse if you are expecting to play a 1280 with this card well, you are not understanding this because they must be compared at 1920*1200 resolution and the 8800GT at that resolution is simply crushed IMO.
If you state that HD 3870X2 is almost equal to 8800GT OC, I can state that HD 3650 is almost equal to 8800GT. Plain and simple....
can someone please test it against a 3870x2 with GDDR4? :(
Not everyone games at ultra high res. I am looking at the overall picture, not just the high res performance. The only resolution the 3870X2 consistently beats the 8800GT is 1920x1200. This means the 3870X2 is really only good for gamers with 24"+ LCDs.
At 1600 x 1200 (which is close in pixel count to 1680 x 1050, a common 22" LCD resolution) the 8800GT performs just as well in most games.
This argument applies to both the 9800GX2 and 3870X2 as well, so don't think I'm just bashing your beloved ATI for the heck of it. ;) I think both cards are poor value unless, of course, you game at 1920 x 1200 or higher.
That review with the 174.51 driver is interesting. I don't understand some of their testing though.
1. Why test Crysis at medium? Who is there right mind would buy a $600 state of the art video card and run it at those settings. I want to see it run at high and ultra high with and without 2X AA. If it gets below 30 fps average, no big deal. At least we will know where it stands.
2. Why on earth are people still testing Half Life 2? Why would anyone buy a new $600 video card to boost yourself from 200 to 230 fps? That test is completely meaningless at this point. Older cards run it well already.
3. How about some more newer games that are demanding on the GPU? Dirt, Witcher, Call of Juarez, etc.
No one ever finishes reading these things :P they just look at the graphs and walk on by.Quote:
Since we’re on the topic of Crysis, I took the time to test it at high detail for everyone. There’s a reason I don’t test Crysis at high. When you get an average of only 50FPS at 1920 x 1200 at medium, the performance is going to be too poor at high. So even if the card is 10FPS faster than the other cards, if the average is that low it’s not going to make a difference in terms of it becoming playable.
With that said, we get a very underwhelming 34.4 fps at 1920 x 1200 when set at high. When the numbers start coming up to a decent level we will move from medium to high, and then hopefully to very high; but I dare say that will be quite a ways away yet
Part of the reason they test the same games over and over is that they already have scores from TONS of cards for those games. I would love to see testing with more modern games, but for a fair comparison a lot of the older cards would need to be re-benched which could be time consuming.
I actually read the last page. If it gets 34 fps at 1920 X 1200, it would be even higher at the lower resolutions. My point still stands. A full test for all 3 resolutions at high settings would be a lot more relevant for gamers interested in this card. Current hardware runs in on medium settings just fine.
whens the NDA lifted and we can actually see them retial?
isn't it the 18th somewhere in the world right now???
Quote:
GeForce 9800 GX2 with ForceWare 174.51
http://aycu22.webshots.com/image/461...2677751_rs.jpg
http://aycu35.webshots.com/image/469...1034678_rs.jpg
http://aycu27.webshots.com/image/463...5018400_rs.jpg
http://aycu07.webshots.com/image/474...5261837_rs.jpg
http://aycu26.webshots.com/image/475...2130635_rs.jpg
http://aycu06.webshots.com/image/464...0311863_rs.jpg
http://www.tweaktown.com/articles/13...ion/index.html
.
looks a litle better now :)
regards
to everyone saying "not everyone plays ultra high res" ... then don't even consider these cards.. i mean, thats kinda why they're made right? for people with HUGE resolutions... i couldnt live w/ anything smaller than 1920x1200... and i've just sold 2 8800gtx's and an 8800gt... so i'm in the market for something beefy, and having a hardtime on 3870x2 or 9800gx2
if the 9800gtx overclocks to 900core on stock vcore.. i may very well consider that if it stays below 400 dollars
I'm with you nosboost 30.
I'm on my 1920x1080p 37" FullHD LCD TV as my primary display and I need to know where these cards stack at the high end.
It's the high end guys who care about those high results. 1600x1200 or under guys have plenty of power with today's solution, except maybe crysis.
Those results posted by Mascaras, if true, makes the 3870x2 the value leader at 1920x1200. I wouldn't pay 200 bux more for 10 FPS improvement. For that money, I can add a single 3870 for Triple CFX action and get potentially more performance.
Someone needs to compare a $ to FPS ratio for a bang per buck list: 150, 200, 300, 400, 500 and 600.
forsale in states but way to much,when you can get two gts512 cheaper
http://www.newegg.com/Product/Produc...d&Order=PRICED