Can we expect a 4850X2 toe-on-toe with 2 4870? :)
Printable View
Can we expect a 4850X2 toe-on-toe with 2 4870? :)
Is it just me or does the 2x1GB 4870X2 compared to 512MB on 4870CF seem to make no or only a very small difference, even at very high resolutions + AA.
I'll here mention that 9800PRO 256MB cards didn't use 256MB for about 2/3 of thier lifespan...driver must know to use it, and based on most reviewers thinking cards are 2x512MB, the driver thinks the same as well.
Nope, not a big fan of coverage sampling anti aliasing, it seems kinda 'fake' in my eyes, 16xCSAA does look better than 4xMSAA, but 8xMSAA is truly amazing, especially if you're limited to 1680*1050 rez on 22" LCD or 1280*1024 rez on 19" LCD. But i don't like wide blur and narrow blur method from ATi either, though edge detect + box AA is truly fantastic in my eyes.Quote:
Do you guys agree about 16x CSAA > 8x MSAA?
:ROTF: LOL, denial is quite futile my friend, just accept it and come out of your green closet, we understand. :) You're firmly in the green corner and loves their cards, fine, we got it, everyone is entitled to their own fetish, i personally respect your choice. :clap:
But please, don't thread crap everything related to ATi in this subforum, it's getting tiresome for MANY of us (just count how many peeps that has called you out just in this thread). :cool: Enjoy your highend, enthusiast card, it's indeed one heck of a fast -and expensive, bad price vs performance at launch- card. :up:
edit: Mr. Mod, please join both of my posts altogether, my bad. :o
I wasn't being negative. I was stating what I said before. That it would perform less then 4870 in CF. and In the same post, I said it looks like they improved on their methods of a dual-GPU solution because it does better then the 3870 CF vs 3870x2 method.
So I still don't you see your point. Nobody looks at the good points in my post, just the bad ones....
Funny thing is, they're using 4870X2 to battle GTX 280. :P
Gotta love this too:
http://images.anandtech.com/graphs/r...3626/17191.png
But it doesn't perform less unless you selectively exclude the times where it performs equal or better. If you factor in the range of different games and drivers used along with the difference in memory size, sometimes it performs better and sometimes worse - but on average it seems to perform, *gasp*, about the same as CF 4870s. And that meshes with reality since it seems to simply be crossfire on a card with a high bandwidth interconnect.
*GASP* You need to look better. Even safan80 posted a graph thats a benchmark. the 4870x2 performed worse then it's 4870 CF Counterpart.
Here is the image.
http://www.xtremesystems.org/forums/...1&d=1216057406
That is one game at one site. It just happens to be the game we both play most of the time so we focus on the performance there.
Look at a few others:
http://www.driverheaven.net/reviewim...hs/cod2560.jpg
http://www.driverheaven.net/reviewim...hs/wic2560.jpg
http://www.driverheaven.net/reviewim...hs/hl22560.jpg
http://images.anandtech.com/graphs/r...3626/17190.png
And those were the ones I could cherrypick in 2 minutes on a crappy tablet pc, you will have to wait a few more hours if you want me to write a little novel about it.
Face it, you are wrong in this. It is roughly equal to 4870 CF, sometimes less, sometimes more, but usually about the same.
And?Quote:
Originally Posted by Anxioz
Perkam
I'm not wrong, I'm not the one changing my opinion.And You only had one other picture from one other source. It looks like your just cherry picking the results to wrap around your point.
Only one? and Don't say there is more, I mainly checked all the results, Typically the 4870x2 is 3-10 FPS lower then a 4870 in CF.
I'm with Warboy in this..Some of his points about the min and the avg fps are correct and quite rational..
The most remarkable achievement by DAAMIT is single 4870 undoubtedly.. Helluva card for the bucks and single card..The card that made Nvidia drop prices drastically.. I'm still amazed as to how it can catch up to GTX280 levels..
Now about the 4870x2 part or Nvidia's x2 solutions..
No matter what most ppl say, some of us actually hate the SLI/Xfire solutions and the sandwich cards..
I paid 500+ euros for the GTX and I don't regret my choice because I knew more or less what I was getting and what would be the benefits/costs of this choice. If Nvidia releases the 55nm counterpart withing a certain timeframe I'll step up to it IF it will be a single chip. If not I'm gonna stay with this one..
And I can't bare hearing about depreciation costs(and babycries) in the bleeding edge. High end and price/perf are irrelevant terms as all of us know..
I wanted the best possible single card out there (not a single sandwich) and the fact remains that Nvidia holds this title and will probably hold it for some more time till the new generation where ATI seems to have an ace up their sleeve.
Well no duh. I just did what you are doing in reverse.
Quote:
Originally Posted by Solus Corvus
I read all the reviews here so far, and I concluded that they are about equal. As drivers mature for the X2 I'm sure that will only solidify my conclusion. With equal core and memory speeds and more memory why on earth would you expect the X2 to be slower, besides driver issues?Quote:
Only one? and Don't say there is more, I mainly checked all the results, Typically the 4870x2 is 3-10 FPS lower then a 4870 in CF.
It bothers me that only a few sites feature min fps, but going by the previews, CF and X2 are neck and neck MOST of the time, as it should be. A 1-3 fps delta is insignificant. In a couple of games, the delta grows...noticeably Assassins Creed has a ~15 fps delta (Muropaketti preview) in favor of CF in the max fps (although the avg and min are actually better for the X2, at which point I'd probably give the X2 the advantage for offering a more consistent performance). Bioshock, HL2, and ET:QW are the other noticeable games (and by noticeable, I mean noticeable on a bar chart, not in game).
Then again, there is the fact that these are engineering samples with potentially crappy drivers and that the X2 will cost ~$100 less than CF.
You're quite right with your observations in general m8.. Then again I'd expect it to be significantly more "effective" in these areas not at best equal to the XFIRE.. Could be the drivers or the 512mb versions (highly doubt that but we'll see) or something else..
Anyway as I said they are a viable solution for me so I don't care that much..
I'm going to play devils advocate here and almost back track from my previous post, perhaps even kick up some dust and annoy that cat in Cooper's Avatar
BUT
As much as this might sound bad...but we should take these benchmarks with a Pinch of Salt.
1) The Card is not due out until August the 13th...and COULD be delayed. This gives the Drivers and the hardware itself more time to mature before it hits retail.
2) We have no idea the level of functionality the card is at, for example the fancy Crossfire Sideport might not yet be enabled at a Driver or Hardware/BIOS level and the card could still be in fall back/compatibility Crossfire Mode
3) Not enough games were tested.
4) Official HD4000 series are still not out yet so lets see if they bring anything new to the mix.
As controversial as this sounds, on the whole I am not impressed with the minimum or average FPS gains over the Crossfire HD4870 or GTX280, however the MAXIMUM FPS shows that there is some potential to this puppy.
Hopefully we will see more previews over the next few weeks with various hardware and driver revisions.
John