Haha... Just saying, I don't think you have ever really led us astray.
Obviously no source can be 100% right.
Printable View
Hes writing about a "best case scenario" and 60% would be way too low for that.
:rolleyes:
Feature-set, bro...
Your over-all point is irrelevant. It offers MORE, it's not going to be better for us because its faster, but because it does more.
You're too caught up in FPS figures and not about how that chip space is being used. If all you want is better FPS, they could push that front very easily. But then games like ARMA II or OFP2 wouldn't have HDR, Bloom, shadows, etc...
A modest 60% increase has never happened so quickly. (less than a year!)
I went from an Radeon X1900XT ~ 8800gts 640mb ~ HD4870 1gig... over a 4year period and each upgrade was only a 25% increase in performance, if that!
Honestly, For $299 and a 60% increase over the $270 HD4870 I bought last October... is fook'ing incredible !!
I still think hd4870 -> hd5870 is going to be like ~80% faster on average, rather than 60% or 160%. 11 days to go!
X1950XTX to a 8800GTS 640mb was a good +25% increase, +50-60% in some cases.
8800GTS 640mb to a 8800GT was a 10-20% increase, 8800GT to a 8800GTS/9800GTX was about the same 10-20%, 8800GTS/9800GTX to a 4870 was also a ~15-20% increase.
So all those upgrades were easily +25% and some cases 50-60%.
all i know is the new cards are going to eat Modern warfare 2 for breakfast, asking for seconds.
Just 1.6 times faster then 1.0*HD 4870? So much for the 1600 shaders rumour...and I doubt 5870 is going to be able to surpass HD 4870x2 at high res with high levels of AA.
Rv870 probably will still have 1600 shaders but that doesn't mean it will offer double the performance cause no drivers are perfect.
http://www.xtremesystems.org/forums/...2&postcount=35
I posted that a week ago.:p:
Science H Logic! Don't do it! :p:
I hope its more than 60% as I expect GT300 to be more than that. We shall see...
As far as 1600sp not being 100% faster than 800 its down to their efficiency, clock speed and where the cards biggest bottleneck rests. We are talking an improved arcitechture here, not simply HD4870 X 2. 100%+ could be possible in certain cases but I'll be quite impressed if that turns out to be the the case. Still I'd expect better than 60%.
I think the 5850 will be around $299 usd and the 5870 are $399
i hope not and i doubt that will be the case......i think they still will be priced at 250 and 350 respectively to give there 4800 series some room to breath...it would be hard to sell the hd4800 series with the HD5800 bieng priced at 199 and 299 but i still think ATI wants to price there 5800's very competitively at the same time....i think we win either way :up:
oh well thats what prices one of my 'contacts' have for RRP
If the GT300 is 10% faster then a 295 like rumoured, then where would that leave the HD5870? I hope it will be fast enough.
I just know that this is going to be very interesting, since funds are at an all time low, and if AMD comes out with a lower priced card, they could have secured my money for this year! :)
and..?
That^^ took place over a 5 year period. My point was the 5800 series is supposedly a 60% increase, in just one years time.
So, those people who are complaining or demanding a higher gains might want to check themselves... if they want a 100% increase, they may have to wait a little longer and pay a tad more.
ATI's claims of 60% better is not disappointing at all.
I heard $349 for HD5870 and $269 for HD5850 at launch :shrug:
A 60% performance improvement would be very good, and even more than what I originally expected. Usually, no more than 50% increase is given from a generation to the next, usually less than that, and last generation HD4000 was more than an exagerated +100% increase or so.
If it would be a number to the average real world cases, it would be believable too. What I see strange is what w0mbat has mentioned: the article is talking about a "best case scenario", and that doesn't match to the other info that has leaked around (a 100% increase in computing/shading power should correspond to a minimum of 100% best case scenario performance increase in a case where computing/shading power was the bottleneck).
I would say that either they don't know what is a best case scenario, or the number is directly wrong. I don't believe that a chip that is 20-35% bigger than RV770 (300-350mm2 vs 254mm2) in a half the size manufacturing process (40nm vs 55nm) is going to be increased only 60% in the most increased aspect of the chip...
And if it's not a best case scenario, then the number doesn't mean anything, because we wouldn't know what that number matches to...