Radeon HD 7000:  HD 7790 |  HD 7850 |  HD 7870 |  HD 7870 XT |  HD 7950 |  HD 7970 |  HD 7990 |  Radeon HD 8000:  HD 8750 |  HD 8770 |  HD 8850 |  HD 8870 |  HD 8950 |  HD 8970 |  HD 8990 
GeForce 600:  GTX 650 Ti Boost  |  GTX 660  |  GTX 660 Ti  |  GTX 670  |  GTX 680 |  GTX 690 |  Titan: GTX TITAN |  GeForce 700:  GTX 760  |  GTX 760 Ti  |  GTX 770  |  GTX 780 |  GTX 790 

October 11th, 2012

Sea Islands To Feature Only 15% Performance Increase?

Sea Islands To Feature Only 15% Performance Increase? radeon 8000

Sea Islands Performance Hit By Tropical Storm or Hurricane.

AMD’s upcoming GPU line, Sea Islands (successor to Southern Islands), have been speculated on across many hardware sites since the end of the 7900 series. First, what we are almost certain of is this series of GPUs will not be released in 2012, rather it is being reported as March 2013 by SemiAccurate.com.

Now comes the interesting new information, AMD promised AIBs an increase of around 30% vs. their current Southern Islands generation. However SemiAccurate is reporting this amount of performance increase to be a rather radical claim as they are citing “more measured sources” as showing only a 15% increase in performance. Users with the top of the line AMD product, the 7970 Gigahertz Edition, have seen the large power draw Tahiti XT uses when OC’d. This can also be seen by the lack of any 7990 by AMD themselves; instead AIBs have created their own 7990, all of which require 3x 8 Pin power connections. This makes AMD’s next iteration most likely power bound (as S.A. says) and obviously, just like any silicon based hardware manufacturer, die size bound by cost.

These two pieces of information solidify the final the rumor, which is that the number of Stream Processors will only be increased slightly. As I’m sure many of you know, increase of SPs not only increase TDP but also increases the mm^2 of the die. As we get closer to the launch of the product I’m sure we will get more details on how solid these rumors are or are not; they do call themselves Semi Accurate for a reason.

Written by Keith H.

Author: WhyCry  

  • skr13

    I remember such kind of rumors before Nvidia Fermi appear.
    Only 15% increase is ridiculous, i´m expecting many type of this “news” in the next months, until some HD 8000 benchmarks appear.

  • Jerome

    You guys/gals remember the “Epic Samaritan Demo” GTX680>Tri-SLI 580′s…..ahahha yeah right!

  • BestJinjo

    Damn, that’s disappointing if true. 15% sounds like a refresh generation similar to how HD6970 was to 5870. At least at that time AMD doubled the VRAM and increased tessellation performance that were weak points of 5870 series. This time HD7970 doesn’t have weak tessellation like 5870 did and 3GB of VRAM is more than enough. 15% after more than a year wait is just sad.

    I hope NV launches a much faster GTX780 than GTX680 or we might be looking at another yawn generation ala GTX500 vs. 6900 series which was really GTX485 vs. HD5890 2GB ;)

  • BestJinjo

    Ya, that was marketing BS alright. GTX680 ended up just 30-35% faster than a 580.

  • DavidJamesIsTung

    15% performance is a huge disappointment if this is true.

  • Jerome

    Roughly 15%in gaming performance Hmmm… What about GNC 2.0?? I wonder how much increase in compute performance Sea Islands will have over Tahiti.

  • BestJinjo

    Remember this rumor for HD8970 with just 2304 SPs?
    http://www.tweaktown.com/news/22102/rumortt_amd_radeon_hd_8970_coming_soon/index.html

    So many conflicting rumors, too hard to know what’s what. The mod here said his sources 100% confirmed that HD8970 will have 2560 SPs. Now we are getting estimates that it may only be 15% faster or that would line up closer with 2304 @ 1100mhz.

    Kit-Guru reported that “partners are planning to show off the Radeon HD 8970 XT at CeBIT in March – with a full launch at Computex in June.”
    http://www.kitguru.net/components/graphic-cards/jules/amd-radeon-hd-8970-and-8950-launch-plans-revealed/

    If AMD is really not launching this thing until March-June, they are probably having issues with the design. It shouldn’t take 15-18 months from December 2012 to only squeeze 15% more performance. Not sure what the deal is. They should have just made a large Pitcairn XT gaming chip and left compute for 22nm when die space allowed it. NV ditched it for Kepler which was a smart move for a gaming GPU.

  • Raghar

    Looks like Charlie would be right for once.

    NVidia forced AMD’s hand by forcing them into GHz edition. When you calculate 0.15 increase to GHz edition and compare it to normal edition, it’s significantly better. Of course overclockers would rise the clock to in effect GHz edition anyway. What was original frequency of 7770?

  • BestJinjo

    Of course but HD7970 “original” can overclock to 1150-1200mhz. For the average person 15% over Ghz edition might sound nice but I doubt an average gamer buys a $550 GPU. 15% over 7970 GE would be sad unless they price it at $369 or something.

  • http://www.facebook.com/FwixGamer Anas Azghari

    emmmmmm by experience , I’m not sure about that , wait and see!!

  • thodge84

    Well, the GTX 680 was able to run the demo that took 3 580′s previously. In raw power it’s not greater, but the whole point was that they were able to get the thing running on a single 680 thanks to optimizations, utilizing FXAA, and thanks to a faster and more efficient architecture. Sometimes it’s not about sheer brute strength, but more efficiency.

  • thodge84

    15% is fine with me. There’s nothing that really brings a 680 or 7970 GHZ edition to it’s knees unless you’re gaming at mega HD resolutions. There’s still a ton of untapped potential in the current rounds of cards. It just sucks to see all the potential go to waste thanks to the vast majority of games being ports from consoles on outdated engines with games designed with 2004 level tech & less than 512mb of ram in mind.

  • Pingback: AMD Radeon: 8000 Series Thread (News and Updates ) - Page 2

  • mic08n64

    In light of this news, I’m curious of your thoughts on this. If you remember the leaked slides of sea islands the 88870 and 8850 were equivalent to the GTX 680 and in speed with the gk104 being slightly bigger in chip size chip size.What’s do you the chances of Nvidia just refreshing and optimizes GK104 instead creating a new midrange chip with additional SMX to go against the 6670 and 8850. ie 1536 SMX vs 1792 shaders 1344 SMX s 1536 shaders.

  • maftul

    Throw in some Anti-Aliasing and both of those cards will be brought to their knees even in resolutions lower than 1080p. Of course without AA those cards can play everything in “mega HD” resolution but one might as well play on a console if there is not gonna be any AA.

  • maftul

    This isn’t surprising at all, hardware manufacturers are deliberately delaying development of new technologies to be able to take more out of people’s pockets, everyone is doing this these days. How much faster is Ivy Bridge in comparison to Sandy Bridge at the same clock? 10-15%. How much was Sandy bridge faster than the first generation i5-i7s at the same clock? 10-15%. As long as there are people willing to change their CPU/VGA twice a year for the smallest increase in performance, this is how things are gonna be.

    Let me give a small example of how slowly technology is going forward based on personal experience:

    Radeon HD4870 (Sapphire non-reference) vs Radeon HD7950 (Gigabyte non-reference):

    Price: 220$ vs 330$
    Technology 55nm vs 28nm
    Transistor Count: 965M vs 4313M
    Bus Width: 256-bit vs 384-bit
    Core Clock: 750MHz vs 900MHz
    Memory Clock: 900 vs 1250MHz

    Based on the above info, anyone would expect the HD7950 be at least 10-15 times faster than HD4870 but is it? Of course not, it’s just 2-3 times faster and it is more expensive while not even being the top-of-the-line single-gpu VGA while HD4870 was top-of-the-line single-gpu VGA at its time. How much difference does it make in games? I maximize shadow in games and force 4x-8x MSAA and not even SSAA and get the same FPS as before.

    P.S. Yes, there are other factors involved too but they’re not enough to reduce 10-15 times difference to 2-3 times.

  • Pingback: Sea Islands ofrecería la mitad del rendimiento esperado | El Chapuzas Informático

  • skr13

    Yes, If AMD increase TDP over 250W will be a FLOP flagship card, in the other hand Nvidia can increase VRAM to 3GB, Bus to 384, and TDP until 250W.
    Let´s see whats will happen…

  • http://twitter.com/jeet_shek Abhijeet Mahapatro

    Just got a brand new MSI HD-6990 at $220(due to stock clearance) and it is paired with another HD-6990. I think I can skip this generation too like I did with HD-7000 and NV 600.

  • BestJinjo

    Right that’s the point. They used FXAA and different settings to run the demo which made it not an apples-to-apples comparison. Later NV announced that FXAA will be supported on 500 series as well. At that time the statement was true but if you run GTX580 Tri-SLI with FXAA, I am sure it will cream GTX680 in frame rates in that demo. The other problem is 580s have 1.5GB of VRAM. It could be that the textures in the demo made them perform a lot worse than expected exponentially due to lack of VRAM.

  • BestJinjo

    That’s why everything we see with specs and prices are all rumors. HD8850/8870 being as fast as GTX670/680 could be false or true but based on this rumor sounds like it’s false. I don’t know what NV will do – increase clocks on GK104, drop die harvested 12-13 SMX GK110 parts, create an entirely new chip GK1xx we haven’t heard anything about. What NV does have is an advantage to increase clocks or enlarge the die size as GTX670 has better performance/watt than 7970 and the same for 680 over 7970 Ghz Editions.

    NV also has another advantage since their GK104 chip is memory bandwidth starved. Even without adding a lot more functional units, just going with a wider memory bus and 2048 SPs would make it faster by a lot more than 15% over 680. I don’t think when NV was designing GTX700 they are thinking HD8970 will be only 15% faster. They probably look to estimate where they think the competitor will land and then assess where they can land to maintain their pricing strategy and profitability. I think NV will win next round if HD8970 is just 15% faster.

  • BestJinjo

    The handful of games that cripple 7970/680 are the same ones that have been crippling GPUs for ages – Witcher 2 with Ubersampling (this is basically SSAA), Metro 2033 (cripples even GTX690 SLI with max settings), and BF3 (really to maintain 60 fps minimum in multiplayer you need GTX680 SLI OCed and 6-core @ 5.0ghz). There is also Sleeping Dogs but it uses advanced HDAO and a form of super sampling AA. Other that, most games can be maxed out rather easily on 7970/680 at 1080P. Still, next year I am sure we’ll get more demanding games like Crysis 3 and Metro Last Light, etc. and cards faster than 15% over today’s would be welcome.

  • BestJinjo

    How did you arrive that HD7950 should be 10-15x faster than HD4870?

    1) The transistor size alone tells us 7950 should be at most 4.5x faster, but it will be less since we know the transistors in 7950 are used for compute functions that are not as advanced in 4870. The actual transistor space allocated for SPs, TMUs and ROPs are less than 4.5x more in 7950.

    2) HD7950 is also not a fully unlocked flagship 4.3B chip. HD4870 was a fully unlocked 965M flagship at that time. You should be comparing 7970 to 4870 and 7950 to 4850. 7950 is also very conservatively clocked. It can often be overclocked by 40-50%, which makes its stock performance rather underwhelming since AMD rushed it to market. HD4870 wasn’t rushed.

    3) 4870 came out June 25, 2008. 7950/7970 came out January 2012. That’s only 4.5 years ago. Even based on Moore’s law there is no way 7950/7970 should be 10-15x faster than 4870. The number of transistors roughly doubles every ~2 years and it looks like it’s going close to that path. Since performance/transistor shouldn’t improve dramatically, why would you have expected 7950/7970 to be 10-15x faster?

    I agree with your post that as long as people are spending $$$ for 15-20% more performance, Intel and AMD/NV don’t need to push for 50-100% faster as they have in the past because node shrinks are more expensive than ever and most games are console ports. Just not sure about your 4870 vs. 7950 comparison. It seems your expectations of 10-15x faster every 4.5 years sounds way too optimistic.

  • Bobby Waldon

    Wonderful, if this happens Nvidia will be able to hold back and overcharge customers for a midrange chip labeled as a high end one again….gk114 instead of gk110 anyone?

  • Jerome

    LOL exceed 250w=flop very bold statement there skr13! It all depends on how it handles the heat and what we get in return for the power consumption. If its 15% increase with a huge power draw + 250w than yeah, we are not sure what optimizations Sea Islands will have. 480, 6970, 580 have all been very high power consuming cards and all of them not flops buddy!

  • BestJinjo

    I don’t think it’s that sample. NV doesn’t design 2 separate GPUs and then holds 1 back on purpose. GK100 never launched because it was unmanufacturable at 28nm wafer shortages that NV faced. GK104 was used not because NV could, but because they had to. Even by Q4 2012, NV barely shipped any GK110 chips for science purposes, we are talking 1000 GK110 chips shipped last month to Oak Ridge that ordered them in March 2012, but their facility needs 14,000 of these chips and they won’t be delivered until March 2013! Think about, they just sent out 1000 chips for initial testing phase. No way could NV have launched GK100 in millions of units for gamers if they are having such a hard time delivering GK110 to its corporate clients. If AMD runs into a TDP wall at 400-420mm2, so will NV since they use the same 28nm node. The difference is NV’s chip doesn’t waste transistor space on compute leaving them more room to add graphical units before they reach 420mm2 die. This gives NV an edge since they can add more performance into a 400mm2 die than AMD can.

    NV doesn’t sit there and hold back its fastest chip thinking HD8970 will be only 15% faster. If NV can deliver 30% faster GPU at reasonable die size and margins, they will, but people who think NV isn’t bounded by TDP as well are dreaming. GTX680 already uses 185-190W of power at peak and NV can’t just drop a 550-600mm2 die chip with 2880 SP as many fanboys believe and not go over 250W TDP itself. I think most of us expected next generation to be 30-40% faster at most. Looks like NV may still deliver on that rumor but AMD might need to lower prices.

  • emyyhh

    You are the voice of logic on this site BestJinjo

  • BestJinjo

    I think I agree with Jerome on this one. We have to look at the performance increase vs. the power consumption. If HD8970 is 30% faster and uses 250W, that’s fine. If it’s 15% faster and uses 250W, that’s underwhelming. GTX480 peaks at 270W in games and uses on average 250W and it still sold despite an obnoxiously loud reference cooler. The other thing is I think people buying $500 GPUs care less about power consumption as the mainstream and lower end GPU buyers. Someone who is dropping $500+ on a new GPU is unlikely to have a cramped case, a 450W PSU and a stock Core i5 CPU and care about minor 30-40W differences on a rig that already uses 300-400W:

    http://static.techspot.com/articles-info/572/bench/Power.png

  • skr13

    AMD GCN 2.0 can´t change much compared with previous architecture, so they need to focus on Power Consumption, like previous rumors said.
    http://news.softpedia.com/news/AMD-Radeon-HD-8000-Oland-Is-33-More-Efficient-than-Radeon-HD-7000-Tahiti-Part-1-293589.shtml
    If HD 8870 was supposed to decrease TDP to 160W and perform better than 7950, i believe that AMD will care also about 8900 segment Power Consumption.
    So, if HD 8970 exceed 250W TDP of course will be a fail, if Nvidia counterpart will have lower TDP for same of better performance.
    HD 5870(188W), HD 6970(250W), HD7970(250W), if the next flagship overpass that 250W mark, sure many people will think twice before buy it.

  • BestJinjo

    Performance/Watt/Performance is very important for most of consumers.

    If this was true, then why did NV have more market share during HD4800-6900 generations? In all those generations, AMD had superior price/performance and performance/watt. Clearly performance/watt is not the most important factor in purchasing decisions for consumers then. Also, it doesn’t explain why GTX460/470/480/570/580 sold well as all of those had terrible performance/watt, especially all the high-end cards. GTX480 and 580 even used more power than HD7970 GE and sold well.

  • skr13

    Performance/Watt/TYPO…
    Because there are fan boys, that purchase Nvidia in every generations.
    But budget-conscious consumers made their decisions based on reviews and not Brand/Company.
    GTX 400 Fermi was better than HD 5000 in DX 11 games that´s why they sold more.
    I think after ATI change to AMD, made some influence in the market for less informed people.
    GTX 480 or 580 used more power, but why this comparison? 7970 GE is a 2012 card, and those from 2010, besides different manufacturing process.
    Drivers problem are one decisive factor behind lower AMD/ATI less market share.

  • BestJinjo

    Thanks! I try to look at different sides. I think from a business sense, NV would make more $ selling GK110 for thousands of dollars and making a gaming chip for the consumer market that’s more efficient and smaller. Worst case if this generation is does not have substantial performance increase, we save our money for 2014.

  • thodge84

    meh. don’t need ubersampling. regular AA is fine with me. metro 2033 is just old now. obviously if they went back and made the game with the hardware in mind they would be able to get more out of it, but the game seems like crysis to me in that much could be done to optimize it. Crysis 3 is going to be the game i’m looking forward to the most technically. Crytek is willingly pushing the envelope and i’d love to use that game as a measuring stick of what the cards we have on hand are truly capable of.

  • skr13

    http://www.obr-hardware.com/2012/10/market-specialist-comments-amd.html
    “So if AMD will continue to have the same low gross margin Q4/12, goes bankrupt or is bought by someone. AMD is rushing into the abyss, it is actually there and not get out.”

  • BestJinjo

    But what driver problems is what I don’t understand? I’ve used AMD and NV cards for 10+ years now. I find the driver issues complaints come from NV users who either never used an AMD card since 2005 or they had a problem in 1-2 games and never waited for a new driver update. Look at Shogun 2 or driver problems in BF3 with textures for GTX650Ti? Why are those forgiven but when AMD improved driver performance by noticeable amount since original HD7000 series launch “they have shitty drivers”. It’s expected that GCN like original Fermi weren’t performing optimally and it would take 6+ months to get the most out of those new architectures. Kepler = Fermi enhanced which is why it was expected to be fast out of the gate, but future performance updates were minimal which is exactly what happened for 680.

    Keep in mind AMD gained market share from NV last quarter.
    http://www.techpowerup.com/171198/Graphics-Add-in-Board-Shipments-Seasonally-Down-from-Last-Quarter-Reports-JPR.html

  • BestJinjo

    The graphics division is making $. It’s the computational unit (processor unit) division that is losing $ and that division includes APUs. Unfortunately, AMD may take the awesome ATI down with them. I hope it doesn’t happen as we need more competition but AMD has suffered from slowing PC market for laptops and desktops worldwide more than Intel since their CPU products are far less competitive. Look at the global slowdown in traditional PCs – desktops and laptops:
    http://www.xbitlabs.com/news/other/display/20121012154227_Lenovo_Becomes_World_s_Largest_Maker_of_PCs_amid_Market_Decline.html

    Most consumers now buy tablets and smartphones and even HD5850/5870 can play games. The market for GPUs has also shrunk last quarter overall, with NV losing 10% shipments. AMD’s $100 million inventory writedown is more or less the Llano / Trinity launch delay / inventory turnover disaster!

    OBR is an AMD hater overall. Never said a single positive thing about the company. I guess the guy has no clue that competition drives innovation and reasonable prices to consumers.

  • skr13

    When GTX 670 appeared, many HD 7970 owners switched to Nvidia card, due problems with games, like crashing, fps drop etc, the usual complains.
    That´s a fact… and rarely saw one Nv user go to AMD because of drivers.

    7970 GE is a bit faster but consumes way more than GTX 680, according to TPU is 48W. Even for a low margin, AMD card consumes more than GTX 580 also…
    http://www.techpowerup.com/reviews/AMD/HD_7970_GHz_Edition/26.html

    AMD need a new breath or will dissapear, what is gonna happen if people buy AMD card now and next year company won´t exist anymore?
    AMD CPU´s are to weak to compete with Intel High-end, they simple can´t made the jump to another performance level.
    AMD GPU´s are true competitor to Nvidia, but company can´t survive with that small market percentage.

  • BestJinjo

    Ya, but NV driver thing has been discussed for 10 years now. It’s like the IBM deskstar hard drives. HD7900 cards did have problems early on but HD5800/6900 cards didn’t. That’s my point. If people bought HD7000 series right away and they had driver issues, too bad. I had driver issues with my Fermi cards getting rocked by GTX285 in OpenGL games in Windows XP. I waited until NV fixed those problems and didn’t dump my cards.

    I agree that AMD has cross-fire driver problems and HD7000 series had driver issues but HD4000/HD5000/6000 series drivers were as bullet-proof as anything NV had those generations.

    I think a lot of it an AMD brand perception problem too. Why did ATI manage to sell GPUs for $500-600 and no one complained? You think ATI had better drivers than AMD 5-10 years ago? Even when NV got killed with GeForce 5 and 7, people still bought NV. That’s the point. Cards such as FX5200, FX5900, 8600GTS and GTX550Ti sold well and yet there were terrible.

  • Raghar

    The main problem with AMD drivers is they don’t give a damn about uncommonly used GFX cards. When majority of users are on W7 Dell systems, they release drivers for W7 Dell systems. (Which can be flaky anyway.) Basically AMD problem is when they release something new, they have problems. Regularly.

    Hey I have 8600 GT, it’s definitely not terrible, it’s able to run Dishonored at reasonable settings. (And overclocks to 1370+ shader clock on passive)

  • skr13

    Previous generations had drivers problems too, i speak for myself, with 4870 that i had.
    HD 5000 were affected by GSOD, the most common problem also present on HD7000, that obligates user to reboot computer suddently while gaming or even browsing internet, appear random.

    There are computers shops here that still sells 8800GT cards for 100 EUR, for example and people still buy them, totally uninformed of current market, shame!

  • xtco

    Its a realistic speculation imo.

  • xtco

    lucky lucky! i wish

  • BestJinjo

    I’ve used NV and AMD cards. I don’t know what these major problems AMD cards have. I seriously don’t. And most of the time when I got AMD cards, it was very close to launch. What is this “uncommonly used” GFX cards you speak of? Honestly you just keep repeating the same thing other NV users say. Have you actually tired using AMD cards regularly in the last 5-6 years?

  • BestJinjo

    What about the constant BSOD and RSOD of GTX670 cards?
    http://www.newegg.com/Product/Product.aspx?Item=N82E16814130787

    I find NV users almost never talk about driver problems. NV drivers were very poor for Shogun 2 and didn’t work at all for months in SLI scaling for Guild Wars 2. NV users didn’t open talk about it. Adaptive VSync stutter fest for a month? NV users kept quiet and few enthusiasts brought it up. NV users hate to talk about issues on their cards because it dispels the myth that NV drivers are better so they keep quiet.

    Even this generation, NV has not fixed performance issues in a single game where it trailed AMD. Why is it NV users don’t talk about it?

    Remember how earlier in the year you told me you cared about CryEngine (Crysis 2) and Frostbite 2.0 performance (BF3)? Well whatever advantage NV had is gone in Crysis 2 and BF3:

    http://static.techspot.com/articles-info/572/bench/Crysis_02.png

    http://static.techspot.com/articles-info/572/bench/BF3_02.png

    What about games where NV was trailing to AMD? Did NV improve performance there? No, it didn’t. So whose driver team is better actually? Where NV loses by 20-30%, it continues to do so 6 months after these issues were identified (Metro 2033, Crysis 1, Batman AC with MSAA, Skyrim with MSAA + Mods, Alan Wake, Bulletstorm, Serious Sam 3, Arma II games, etc.). NV has done nothing to improve performance in those games against AMD, while AMD has closed all the major gaps and even regained the lead in games like Dirt 3, Batman AC and Skyrim.

  • skr13

    AMD 12.8 Drivers notes:
    “Crysis Warhead: No longer crashes when launched in DirectX 10 mode”
    “Call of Duty- Black Ops : No longer hangs when loading a mission in DirectX 9 mode with Crossfire enabled”
    Only after so much years they fixed Crysis or Cod issues, that´s terrible…
    After announced 20-30% layoffs in Engineers, what you expect? Better Driver support? I don´t.

    Nvidia problems exist, but only affect few games and situations, in the other hand, AMD drivers can trouble with Internet browsers, Video Players, or freeze in IDLE state… and like i say before, many AMD users went to Nvidia mainly because drivers or something.
    IMO AMD knows that exist a MAIN problem in every generations, but can´t fix it, some incompatibility with chipsets, motherboard, but in the end they keep mouth shut.

  • kn00tcn

    stop pretending those 2 release notes existed on every driver since release, i have first hand experience with just fine warhead & of course if blackops CF was crashing, everyone would have heard about it!

  • kn00tcn

    blah blah blah, what kind of stupid conspiracy theorist are you

    other than 3->4->5 series, when has there been double performance in a generation? dont forget 3 series was the same performance as 2, so ati was behind in addition to the delay that the 2 series had

    something wrong with 15% ivy? i dont see how cpus can magically speed everything up per core, the tick tocks havent changed, 45nm quads were a nice efficiency boost over 65nm which is exactly what ivy is…. while nehalem, sandy, haswell, those are supposed to be the drastic architecture changes

    now is apple taking money out of people’s pockets with DOUBLE GPU performance after half a year? are you gonna complain about how it’s too soon? (& it is indeed too soon, even though ipad3 struggles with that retina resolution)

  • jon

    i dont get all the people crying about how a card uses 25 watts more. unless your talking about a laptop and battery life. if your so worried about power go use the intel HD 4000 igp or amd apu. give me a 1000 watt power hungry beast gpu.
    as long as it has performance that justifies the power im all for it. if
    im spending $500+ on a video card all i want is MAX performance at
    pretty much any cost.

  • skr13

    LOL, I just posted AMD notes, not my fault that they took many months to release a proper driver for games…

  • Wifebeater666

    If this is true, Nvidia will stomp all over AMD this generation.