The power/temp requirements make me think that they knew they were not going to perform as well as the ATI 5870 and so they overclocked and then got stuck trying to deal with the power/heat issues. .
Quite the opposite in fact. The original specification was for (I believe) an SKU with clocks of 725MHz core/ 1450MHz shader/1050MHz effective memory. The fact that, with review samples at least, the cards can fairly easily reach those clocks through overclocking obviously indicates that the cards have been underclocked to keep them within TDP.
.. I know there were yield and other issues as well, but it seems odd that they would have let this pass through given all the the recent attention to higher efficiency and lower power requirements.
The alternative being what? Having no enthusiast class cards in retail for another 6to 12 months ?
At least with these releases enthusiast grade card consumers now have a degree of competition in the marketplace, and for some of us we can see where gaming might be headed in future.
Before you ask "what competition?", I'll add a few observations.
Power hungry....Yes, but hasn't stopped SLI and Crossfire ownership (nor single card versions: HD 5970 -294w TDP, GTX 295 -289w TDP, HD 4870X2 - 290w TDP , HD 4850X2 -250w TDP) and overclocked CPU's/GPU's.
Noisy....Yes, kill two birds with one stone and buy the waterblocked version, or buy some noise cancelling headphones/headset.
Expensive...Yes, but thats the nature of enthusiast/performance components
Cost effective...Hell no, but tell me of a enthusiast grade product that is?
I could buy the cost effectiveness argument if PC gaming was in any way cost effective outside of a budget system at medium to low resolution-But it is not.
So would I buy one (or two) ? No, not until DX11/tesselation makes a quantative difference, and by that time both AMD and nV will have more refined products available.
My personal advice to customers requiring good GPU performance is to Crossfire HD 5770's. If they can afford to game on a 2560x1600 IPS monitor then in the main, cost becomes somewhat immaterial. Prime consideration becomes "is the new game playable on zero-day ? and do I have to lower any settings?"
One thing I've found is that people who hand over large wads of cash for both components and games get really irked over having to turn down the settings.
Would I consider nV's future offerings based on these cards? Definitely.
AMD have had GDDR5 for 2+ years. By now the memory controller should be refined, yet the number of cards that have a memory bus larger than 256 bit...zero.
At their second attempt (after the GT 240) nVidia have a working 384-bit GDDR5 memory controller-the difference is vivid in games such as
Metro 2033 and/or heavy antialiasing, they also have 8 x multisampling AA working out of the box....no too bad a start for what is essentially a proof-of-concept part.
Ultimately, the argument is less about the hardware than how it will be implemented.
All things being equal, AMD and nV will both be sponsoring game development optimizing for their own products. The difference is that nVidia's TWIMTBP is both well established and linked to many AAA titles, so it's a fair assumption that nVidia will be putting development funding into games that feature options for heavy tesselation, PhysX and DirectCompute- area's where AMD's 5 series don't have an answer. AMD's answer was the recently announced game dev funding at GDC 2010 - the "
Gaming Evolved" brand. If the initiative is followed through with sustained funding then the balance tilts clearly towards AMD. If however it turns into another smoke-and-mirrors campaign...
Richard Huddy and earlier defunct “Get in the Game” (GITG) program
How active are ATI in helping GITG developers write general code and shaders that runs well on that ATI hardware, using those drivers?
[Richard Huddy]: We've made quite a few changes recently to how much emphasis we're putting on GITG, and as time goes by you should see more from us. But it's really important to us that this genuinely helps games developers and players. We don't intend to invest money in getting our logo into the splash screens of games since we thank that is amazingly unproductive. Advertising like that is only aimed at making the advertiser richer
28 June , 2004
then the gaming future might be a little less clear cut than some are expecting.