I was just pointing out my view on the whole theft of Nvidia's tech comment... I'm thinking Sapphire right now, seeing as how they're the ones who have really worked the closest with ATi in the past. Who knows who they'll open the opportunity up to though... Could be a rather large group if the technology is accepted by the general public.
There's one thing that bothers me. If Ati is ready with their dual cards, game developers will start making the games to take full use of their power. Which means, that you will only be able to play on everything max. if you got two cards. Us poor fellows with a single card will have to play on medium or low settings.
Well, maybe, but that means much more realistic looking games. I'm sure even if you played on the lowest settings you would be able to get at least the same quality that you do now. I'm one of those people who won't mind paying a lot for a rig, including 2 video cards, if it means better looking and playing games.
hmm i don't know if i can agree with the iq on Nvidia drivers coz when i had my 5900 XT i was disapointed in image quality and fps and performance from the drivers and the card seemed poor to me but hopefully thing changed but that could just be me but i do admit ATI drivers also need work done to improve coz i have noticed on older games that i play, quake 3 performance and lagg problems occur so non
i kinda agree, how will this work because if the performace is scalled down to the lower end card then that mean a lot of potential perfrance is lost and if ur in the scenario with a 8 pipe card and a 16 pipe card then that mean ur losing almost half the performace of the 16 pipe card and what this means is you would be better of getting 2 8 pipe cards or 1 16 pipe cards.
Aparently the article says that it would be possible to run two different cards (although not supported right now), but you really wouldn't get your money's worth in performance, so I'm sure we won't see this kind of support.
ATI's method will be SuperTiling. ATI has more experience than NV when speaking about rendering with more than 1 card. ATI has been working with E&S and there are also SGI plataforms that uses 2xFireGL cards with Supertiling. This happened before NV SLI hit the market. http://www.ati.com/companyinfo/press/2001/4434.html http://www.theinquirer.net/?article=22074 One of their products: http://www.hwacreate.com.cn/english04/Products/SST/simFUSION.htm With that said, I think ATI will just implement their experience with multi graphic cards into the gaming area. There are known problems with Supertiling... but ATI may have better answers to those problems than ourselves (and NV marketing managers of course... ).
Yeah, but we'll have to see how ATI product will fare with the gamers, as that is what really counts.
Hmm... Couldn't they theoretically include some sort of utility in their drivers where you could adjust settings according to the power of each card? Like you could say, well, I've got a X800 Pro in Slot 1 and an X900 XT (*drools*) in Slot 2, so maybe I want to tell the system to only request that 25% of the tiles be rendered by Slot 1, and request the other 75% from Slot 2. Dunno if that's possible or not, but it would certainly be a good way of doing it in my mind.
Probably not, at least not right now. I'd think you'd have to give the cards the same amount of work.
True, it's just a matter of time What I'm trying to say is that the method ATI is using is not bad at all. NV tried to use Supertiling without luck (don't know the real reason about that). We always have 2 options: It'll work as expected or it'll work worst than expected There's an interesting discussion in other forum (I don't know if I'm allowed to say where ). About using 2 different cards, I personally think it's a bad idea (if the cards are too different of course). If you are planning to use a X800Pro with X800XT PE, it should work (I'm not saying it's easy or possible), but you'll get 2xX800Pro performance which is way better than 1xX800XT PE. Conclusion: we always have to wait for the results (I hate waiting ).
I think nVidia didn't like supertiling because many of the edges of the tiles sometimes share pixels between the two cards, which can cause unneeded calculations on both cards.
Yeah, but I don't think he read the article. Where did you think I got that information, as I surely didn't know myself?
Yes, I've read it But I don't believe any marketing manager from any company I just don't trust them. If something goes wrong, the market manager ALWAYS has the correct answer not to blaim its company.
Exactly When AMR is ready we'll see how it behaves with games that use effects that requiere the information of one tile that is being rendered by the other card. I don't think it's a big deal. We should research how supertiling behaves in E&S or SGI products, but I'm not sure if we'll be able to find anything
