Straight up I think Nvidia is going to bring us a really great card. No interpretation necessary. Gut feeling.
Meh, I don't think the chart is real just because of characteristics of the person that posted it (1 post, banned a chiphell), but anyway.... the real question is does it beat 470 sli?
EDIT: Let me rephrase that, I guess you technical can play - but unless you're going to finish a whole game in one sitting, removing the Save option is a huge dick move.I don't use any of the online / multiplayer features either. I just use a offline account and save achievements locally.
I guess there is a free online F1 multiplayer game comming from Codemaster's this year.
Exactly what I was going to say. I see a lot of posts in this thread and others that echo "oh yeah, well Kepler's going to be even better!", like a threat from a deprived child. NVIDIA already lost this round to me, and many others from the posts I see all around forums. By the time Kepler comes out, I will have been enjoying great performance for months, and there's no way GK104 can compete with the 7970 I've already purchased. I'd be interested to see how many people cheerleading Kepler's release will actually be purchasing one.ehh... I think you meant it's fun watching the early adopters currently enjoy their real cards, who's performance (and existence) has been proven, not rumored, faked, hoped for, dreamed of...?
Exactly what I was going to say. I see a lot of posts in this thread and others that echo "oh yeah, well Kepler's going to be even better!", like a threat from a deprived child. NVIDIA already lost this round to me, and many others from the posts I see all around forums. By the time Kepler comes out, I will have been enjoying great performance for months, and there's no way GK104 can compete with the 7970 I've already purchased. I'd be interested to see how many people cheerleading Kepler's release will actually be purchasing one.
You miss my point though: I've already had that performance for a month, and will have had it for several months by the time Kepler is here. So yes, if Kepler comes out and it's as fast as a 7970, it doesn't really affect me. If it's slower but has better price/performance, again, still doesn't really affect me. Therefore I, and many other 7970 owners, aren't impressed by Kepler, understand?Regardless of this "leak", all things point to the fact that GK104 will indeed compete with the 7970. We will find out very soon, some say next month.
Not really, most of the "leaks" have been determined to be fakes. About the only semi-concrete (see what I did there) info is from Charlie, and that is very suspect to some.Regardless of this "leak", all things point to the fact that GK104 will indeed compete with the 7970.
Not really, most of the "leaks" have been determined to be fakes. About the only semi-concrete (see what I did there) info is from Charlie, and that is very suspect to some.
Just hurry up Nvidia I can't take much more of silly season.
You are right except the specs that are linked to this thread. Since these have been reposted by numerous respectable sites we can assume they are as close as they can be to the original. Based on this we can debate the possible performance of the chip. Of course all we have are speculations but as Charlie and others reported NVs' AIB partners in the far east have the goods, NV's marketing is in action for pre-launch so we can expect more of these "leaks" pretty soon.
You miss my point though: I've already had that performance for a month, and will have had it for several months by the time Kepler is here. So yes, if Kepler comes out and it's as fast as a 7970, it doesn't really affect me. If it's slower but has better price/performance, again, still doesn't really affect me. Therefore I, and many other 7970 owners, aren't impressed by Kepler, understand?
Wow, I guess those who have owned GTX580's and HD6970's for the past year are super fortunate then. I understand the need to justify/defend a purchase for a consumer. Nobody wants to feel like they just threw money into the garbage. I do not think this "argument" has much relevance though because like I said, people have been enjoying other cards for a year or more. And here you are touting one month soon to be 2 or 3. Not flying much.
There is a difference. GK104 isn't likely to be on average faster than 7970, but it's quite likely to bring the price down. For many 4 months are preferable to 50 or so bucks, because they're going to jump on the next best thing in a year anyway.
But where are you getting any of your information from? Surely you don't believe any of the rumors you've seen about it?
You saying GK104 isn't likely to be on average faster than 7970 is based on what? die size? We don't know that. Shader count? We don't know that either. We know nothing so your comment has no relevence at all. There isn't even enough info for an educated guess. Nothing to go on. I wish we all knew more.
Easy. What we currently more or less know:
1)Power efficient
2)~300-350$
3)Comes out slightly more than a month from now.
4)Performance generally around the 7870 except in highly optimized scenarios, where it can perform as high as a 7970.
Adding 5)Is clocked as high as possible to meet the 7970 in performance, yields and power draw be damned, goes against 3 of 4.
We don't know any of these things at all. (1)Could be a power hog, (2)could be just as expensive as 7970, (3)could come out next week or six months from now for all we know, NO idea where it will perform. (4 & 5).
Square 1?
We don't know any of these things at all. (1)Could be a power hog, (2)could be just as expensive as 7970, (3)could come out next week or six months from now for all we know, NO idea where it will perform. (4 & 5).
Square 1?
Crap Daddy said:If one really wants a card, is a first day buyer with all the consequences (driver related) and has the money to act there's absolutely no problem. Out of all the variables with this first Kepler chip the least we know is the price which was referenced only once before all these predicted specs and performance leaks. Let's say it's between the 7950 and the 7970 (I really can't think it will be slower than a GTX580), based on the market right now it shouldn't come as a surprise a 500$ price tag.
I don't see you advocating everyone to buy Tahiti, because hey, Nvidia might not even release their 28nm cards this year, eh?
From the looks of it, Kepler brings quite large arch changes and so is just as prone to problems with drivers as SI.
And you haven't seen me tell anyone not to either. Have you. So what is the relevance?
From the looks of WHAT? LOL. Jeebus.
Keysplayr said:Wow, I guess those who have owned GTX580's and HD6970's for the past year are super fortunate then. I understand the need to justify/defend a purchase for a consumer. Nobody wants to feel like they just threw money into the garbage. I do not think this "argument" has much relevance though because like I said, people have been enjoying other cards for a year or more and able to game the latest on highest settings. And here you are touting one month soon to be 2 or 3. Not flying much.
Keysplayr said:We don't know any of these things at all. (1)Could be a power hog, (2)could be just as expensive as 7970, (3)could come out next week or six months from now for all we know, NO idea where it will perform. (4 & 5).
Spinning like a windmill.
And two posts later you managed to prove his point with this:
You can't selectively believe rumors in one post, and then don't when it's suitable to your rhetoric.
Certainly possible. But I do remember prior amd releases where it was suspectednot that I'm holding my for this to happen, but... There may be a reason why there are no WHQL drivers ready for the 7k series. Drivers could easily plop another 10-15% performance increase just in time to ruin the Kepler realease.
I know this gets said almost every generation but maybe, just maybe.. /speculation
I don't get excited about mid-range cards, only parts that deliver new levels of performance not seen before. So the original OP and specs of this thread are not exciting, as that card is not going to do that with a 340mm die.