Quote:
Originally Posted by Feuer Frei!
Not really.
I was posting it in the hope that people would get caught by the April Fool's joke.
I guess my mission failed   |
I knew this was a joke right away for several reasons:
First, the launch of the 6990 can hardly be called a "huge success," as it is a niche product with a price range which is above what all but the hardiest of computer enthusiasts will avoid. It was a success in terms of giving AMD the ultimate performance crown in certain applications however, so a biased site might have included the reference. Definitely not outside the realm of possibilities here.
Second: AMD's current naming scheme uses the first digit of the four to indicate series, the second to indicate position within the series and the third to indicate clockspeed in that position, hence: 6970 is the top end of the 6000 series, while the 6950 is a downclocked version. The 6870 is of a lesser design and is slotted in below the 6900 series cards, etc. For a card that is supposed to be better than anything else out, 7770 is an odd designator. Keep in mind that the 5770 was AMD's high-mid-range card for the 5000 series. The joke would have made more sense if you had suggested either 7870 or 7970 as the name.
Third: "Cuban" core
For the record, the code names for the next set of AMD cards are based on Southern Islands.
Fourth: "6GB of GDDR6 on a 384-bit memory interface"
OK, most single GPUs are currently being released with 1-2GB. Any more RAM than that simply doesn't make all that much sense, especially given that it would serve no purpose. Most displays today are 1920x1080 and you don't need more than 1-2GB to buffer for that resolution. 6GB would simply eat into the profit margin for AMD and raise the card's price to astronomical levels. Aside from that, 6GB of RAM chips would likely make the card large enough so that it would not fit in an ATX case.
GDDR6 does not currently exist. GDDR5 is advanced as it gets.
As for the 384-bit memory interface, AMD has traditionally stuck with buses that are power of 2 friendly (i.e. 32, 64, 128, 256, 512). They've never used an off power bus width. nvidia, on the other hand, has used 320 and 384 bit buses from time to time (e.g. 8800GTS, 8800GTX and GTX 4/580). Thus, a bus width of that size makes more sense to come from nvidia.
I'm not trying to rip into your joke here (and in all seriousness, props on making the attempt), I'm just trying to explain how I knew it was a fake story so readily.