So the memory is 1280mb x2?
I'm a little confused by the description.
Printable View
So the memory is 1280mb x2?
I'm a little confused by the description.
so what is it??? dual 560?? or dual 580 .... LOL
1GB total memory? WHAT?
Take a look a the back of the card:
http://www.techpowerup.com/img/11-01-07/71b.jpg
Notice the area which coresponds to the position of the two GPUs. Now compare to:
570: http://www.techpowerup.com/reviews/A.../back_full.jpg
460: http://www.hardwareheaven.com/review...x-460_back.jpg
Which is more likely? :)
Someone commenting on the Fudzilla link says "EVGA mentioned GF11x - they did't specifically say GF110"… interesting.
I think it'll be GF114, using cores with clocks similar to or slightly below those of the 560 and cherry picked so two will fit on a 300 W dual chip card.
Stop the dreaming: that chip configuration TechPowerUp makes no sense. How do you have 4 on one side then 8 on the other without those traces/soldering points anyways?
Most likely its dual sided RAM, 8 chips total, so 8 = 1024 = 1GB RAM on 256-bit
Besides, 12 chips is 1536 MB, 10 is 1280, which TUP wrote incorrectly. Besides, the back side pins say it all:
http://img715.imageshack.us/img715/4008/gtx590.jpg
Where does the SLI connector fit in then if its two 114 cores, wouldn't dual 104/114 mean no external sli. The back gpu area doesn't distinctly resemble either a 104 or 110 gpu based card.
Maybe if we could see the back of a 114 based card.
Evga have explain, this card is from it's own, it will not appear on Nvidia listing( anyway not this one, this let open to a dual gpu SKU ). Look like they can do what they want... you are surely right--
Anyway this can be anything: at the moment the card can be cooled efficiently, they can design the card with what chips they want, but i believe more they aim to a card for Surround, more of aim the higher perf.
A good reason for don't use the GF110, is the cost, 2x 580 chips will bring a big price on the cards, then the clock, with lower chips, they can clock it alllready well, and keep some room for overclocking...
What I see is a useless product.
Since it is clearly not 2x GTX580, it will be bu** f**ked by 6990, worse then shower scene from American History X...
What annoys me is...
AMD clearly has a leap that they can actually make products that does not require a powerplant for working, allowing nice multi GPU products... Yet still, having this leap over nvidia, no one has come up with the bright idea "Why not start optimizing for 8 GPU?"..
Who on gods earth would ever want to buy 4x GTX580 if being able to use 8 GPU's from AMD? (4x dual GPU) Answer is simple; Marketing department from nvidia, would be useless to each and every other living person on planet earth...
If its a 4 GB, 2 GB effective card, it can still have uses. Would be a good folding card or cuda or tesla card on the cheap.
I really do hope that this card will have at least more than 2.5GB of memory, because 1GB per core just isn't cutting it anymore.
Hitches that are to be overcome with ease.
Sure CPU limited, not really, depends on scenario.... And there is STILL PR value in delivering more then 40% higher 3Dmark2011 score then competition
My 2x HD5870 SOC kind of disagree...
- Due an old classmate, I actually decided playing a full wooping 10 hours computer in 2011, Left 4 Dead 2 in 6048*1080, 2x AA, 8x AF isnt really a concern, nor is 4x AA, 16x AF
I agree! I have been looking to upgrade from a GTX 295 Single PCB and have been waiting for the next nVidia Dual GPU monster.
If it is a dual 560, providing hey have 2GB of VRAM and the 384pipes per GPU it would not be a bad card at all.
Who knows, maybe nVidia are doing to do a new SKU, perhaps a GF110 with 448 pipes per GPU.
Remember nVidia did that with the GTX 295 and then released the 275 afterwards. I might be wrong, but essentially a GTX 295 is two GTX 275's on one card with a slightly lower clock rate to reduce power consumption and thermal limitation.
I really really do hope that we see a 448piper with ~2GB per GPU. I keep thikning that IF the GTX 295 had 1.7GB per GPU it would have been a monster card, instead of a fast card.
John
Not before PCI express 3.0-4.0, more lanes for the GPU's, and with faster CPU's. This will be bottleneck in all parts...
huum do you think to an Evga quad-8 SLI dual CPU classified based on socket 2011 ? I try imagine the beast with 2x 6 cores on it.