Monday, March 18th 2013
NVIDIA Working on Second GK110-based GeForce Graphics Card for Summer
NVIDIA may decisively hold on to the single-GPU performance lead, with its GeForce GTX Titan graphics card, but at roughly $1000, it could attract a very small market. According to a SweClockers report, NVIDIA is looking to woo gamers just ahead of Summer with the second GK110-based GeForce GTX graphics card. Similar in specifications to the fabled Quadro K6000, the new SKU could feature 13 out of 15 streaming multiprocessors on the GK110 silicon, working out to 2,496 CUDA cores, 208 texture memory units, a 320-bit wide GDDR5 memory interface holding 5 GB of memory, and 40 ROPs. Given that there's a deep ravine between the ~$450 GeForce GTX 680 and ~$1000 GTX Titan, NVIDIA could pick a price-point in the middle. The report claims the new SKU could launch some time between July and August, 2013.
Source:
SweClockers
74 Comments on NVIDIA Working on Second GK110-based GeForce Graphics Card for Summer
I'm not liking this mixture of new cards with the current series of cards as it seems to give them an excuse to charge more for them.
I will Buy 3 for 3-Way Sli
The same mistake they did with Fermi and GTX 400 series ...
I can see the new Nvidia naming scheme now
GTX Titan
GTX Diminutive
GTX Half-Pint
GTX Runt
GTX Pygmy
Only reason I can think of for people to buy these cards right now are professionals who need the compute power, but don't want to spend $4000 for a card to do the same things. Oh and for people who want the sexy looking cooler! :D :laugh:
Other then that there are milking every cent they can get from the GK110.
Sexy looking coolers you say ?
Just get lable paper and print out a picture of Kate Upton and slap that on your current cooler. Bang!!! save $500 dollars right there. :rockout:
P.T. Barnum said it best or was it David Hannum... ;)
I can buy two 680s for the price of one Titan. Nvidea wants to burn through their poorly performing silicon, so they want to sell a "stripped down" version of Titan. Realistically, the stripped down card will be specced higher that the 680 but priced higher than any performance gain it might possibly have.
Who in their right mind will buy this?
Looking at a limited production run, insane costs, and generally unappealing cost to performance ratio I can't see a market. The Titan can claim absolute superiority. The 680 can claim ownership of the single card consumer market. What does the stripped down Titan claim?
The Titan owns all single chips, yes. Then the 7970 can arguably be called the most realistic king of the consumer market. The 7970 at comparative speeds (call it GHz if you will) is superior to the 680.
pointless :shadedshu
For the people willing to pay 600+ for this.......WTF :wtf:
Then they thought oh shit the GK100 isn't that fast compared to 7970 we better build in boost overclocking to make it look better. They had to max out the GK100 chip to get it to compete.
Then they release this titan for $1000? What if the 7970 was 15% faster than it was? Would they have released titan for $699 up against the 7970?
Anyway its all fun and games I guess
Actual story from the original source: Note the complete lack of reference to GK 110 or GK 100 ( a GPU that doesn't exist btw), nor the GK 104 and GK 107 which were the first Kepler series GPUs that achieved series production. 25% ? Generally with a small sample set numbers can mean whatever you like
A quick look at the 28 available reviews I could ( TPU, Anandtech, HardOCP, Guru 3D, Bjorn3D, HardwareLUXX, ComputerBase, Hardware France, Hardware Canucks, PC Perspective, OCC, Sweclockers, Lab501, Hardware.info, PCGH, Alienbabeltech, Tech Report, Hexus, Hot Hardware, HiTech Legion, TechSpot, PC Gamer, X-bit, bit-tech, expreview, Tom's, Hot Hardware, MaximumPC, Linus amongst them), says that the Titan is 33% faster (using highest game i.q. benches only) at 1920 / 2560, and 45.5% faster at 5760...and that taking into account some oddball benchmarking results ( the 7970GE gaining framerate from 2560 to 5760, or not losing any framerate between 1920 and 5760 in AC3for example, or lack of driver support in Tomb Raiderfor example) in small sample sets* that skew the mean value of the averaged results.
* Only three sites benchmarked AC3 at 5760x1080:
TPU: Favoured the HD 7970GE by 44.99%
ComputerBase favoured Titan by 36.73%
HW Canucks favoured Titan by 49.09%
I'd be inclined to delve into the numbers a little further -along with any other possible workload the cards' user base might employ...after all, I seem to remember that many AMD proponents were happy to justify a $549 initial price tag for the 7970 based on its hashing ability.
Besides, if you're using a performance-per-dollar metric based solely on gaming you obviously aren't part of the intended market for the card. Given your distain for anything Nvidia it beats me why you even bother with the argument- it's not as if you'd buy an Nvidia card even if it came with a 100% rebate voucher.
I would have been GF100 all over again, infact there is no GK100 and I'm fairly sure the reason are the very poor early 28nm yields.
I mean 7.1 billions transistors on a new manufacturing node? Would have costed an arm and a leg to produce in a sizable quantity for their GTX 670/680 lineup.
I think that if the 7970 was faster than it is now, like 10-15% more then Nvidia would have launched GK104 as midrange chip and they would have launched the GK110 as GTX 680 when they could have had reasonable yields.
Nvidia would have probably been without a single GPU capable of fighting AMD toe to toe and they would have probably used a dual GK104 solution.
But meh things have gone differently and we can only speculate.