Aorus Also Intros GeForce GTX 1060 6GB with Faster 9 Gbps GDDR5 Memory

[btarunr]

In addition to a new pair of GeForce GTX 1080 graphics cards with faster 11 Gbps memory, GIGABYTE, through its Aorus brand, introduced a new custom-design GeForce GTX 1060 6 GB graphics card featuring faster 9 Gbps (GDDR5-effective) memory. The faster memory lends the GTX 1060 a memory bandwidth of 216 GB/s over its 192-bit wide interface.

The Aorus GTX 1060 6GB 9Gbps Edition (model: GV-N1060AORUS-6GD) features a premium-looking custom board design with a 2-slot thick triple-fan WindForce 3X cooling solution, mated to a custom-design PCB that draws power from a single 8-pin PCIe power connector; and a back-plate. The card comes with factory-overclocked speeds of 1607 MHz core, 1835 MHz GPU Boost, and 9.00 GHz (GDDR5-effective) memory, against reference speeds of 1506/1708/8008 MHz. The company didn't reveal pricing.



View at TechPowerUp Main Site

 

[RejZoR]

This is not looking good for RX 580 with same VRAM as RX 480...

 

[crazyeyesreaper]

The faster memory isnt going to matter all that much. It contributes to an extent but not enough to matter vs say GPU clocks.

On the RX480 25% memory clock gave an average performance uplift of 8%

Put another way while the % numbers seem good the FPS gain average from 1200 Mhz memory to 1500 Mhz memory was 4 FPS. Now considering things such as CPU OC / memory clocks / SSD or HDD affecting data streaming etc. The gains are much smaller. Every little bit helps but essentially even with the faster memory. The gain will be smaller vs the core clock OC which is likely not going to change.

As is most GTX 1060s for example could hit 2400 Mhz memory which already boosts bandwidth to 230 GB/s

Basically it just means the new GTX 1060 reference cards will perform closer to Aftermarket 1060s available now.

With new compressions technology etc the performance gain will be essentially within a margin of error of new drivers vs older. Slightly faster but otherwise indistinguishable from previous designs unless they magically clock higher on the core.

Current designs hit the memory ceiling around 9600 Mhz effective. vs the 8008 Mhz stock

That means the default memory clock will be 2250 Mhz vs 2002 Mhz previously. Again since older designs already overclocked to 2400 Mhz on average. The gains wont really be noticeable unless they clock way beyond that. Even then I doubt the core clock limit will change much meaning miniscule gains its just a marketing bullet point.

 

[ZoneDymo]

This is not looking good for RX 580 with same VRAM as RX 480...

All about pricing, drop that RX580 by 50 bucks or more and boom, competition.

 

[Vayra86]

This is not looking good for RX 580 with same VRAM as RX 480...

RX480 is not as balanced as the 1060 or the 1080 are, and on most 1080's you can hit 11 Ghz anyway yourself. In the case of the 1060 I do see some significant perf uplift in a large amount of games because that card is bandwidth starved. But the RX480 is by no means in need of bandwidth, it's the Core that needs a push.

 

[jabbadap]

Uhm, why it's so long? GP106 is hardly gpu needing that size of cooler bolted on it. To think about it they could have leave those fans off altogether and add more heat pipes and fins to make it passive, even within smaller footprint.

 

[ratirt]

I don't think that comparing AMD and NV products with memory frequency or core clocks is a good idea. They are totally different. Performance to performance is the only thing that those 2 can be compared. RX 500 series is promising I think. It would really matter what those new Polaris based cards can do performance wise. Maybe they have some slight changes in architecture. I guess yeah since they will eventually run a bit faster and still be more efficient in terms of power than RX 400 series. The question is how those changes would impact performance.