ASUS GeForce GTX 1080 Ti STRIX Specifications Surface

[btarunr]

Ahead of its launch, specifications of the GeForce GTX 1080 Ti STRIX graphics card by ASUS, appeared on the web thanks to a leaked press-deck. The one spec on most people's minds is the factory-overclock. Out of the box, in the so-called "gaming mode," the card is clocked at 1569 MHz core, 1683 MHz GPU Boost, and an untouched 11 GHz (GDDR5X-effective) memory. The "OC Mode," a clock-speed preset that you activate with the included GPUTweak II software, runs the card at 1594 MHz core, 1708 MHz GPU Boost, and 11.1 GHz memory.

It turns out that the DirectCU III cooler deployed on the GTX 1080 Ti STRIX is indeed beefier than the one found on the GTX 1080 and GTX 1070 STRIX cards; with ASUS claiming up to 30% lower temperatures, thanks to a 40% increase in surface area of the heatsink. ASUS took advantage of the heavier heatsink to lower fan-noise, which it claims to be significantly lower than the reference-design card. The company is also using a newer thermal interface material (TIM) called "MaxContact," which due to its particle density, can get up to 10 times flatter than conventional TIMs, for lower thermal resistance posed by the TIM, and higher conductivity. Display outputs of this card include two each of HDMI 2.0b and DisplayPort 1.4 ports.



View at TechPowerUp Main Site

 

[jesdals]

This could become my first non laptop Nvidia card if VEGA do not delivery on at least 16GB and good performance vise. My current monitor setup would need at least 3 DVI/HDMI to avoid use of dongles.

Might consider going green if VEGA is to far off performance on this card.

 

[RejZoR]

The way how Vega will work, they don't need 16GB VRAM. That's the whole point why they introduced HBC. They'll have 2 stacks of 8GB HBM2 and the rest will be offloaded to system memory. You've seen the DX Mankind Divided demo with just 2GB available VRAM but enabled HBC. It was smooth. The question now is whether this is 100% automatic or it needs special code in game. Because if it will, that could be a problem. But if it doesn't and it's automatic, then AMD solved whole lot of cost issues while delivering nearly same performance as having full onboard memory. Again, we'll see when we get the card and people poke it to see how HBC system works. But I know the basics from the past and I'm very optimistic about it based on that knowledge...

As for Volta, we actually have even less info than about Vega. But we do know that GTX 2080 can't possibly be slower than GTX 1080Ti so that's already a bar set very high. Which is good for consumers and reason to poke AMD in the eye and demand from them to be even more aggressive either with prices or performance. This is why competition is good

 

[XiGMAKiD]

I can't see where in the image posted that says MaxContact is a TIM, but I can see where MaxContact is being addressed as "an enhanced copper heat spreader"

 

[EarthDog]

Sure they surfaced... the boss has and reviewed it.

 

[Aenra]

This kind of money, if i was out for one, i'd be waiting for a water-cooled model.
Too much weight, the same awful fan positioning (by necessity, but nonetheless), the same consequences in as far as heat and dissipation may be concerned. I can understand the reasoning behind fans, but on cheaper models. At this level, i personally find no excuse.

But of course, none is needed. It's out, people buy. End of story..

 

[mikka1986]

I don't get it, MaxContact is direct metal to die contact without TIM?

 

[Kanan]

I don't get it, MaxContact is direct metal to die contact without TIM?

I think the copper is just better "flattened out" compared to before, improving contact. I don't think it's a TIM, but I'm sure it still needs TIM for better contact as usual.

 

[D1RTYD1Z619]

I really want this card but this SOB is 3 slots. Im going to rebuild my biatch in a In Win 303 case I bought. Was looking to get a matx mb to have MB header clearance and bottom radiator clearance. And in getting 2 of these might not fit properly.

 

[Caring1]

I think the copper is just better "flattened out" compared to before, improving contact. I don't think it's a TIM, but I'm sure it still needs TIM for better contact as usual.

This^^^
It just looks like they highly polished the contact surface to a mirror finish to increase contact area.
People have been doing this for years with CPU's to aid cooling, no reason it can't help GPU's too.

 

[Jism]

This^^^
It just looks like they highly polished the contact surface to a mirror finish to increase contact area.
People have been doing this for years with CPU's to aid cooling, no reason it can't help GPU's too.

Any normal surface that has'nt bin polished, is as described rough. Thermal paste adresses this by filling up those gaps. But better thermal results can be archieved when polished with a shine.

That is the idea basicly. It's nothing new. In best situations it can save a few degrees at max.