Nvidia GTX 1080 reviews out

[VR-Hunter]

NDA was lifted on reviews. Great improvement on fps and 3DMark. Will it beat 980Ti in VR? Time will tell.

 

I have 980Ti and I gonna hold on it until we have some comparison between 980Ti and 1080.

 

Here are some reviews.

 

[ame]

[/ame]

 

http://www.guru3d.com/articles-pages/nvidia-geforce-gtx-1080-review,1.html

 

http://arstechnica.com/gadgets/2016/05/nvidia-gtx-1080-review/

 

http://www.engadget.com/2016/05/17/nvidia-geforce-gtx-1080-review/

[SkateZilla]

As I said in the other thread:

 

After reading 3 reviews (now 7), definitely gonna keep to looking at AMD.

 

 

- Baseline support for Async / DX12, But they Developed their own GameWorks/nVidia Version (Dynamic Load Balancing)

 

- Lack of 3/4 Way SLi Support

(You will have to unlock the ability with key from nVidia. Side effect of continuing to use dated SLi bridge interface instead of xDMA like AMD.)

( Even after unlocking, it will not be supported by nVidia, it will be up to developers to integrate it properly into their engines)

 

- Baseline support for DX12 EMA, While they continue to try and push SLi, even after neutering it.

(Basic Support for DX12 EMA included in drivers, No Support outside of that)

 

- No Support for Adaptive Sync at all (Despite Hardware Capable of Supporting it)

 

 

 

So, While ignoring support for industry standard, they continue to develop their own proprietary tech, which is historically only used in their press event tech demos.

 

 

How many studios actually still use GameWorks as their Primary API?

 

I wonder how well nVidia locking out 3/4 way SLi is gonna go over with the community, lol.

(Can unlock w/ a Key from nVidia, but they will not support it in drivers)

 

its Like nVidia inherited 3dfx's poor decision making.

 

On the bright side the raw performance is there.

 

 

Multiple reviews, same underlying tone, NV is shooting the industry standards the finger and developing their own proprietary tech to try and push onto developers by locking it to their GameWorks API.

Edited May 17, 2016 by SkateZilla

[Enduro14]

Ya I would base my decisions on fact not opinion....

[SkateZilla]

Ya I would base my decisions on fact not opinion....

 

The Performance is definitely there.

[Gladman]

Really, how many people run 3 or 4 cards. The fraction of the market would be irrelevant, so who cares. The dx12 vs DCS dx11 does note some concern but I doubt it will be significant given the fps achieved with much lesser cards.

I'll still wait to see what the competition can do but I see no reason to believe the 1080 won't be an excellent card for our purposes.

[SkateZilla]

In the Consumer Market, not that many.

 

in the Benchmark Market and Pro Sims Market a lot, but they comprise of a small % of total market.

 

The Limit of 2 Cards is a Band Aid slapped on the cards because the SLi Bridge doesnt have the bandwidth to send the 4K/90Hz images fast enough.

 

My judgement is based on business and development decisions, as previously posted about (Large Die Chips, etc, and now lack of Industry Standard while cramming their Proprietary Features down consumers throats.)

Which is What eventually Doomed 3DFx. Who had unmatched 3d Performance in the 90s and Alienated themselves from OEMs while Cramming Proprietary technologies into their GPUs, and eventually being spanked by both ATi and nVidia.

 

Anyone who witnessed the downfall of 3DFx Interactive can match the same dumb decisions being made by nVidia to them.

 

A. Shifting to "Pure Performance Only" (Voodoo 4/5 was supposed to be pure performance and was beat by GF256), Couldnt Compete with Entry Level Price Points.

B. Shifting Production to Retail Performance Cards only, Removing OEMs, Forcing OEMs to Buy from nVidia, 3DFx Loses Income from OEMs which was previously their major source of revenue.

(nVidia Shifted to High Performance GPUs and Alienated Consoles, which make of up 75% of the Market, and that 75% now belongs to AMD)

C. Started Ignoring Industry Standard and MS DirectX (When they announced they'd only support OpenGL or Glide), nVidia is doing the same w/ Basic Implementation of DX12 Features, while the more advanced features are part of the GameWorks API

 

Lets see what happens when they convert a console port to a GameWorks DX12 Title... and hopefully it doesnt completely botch like other titles the last 2 years have.

 

 

Im not gonna wave I told ya so's in anyone's face,

But I can almost guarantee a few months after the 1080s are out, everyone that spent $600+ on Reference or AIB Partner Cards are gonna go donkey kong when they finally have enough laser cut GP100 Chips to sell 1080Ti's for $700 - $1000 w/ nearly 80% more performance over the GP104 Chips.

Edited May 17, 2016 by SkateZilla

[Toxin1]

Interesting analysis by RoadToVR on the 1080s current relative VR performance to the 980Ti.http://www.roadtovr.com/nvidia-gtx-1080-benchmark-review-performance-head-to-head-against-the-980ti/3/#vr-benchmarks

[j0nx]

Doesn't look like currently the 1080 will offer much over the 980ti even in VR. My Gigabyte gaming 980ti runs all day at high 1400's at 67c with my 75% fan curve. It will run 1500 easily and I have but choose not to do it on a daily basis because it didn't give me anything over what I feel is safe. I recently sold my 970's and have some change in my pocket from those so upgrading to the 1080 would only cost me $300. I just don't think the price/value proposition is there with this card at $700. I'll revisit this when the AIB cards come out next month. $550-600 is all this card is worth.

[SkateZilla]

wait for the 1080ti

[j0nx]

Well I need another high end card for my daily gamer PC because I have to pull the 980ti out of it to put it in my VR rig which will be a dedicated VR/sim machine when my rift arrives tomorrow. I am taking that machine and putting it in my Volair simming cockpit that I am picking up next week. That leaves me with a single 970 that won't cut the mustard on my daily driver 3D surround PC that I use for all my other games. I don't mind spending money but I refuse to throw it away on gimmicky FE overpriced cards. I know, first world problems...

[SharpeXB]

There are some VR tests here

http://www.roadtovr.com/nvidia-gtx-1080-benchmark-review-performance-head-to-head-against-the-980ti/

 

"If your focus is purely VR however, the upgrade argument is perhaps less compelling at this stage. With application developers judiciously targeting performance well below the capabilities of the GTX 1080 for very good reasons, it’s hard to see the need for 1080 levels of raw grunt for a little while."

Edited May 17, 2016 by SharpeXB

[Vivoune]

It is so highly priced, I'll just wait and spend the same kind of money on the ti at the end of the year. And it gives AMD a chance to win me over in the mean while, but I'm not holding my breath.

 

"If your focus is purely VR however, the upgrade argument is perhaps less compelling at this stage. With application developers judiciously targeting performance well below the capabilities of the GTX 1080 for very good reasons, it’s hard to see the need for 1080 levels of raw grunt for a little while."

 

"Hey, you don't need a powerful GPU, developers only code games for people with less powerful GPUs, and since people with less powerful GPUs can max out all they need they won't upgrade thus developers won't develop more GPU hungry games." If everyone would follow that guy's logic we'd still all run on 3dfx.

Edited May 17, 2016 by Vivoune

[vicx]

From the RoadToVR review.

The upshot of all of this is that both the GTX 980 Ti and GTX 1080 are supremely overqualified in terms of raw grunt to handle the majority of VR experiences out there.

 

Absolute nonsense. This must be a review from someone who doesn't play games because that statement is not correct.

 

Sure many of the highly optimised VR launch titles (mostly single player) can run on a 980TI but sandbox engine games, multiplayer shooters and flight/space sims all have highly variable and demanding GPU loads and even the optimised titles can benefit greatly from 8xAA and 2x oversampling; much harder to sustain at 90fps.

 

I doubt a reviewer will test DCS but I hope someone has at least seen what a 1080 can do for Elite Dangerous.

[Swoodle]

As my current build has NO graphics card at all, the 1080 will fit just nicely. However either the end of the year or early next year I'll upgrade to the ti or an AMD card if they have big performance upgrades in the VR world.

[SharpeXB]

Absolute nonsense. This must be a review from someone who doesn't play games because that statement is not correct.

Well I'm sure he's referring to such graphically sophisticated and demanding VR games such as Lucky's Tale and Job Simulator... :noexpression:

[Chivas]

VR is in desperate need of faster highly VR optimized GPU's, especially if they want to attract AAA developers..

[Vivoune]

VR is in desperate need of faster highly VR optimized GPU's, especially if they want to attract AAA developers..

 

Absolutely, the sooner people can get their hands on affordable VR capable cards the sooner we'll see "real" VR games/sims and the sooner we'll see 8K HMDs.

[Gladman]

Wow,are you guys ever happy? ;)

[Vivoune]

Wow,are you guys ever happy? ;)

 

^^

 

The 1080 is "only" about 15-25% faster than the 980ti and similarly priced. So it's not of interest for a lot of people that's all.

 

It's a great card, but I think a lot of people here will be much more happy once the 1080ti gets released (or any AMD equivalent). If the price isn't too outrageous.

 

That said, it supposedly has good OC capabilities, so maybe with the 3rd party cooling and factory OC it can get as far as +50-60% over a 980ti. That might motivates a few people with the disposable income. That and the new features it brings, though the fact most of it is proprietary raises serious doubts about mass adoption from developers.

Edited May 18, 2016 by Vivoune

[vicx]

Glad, gotta disagree in spirit with Viv, I actually think the 1080 is a great step from Nvidia.

 

They move VR forward a bit more with the 1080. I think an ideal VR card for current software architectures is one that can support higher clocks and Nvidia have sort of kept shader performance per clock steady and focused on increasing the clocks with the 1080. I think if NV disregarded VR they could have easily boosted perf for traditional GPU loads lots more, so I'm glad that they focused on VR as much as they have. People who complain that the +1080 is not much of an upgrade from a 980TI just need a 1080 on water. It will slay.

[SharpeXB]

People who complain that the +1080 is not much of an upgrade from a 980TI.

The GTX 1080 out performs the TitanX by 20-30% so I'm not sure what "not much of an upgrade" means... :music_whistling:

[vicx]

The GTX 1080 out performs the TitanX by 20-30% so I'm not sure what "not much of an upgrade" means... :music_whistling:

 

Some people who OC their 980TI claim they can match the 1080. They might be right. I think this is where the noise is coming from. But if you OC 1080 then ...

[Vivoune]

The OC capabilities of the 1080 are still to be determined but it looks quite promising, sadly it's on a whole new level in terms of complexity apparently.

 

But yes I also think it's a great (long overdue) step forward, but even 20-30% is not worth spending that kind of money for a lot of 980ti/Titan owners. I really think we're almost there, in a few more months the 1080ti will hopefully show its nose and if the price is right should offer all the power one would need for years to come, and even further through SLI if need be.

Edited May 19, 2016 by Vivoune

[SharpeXB]

As the proud owner of two Titan X cards myself :ermm:

I'll probably hold out for the 1080Ti too.