The GTX 1080-Ti Thread

Nice! Mine shipped out yesterday and will be getting it tomorrow.

 

No really i still have smooth over 60 fps with witcher 3 with my sli 1080,s and hairworks.

Nevermind my bad.

 

This guys speaks the truth.

 

So far for the Vive, the 1080ti is completely worth it. Everything is far smoother than with my 980ti (and his lazy brother who stopped doing anything after a year of releases).

 

So I got my new GTX 1080 Ti SC2 the same day NVIDIA released a new driver, v381.89!

I wasn't going to benchmark since I find it generally dull but since I'd installed a new driver I thought I'd run a few random game benchmark with the new driver prior to installing the new card. As such I didn't actually install the GTX 1080 Ti until nearly 9pm but I ended up not going to bed until 2am anyway so I had a good opportunity to test the new card.

These games were ran at 2560x1440 with completely maxed out settings except for SSAA and with NO overclock on the GTX 1080 Ti but a mild overclock of +75 MHz on the core/+250 MHz on the memory on the GTX 1080 FTW (this is already factory overclocked remember). This resulted in a boost to around 2,038 MHz on the GTX 1080 and an impressive boost to 1,911 MHz on the GTX 1080 Ti where is remained until the temperature rose above 71 C. All I did on the GTX 1080 Ti was increase the Power % from 100 to its max. of 120.

The results are as follows:

Sniper Elite 4 - DX12, Magazeno Facility, noted FPS immediately after the save game loads
1080: 70.0 fps
1080 Ti: 101 fps

Resident Evil 5 - DX9, in-game Variable benchmark
1080: 239.7 fps
1080 Ti: 240.6 fps

An oldie, yes, and one that appears to be CPU bound.

Middle-earth: Shadow of Mordor - DX11
1080: Ave. 102.88/Max. 140.96/Min. 65.50 fps
1080 Ti: Ave. 131.85/Max. 187.53/Min. 81.31 fps

Tom Clancy's The Division - DX12, in-game benchmark
1080: [N/A as the game was updating]
1080 Ti: 96.1 fps

Rise of the Tomb Raider - DX12, in-game benchmark
1080: 93.24 fps
1080 Ti: 114.5 fps

Far Cry Primal - DX11, in-game benchmark
1080: Min. 58/Ave. 78/Max. 93 fps
1080 Ti: Min. 57/Ave. 89/Max. 108 fps

For Honor - DX11, in-game benchmark
1080: Min. 68.69/Ave. 92.14/Max. 134.45 fps
1080 Ti: Min. 80.34/Ave. 116.54/Max. 190.48 fps

Tom Clancy's Ghost Recon: Wildlands - DX11, in-game benchmark
1080: 52.14 fps
1080 Ti: 59.77 fps

Hitman (2016) - DX12, in-game benchmark
1080: 103.61 fps
1080 Ti: 104.7 fps (???)

As you can see there's a 20 fps boost in framerates to the better optimised games such as Middle-earth, Rise, For Honor and Sniper Elite 4. Resident Evil 5 appears to CPU bound and Far Cry Primal basically maxed one CPU core so I'm not surprised that the framerate only improves by 11 fps. As for Wildlands, well I don't think it's the most optimised game by far but nevertheless the game feels smoother now it runs at 60 fps more of the time.

 

The Ti does so well under water although whatever is set on the core, everything is eventually limited by the 120% power draw.
My FE idles at case ambient temps or just above and even under full load rarely ever goes above +10C of ambient.

I have gone from a significantly over clocked/volted watercooled 780gtx (@1440p) so have noticed some major gains in performance, most notably with the 3gb VRam getting maxed out on the 780.

 

@ pegasus1 - Liquid cooling makes me nervous so it'll be a few years before I ever use that in my PC. Air cooling is fine IMO. It's not like I heavily overclock everything anyway and fan noise isn't an issue either.

Looking forward to putting my new card through its paces tonight. What I like about this card is that it has three LEDs on the side that indicate how hot various areas of the card are. Currently mine are blue which means temps are cool as they get warmer they go to green (warm) and, hopefully, something I will never see: red which equals hot. It's a neat way to check card temps without the need for any monitoring software, although you can use EVGA Precision X OC, I believe, to include the readings in the on-screen display if you want.

 

You would get much better results, far as the gap between two cards, if you go up the resolution to 4k, with all bells and whistles intact + nvidia inspector hardware forced aa.

But Thanks for sharing though.

 

Has anyone tried the Gigabyte 1080ti Auros??? Not the extreme version, if so, any good compared to others out there?

 

Yes and it seems to be a great card. I have it oc'd to the Extreme specs but want more info on how far I can go without messing with voltage. Using the Aorus Graphic Engine for now.

 

Hi. My new GTX 1080 Ti from MSI is getting only 8000pts in Time Spy. I have chcecked may reviews and this card should have at least 10000 pts . Same problem with other benchmarks like Fire Strike where I have 19500pts but in reviwes is around 22000 pts . I have 700W PSU from cooler master, i5 6600k ovelclocked to 4.6Ghz When im using MSI Afterburner and monitoring status of this card ,it show me that this card is loaded during benchmarks only 80%. Im getting some info that voltage is limited or sometimes power. Is the a problem with my PSU or maybe my i5 6600 is too slow for this card ?? Any help ?

 

With TimeSpy its just the Graphics Score you need to go by as a yardstick as we all have different CPUs etc but yes if 8000 is your graphical score then its low.
My card is an MSI Fe and its scores are roughly
Out the box = 9450
Stock Core, Stock Mem, Power Limit 120, Temp Limit 90 = 9650
Core +175, Mem 400, Power Limit 120, Temp Limit 90 =10660

 

I got 8,144 in Time Spy with my GTX 1080 Ti, which I admit was disappointing but since when have benchmarks reflected real-time game performance? In my experience, they haven't so while that score might be underwhelming compared with Guru3D's review testing it is a leap over the 6,698 I score with my overclocked GTX 1080.

*EDIT*
Just did a quick search and someone else scored 9,096 with a 4.6 GHz i7-4770K and a GTX 1080 Ti albeit they had 32 GB of 1,333 MHz memory (I have a 4.0 GHz i7-4770K with 16 GB of 1,6000 MHz memory). Looks like the 900 point difference may be due to the higher CPU overclock?

 

An X% jump in bench performance will be reflected in a game.
Your card out of the box is running about 15% slower than mine also out of the box, that will be reflected in both a bench and real world games (and benches within games).
A benchmark is exactly that, its a standard that can be accurately used as a unit of measure.
The overall score isnt important as CPU, Mem and other factors come into play but using a benchmark to identify improvements against base scores is very useful, certainly with the 1080Ti, how the Boost system works and thus finding the right balance.

As above, i was astounded just how much quicker the card became as it was watercooled therefore temperature throttling didnt happen, also the jump when i maxed out both the power limit, and again when OC it.

Maybe the score of 8000 is typical for an FE crad out of the box with stock setting for temp and power draw and a stock fan profile. maybe if you cranked up the the temp and power limits plus went 100% fan (just as a test) you will see a gain of maybe 1000 points (8% boost), that may show that the card is limiting itself due to thermal profiles.

 

Not if the 900 points is the graphics score rather than the overall score. The graphics score is independent of the CPU/Mem.

 

This, my 1080 Ti at stock sits at ~10k (a FE may be lower out of the box) and with max OC ~11k graphics score in Time Spy. If it's significantly lower then that then there might be a problem, which for sure will also reflect in games too like pegasus1 said.

 

Great score, is your 10k with the temp and power limit raised?

 

I'm sad... my waterblock won't be around until next week. No buidling this weekend

EKWB listed it as shipping on the 21st during order, then NOTHING happened for a week until I asked them on any input since it was NOT shipped last week. Two lines coming with the info that they will probably ship it tomorrow... not very convinced by EKWB's practice when ordering from them directly for the first time.

 

I ordered mine as soon as i found out the Titan blocks would fit, way before the Ti was even released.

 

It's 100% how it came out of the box, didn't even use my custom fan profile. Note however that it boosts at 1940~ average at stock.

This is the score I got the very first time I popped the card in my rig:

http://www.3dmark.com/spy/1593982