GeForce GTX 1080 Ti Discussion Thread

heatlesssun said:
So I just started doing some overclocking:

View attachment 19549

Ran Heaven 4 and Time Spy, got about a 10% boost in those with these settings. But, wow, gets pretty loud, just about at the edge of being uncomfortably so and a lot of heat. But pretty insane performance. This making think about AIO cooling.
Click to expand...

AIO... Wut!? Come'on, this is [H]... you want those puppies in a full, custom loop! :D (Especially since they are SLI'ed!)

+196 on my GPU and +400 on my VRAM is my 24/7 setting (Pascal Titan X Pascal though).

Anyone with a 1080Ti have it waterblocked yet and holding a solid 2.1Ghz on the GPU yet? The OC numbers I'm seeing so far for most 1080Ti OCs tend to hover just below 2Ghz.
 
Last edited:
I'm topping out around +125 / +350

FaXmijcs.png


http://www.3dmark.com/3dm/18662978?
 
For the Nvidia branded 1080Ti FE, what app are you guys running to overclock? One better than the other? The last several cycles I have ran EVGA and always used theirs.
 
heatlesssun said:
Hmm, we'll I seem to off in this game as well. Didn't really do a lot of testing just spot checked games with canned benchmarks. Actually played a good bit of Forza Horizon 3. The 1080 Ti definitely is smoothing things out, almost able to keep a constant 60 FPS with brief dips in to the upper 50s. OCing should help out.

So far very impressed with these cards. And it actually is a pretty significant upgrade even over the 1080, at least at 4k.
Click to expand...

Syphon Filter said:
I averaged 56.something on the Ultra-preset but with AA off and Shadows set to Very High.
Click to expand...

I found one review where the numbers were identical to what I'm seeing. Using the ultra-preset without touching anything I'm getting the following results:

FPS:
Average: 59.47
Min: 42.96
Max: 68.18

This is what I was getting a couple days ago. I'm not sure what changed. According to the benchmark reports the settings are the same. I suspect this is because my boost clocks probably weren't very good due to me running the fans at stock speeds. I've increased their speed somewhat in MSI Afterburner. I didn't verify this so I'm just guessing.

FPS:
Average: 52.30
Min: 35.82
Max: 59.76

Here is what I got with Turf turned off and FXAA instead of Temporal AA.

FPS:
Average: 59.37
Min: 43.78
Max: 66.60

There is almost no change. The CPU usage reports as 21.3% consistently. GPU usage fluctuates between 78% and 92%. RAM usage is around 4.5GB.

Again, I'm seeing benchmarks online that are generally within the same performance range as what I'm seeing. The video in post #262 seems to indicate their test was in a different part of the game than where I'm at. I'm thinking this accounts for the difference. What's interesting to me is that they are using a non-OC'ed Core i7 7700K and DDR4 3200MHz RAM. That suggests in this particular case that Kaby Lake at higher RAM frequencies might be better than having more cores in this game despite it reaching into more than four. We've seen similar behavior out of other games so it isn't entirely unexpected.
 
I think I mentioned earlier in this thread that even w/ the 1080ti Forza Horizon 3 can't maintain 60fps. Well... I turned off MSAA and downsampled from double my native resolution (5120x2880). The game now runs at a damn-near rock solid 60fps and looks way better than it did with multisampling. It drops a frame or two here and there in the jungle but it's imperceptible from a gameplay perspective.



 
Last edited:
Personally I wonder how much CPU bottleneck I would have with this and an i7-4770K at 4.4 GHz... My target would be 1080p (preferrably with some DSR) @ 144 Hz (or at the very least above 100 Hz). My 980 Ti is still pretty good for 60 Hz.
 
I can hit 2000 stable at 65 % fan speed but that noise is killing my ears. This temp card is such a POS.

Need those Strix asap.
 
KickAssCop said:
I can hit 2000 stable at 65 % fan speed but that noise is killing my ears. This temp card is such a POS.

Need those Strix asap.
Click to expand...
This is why I've been water-cooling video cards in my custom loop for the last 5+ years. You end up paying $50 premium for the AIB custom cards for improved noise however a $100 block on a founders edition for quietness is the way ago, assuming you already paid for your custom loop which is a big investment.
 
I'll be interested to see what you can hit. I think I'm stuck on +125 / +400. Something is holding me there but I can't work out what. I'm seeing max temps around mid-70s. Maybe I just have shitty clockers.

I was hoping for +200 / +500.
 
Syphon Filter said:
I'll be interested to see what you can hit. I think I'm stuck on +125 / +400. Something is holding me there but I can't work out what. I'm seeing max temps around mid-70s. Maybe I just have shitty clockers.

I was hoping for +200 / +500.
Click to expand...

I seem to be topping out at +125 / +400 as well. That's still about a solid 10% bump over stock. The only problem is the NOISE. The loudest cards by far since the GTX 480 I've had when overclocked at least.
 
According to Afterburner OSD monitoring, I'm hitting power and voltage limits. I guess that's that my lot.
 
Syphon Filter said:
According to Afterburner OSD monitoring, I'm hitting power and voltage limits. I guess that's that my lot.
Click to expand...

This is still a solid 10% boost from my observation and that's pretty good. It's the damned noise from the overclock that I have a problem with. What's your take on that?
 
heatlesssun said:
I seem to be topping out at +125 / +400 as well. That's still about a solid 10% bump over stock. The only problem is the NOISE. The loudest cards by far since the GTX 480 I've had when overclocked at least.
Click to expand...

At what speeds does the noise become an issue for you? Edit.. not very clear but referring to fan speed.
 
Not quite as extreme, I've limited mine to 50% fan at +120/300 overclock.
This gives approx 1911MHz once heated up fully.
I'm getting an Accelero Xtreme IV cooler some time soon, gotta make sure it has no issues first.
 
Max.
I could probably manage +125 and a bit less than +350 memory but it failed at 130/350 so didnt test further.
Often the GPU clock is over 1920 but under high stress after heating up it is most often at 1911MHz.

I'm not a fan of over 80C, it sometimes reaches 86C.
But I'm testing it before losing my warranty and would rather it failed now from any basic problems.
 
I managed +145 on the GPU and +450 on the ram. I've only played FO4 with it (just a short term heaven test) and the card runs nice and quiet. Never seen the fan spin up over 50%, and I'm seeing boost speeds right around 2k. Clock speeds are pretty erratic, they jump between 2k and ~1400 when gaming, but FO4 seems to be really inconsistent as far as GPU load. Solid 60fps @3440x1440, it's not really taxing the card too much.
 
B2BigAl said:
I managed +145 on the GPU and +450 on the ram. I've only played FO4 with it (just a short term heaven test) and the card runs nice and quiet. Never seen the fan spin up over 50%, and I'm seeing boost speeds right around 2k. Clock speeds are pretty erratic, they jump between 2k and ~1400 when gaming, but FO4 seems to be really inconsistent as far as GPU load. Solid 60fps @3440x1440, it's not really taxing the card too much.
Click to expand...
Other games that push the gpu might not be stable without at least ramping the fan up.
Try with vsync off to get some idea.
 
If you want to try out some games that put 100% load on the card, some good examples of that are:

Witcher 3
Call of Doody Infinite Warfare
Mad Max
Ass Creed Syndicate

Try those with 50% fan speed and 2000 MHz. Your card will blow up or run at 90C.
 
  • Like
Reactions: elvn
like this
1080 Ti SLI currently under water. So far, I've managed +160 / +500 (shunt mod on both cards) with clocks as high as 2063MHz...still testing.

33363088132_46f71a1f63_c.jpg
 
Shunt mod? Is that software or hardware?

Also, those running stock (i.e. blower) FEs, are you using custom fan profiles or just the automatic?
 
Syphon Filter said:
Shunt mod? Is that software or hardware?

Also, those running stock (i.e. blower) FEs, are you using custom fan profiles or just the automatic?
Click to expand...

I think that's a hardware mod. Basically applying liquid metal TIM over resistors to change their reading, allowing for more power.
 
nwkrep82 said:
1080 Ti SLI currently under water. So far, I've managed +160 / +500 (shunt mod on both cards) with clocks as high as 2063MHz...still testing.

33363088132_46f71a1f63_c.jpg
Click to expand...
Looks like you never touched the voltage slider and your already at 2063. Is that correct? Your voltage is already higher with your mod. Now what if you raise it further?
 
KickAssCop said:
If you want to try out some games that put 100% load on the card, some good examples of that are:

Witcher 3
Call of Doody Infinite Warfare
Mad Max
Ass Creed Syndicate

Try those with 50% fan speed and 2000 MHz. Your card will blow up or run at 90C.
Click to expand...

I agree and am personally more interested in the frame rate differences in stable real world stressful game sessions stock boost vs stable overclock at 2560x1440 and 4k.

The stressed by game temps stock boost vs AiO default boost vs manual overclock without volt mod would be interesting to me too.
 
Last edited:
Syphon Filter said:
Shunt mod? Is that software or hardware?

Also, those running stock (i.e. blower) FEs, are you using custom fan profiles or just the automatic?
Click to expand...

These are FE cards with EK blocks and backplates.

trandoanhung1991 said:
I think that's a hardware mod. Basically applying liquid metal TIM over resistors to change their reading, allowing for more power.
Click to expand...

Yes, that is correct. This hardware mod tricks the card into calculating a lower power limit.

Syphon Filter said:
What are load temps under water???
Click to expand...

So far I haven't seen temps above the mid 40s overclocked. I still have more tests and games to run.

Nirad9er said:
Looks like you never touched the voltage slider and your already at 2063. Is that correct? Your voltage is already higher with your mod. Now what if you raise it further?
Click to expand...

Yes 2063MHz with no voltage adjustments is correct. This mod does not increase/unlock voltage or voltage limits. It effectively raises the power limit. The major benefit to this is (and better cooling), is it allows more consistent and stable clock speeds. Raising the voltage may/may not help as there are still limits with the silicon itself and the max voltage allowed. Also, chips will hit that silicon lottery wall no matter what you do.
 
Last edited:
So, playing Titanfall 2 single-player (highly underrated game IMO) and the system froze. No crash, nothing. Just locked up. The 1080Ti and RAM are the only recent changes. Ran 3DMark, Time Spy, ect... no issues. It stayed like that for a minute or two until I hard rebooted the system. The only OC was on the CPU @ 4.6ghz and now I have dropped it back to 4.4ghz. Temps looked ok but I will say this 6700k runs hotter than my old 4770k by about 10-20 degrees. CPU and GPU on same loop. Time to separate them? When I rebooted I got the following error?

IMG_20170319_125109.jpg
 
Etherton said:
So, playing Titanfall 2 single-player (highly underrated game IMO) and the system froze. No crash, nothing. Just locked up. The 1080Ti and RAM are the only recent changes. Ran 3DMark, Time Spy, ect... no issues. It stayed like that for a minute or two until I hard rebooted the system. The only OC was on the CPU @ 4.6ghz and now I have dropped it back to 4.4ghz. Temps looked ok but I will say this 6700k runs hotter than my old 4770k by about 10-20 degrees. CPU and GPU on same loop. Time to separate them? When I rebooted I got the following error?

View attachment 19706
Click to expand...

Good luck with that.. hard to read the writting in the pic.

Take it back down to basics.. keep cpu at default and mem if overclocking.. do the same for the 1080ti.

Start over with cpu overlcock.. find the stable overclock first.. thats with testing it on different programs on games.. then move on to overclock your card testing that in different tests as well.. just cause one is fine dont mean that one program put a good strain on your hardware.

Walk before you run with overclocking.

And yes test the ram as well to make sure its good.. test it with the cpu stock.
 
eclypse said:
Good luck with that.. hard to read the writting in the pic.

Take it back down to basics.. keep cpu at default and mem if overclocking.. do the same for the 1080ti.

Start over with cpu overlcock.. find the stable overclock first.. thats with testing it on different programs on games.. then move on to overclock your card testing that in different tests as well.. just cause one is fine dont mean that one program put a good strain on your hardware.

Walk before you run with overclocking.

And yes test the ram as well to make sure its good.. test it with the cpu stock.
Click to expand...

Also no need to seperate the loops.. much cheaper to just add more rads of possable.
 
You must log in or register to reply here.
Back
Top