GTX 1080 Ti Performance and overclocking
How we test
We’ve run all three of these cards on the open test bench that consists of the following components.
- Motherboard: Asus Z170-Deluxe
- Processor: Intel Core i5-6600K (Overclocked to 4.8GHz)
- RAM: Corsair Vengeance 2666MHz, 16GB DDR4
- Cooler: Corsair H60 liquid cooler
- PSU: Corsair CX750M
- SSD: Samsung 850 EVO
- OS: Windows 10 Pro 64-bit
I’ve run the numbers against the GTX 1080 and the Titan X, all with the very latest March 2017 drivers to give each GPU the fairest crack of the whip.
The games were run at their maximum settings with a healthy dose of anti-aliasing. Some might find the addition of AA controversial, but you’ll always get the best comparisons when pushing cards right to their very limits. I also ran each game at 5K, although here I’ll focus on the 4K results since this is what most people will be expecting.
GTX 1080 Ti – Performance and Benchmarks
(average fps)
To make the most of the 1080 Ti, we’ve added some benchmarks that represent a couple of more recent games to ensure it’s pushed to its limits. We start with a hit from last year.
Battlefield 1
DICE’s World War I epic is a great benchmark for high-end GPUs thanks to incredible detail, particle effects and GPU-sapping textures. I ran it at 4K and at 5K, although realistically most people will be maxing out at 4K for the time being.
Battlefield 1 doesn’t have a repeatable built-in benchmark, so the test is run during the on-rails section of the mission “Mud and Blood”.
The GTX 1080 Ti managed an average frame rate of 72fps, which is superb when you consider I’d turned everything to Ultra and whacked up anti-aliasing as well. Tweaking the graphics settings that aren’t as important to you will yield even better results.
Related: Best gaming PC specs to build yourself
Compared to the GTX 1080, which managed 55fps in the same test, the 1080 Ti is 24% faster.
Middle-earth: Shadow of Mordor
It’s almost time to retire Shadow of Mordor from our benchmarking suite. 2014’s adventure epic has an excellent built-in benchmark, but with sequel Shadow of War coming soon, the 1080 Ti benchmark is perhaps its last hurrah.
At 4K and Ultra settings, 86.6fps is another good result, with no noticeable dips in frame rate during the test. Surprisingly for such an old title, Shadow of Mordor was a huge outlier in my tests, generating a 1080 Ti score that was 44% faster than the regular GTX 1080.
Rise of the Tomb-Raider
One of the toughest tests at high resolutions, Tomb Raider’s beautiful benchmark was where the 1080 Ti came closest to nudging the sub-60fps barrier, with an average score of 61.5fps at maximum settings in 4K. This was 20% faster than the GTX 1080, which is a notable and noticeable improvement.
Drop a few graphics settings, including shadows and anti-aliasing, and you’ll have 70fps plus in Tomb Raider without any issues.
Hitman
Hitman released its final set of missions last year, but it still has a resource-hogging benchmark that really tests the mettle of GPUs.
In 4K, the GTX 1080 Ti managed an impressive 73.6fps at maximum settings, maintaining a relatively stable frame rate throughout the run. It wasn’t significantly faster than the GTX 1080, though, which put in an average frame rate of 59fps.
GTX 1080 Ti vs 1080
(average fps)
GTX 1080 Ti vs 980 Ti
I didn’t have a 980 Ti test with the latest drivers, but our benchmark results from last year point to 1080 Ti performance that’s between 30% and 50% faster at 4K in Hitman, Tomb Raider and Middle-earth: Shadow of Mordor.
GTX 1080 Ti vs Titan X
(average fps)
It was the comparison Nvidia made when it launched the 1080 Ti – and it is indeed faster for £300 less. If you own a Titan X, it isn’t exactly a tragedy; you’ve had world-beating performance for a full seven months, so think of it as a Titan Xclusive. Obviously, you’ve no need to upgrade.
GTX 1080 Ti – Overclocking, Heat and Power Consumption
The GTX 1080 Ti is a much more powerful beast than the GTX 1080, and considerably more so than the 980 Ti. Its TDP (thermal design power) is rated at 250W, the same as the 980 Ti and Titan X, while the GTX 1080 clocks in at just 180W. This could be all the difference for those with builds that have 500W power supplies.
Indeed, peak power draw for the entire TrustedReviews test system without overclocks was 412W, compared to the GTX 1080’s 340W in the same test. If you have a more powerful processor than the one we used, take a look at your power consumption readings to check you won’t over-stretch your PC’s power supply.
Running the Unigine Heaven benchmark on a loop in 4K, the GTX 1080 Ti was able to boost well beyond its stated 1,600MHz clock speed, happily sitting at around 1,800MHz at around 84oC.
Overclocking is as easy as ever, with the GTX 1080 Ti receptive to around 1,950MHz. I was quickly able to get a stable looping Unigine benchmark at 1,950MHz without the GPU throttling back. Temperatures maxed out at around 86oC on the automatic fan profile; I’d probably tweak it so it runs a little louder and cooler. Popping the fan speeds up to 70% reduced the temperature to a more comfortable 77oC.
It will be interesting to see whether third-party card manufacturers can encourage the 1080 Ti beyond 2GHz and what sort of cooling will be required in order to achieve such performance.
Memory overclocking was nice and easy: the 11GB of 11GHz GDDR5X memory was happy to receive a boost up to 11.8GHz without instability kicking in.
Nvidia says it has improved the fan design to run slightly quieter, but I didn’t hear a difference – and neither could my decibel meter. At 28,000rpm, both fans put out between 55 and 56dBA of noise, which is perfectly acceptable. In a compact, poorly insulated case with overclocks applied, it will likely become louder, but not significantly.
Should I buy the GTX 1080 Ti?
Buying the most expensive graphics card you can afford has long been a reviewer’s’ trope, but it’s almost always true. The 1080 Ti stretches this advice to its very limits with its super-high price, but with no competition from AMD at this end of the market, Nvidia has reached new performance territory.
When buying a 1080 Ti, don’t think about the Titan X – for what it offers, it’s rather overpriced at this point – but think more about the GTX 1080. With the 1080’s big price drop to £500, it’s become the new de facto choice for 4K gaming at High settings.
However, with 3GB more memory and at least 20% better performance, it’s hard not to recommend the 1080 Ti for those building the best system they possibly can for 4K and multi-screen gaming.
If you’re thinking about a direct upgrade from the 980 Ti, the performance increase is substantial; the 1080 Ti is up to twice as fast. If you have a 1080, your card has a lot of life left in it and it’s probably wise to stick with it for now.
Related: Best graphics card
If you don’t have a 4K monitor and don’t plan on buying one, the 1080 Ti is overkill – and even if you have a 1080 Ti but don’t always find yourself playing the very latest AAA games, you can save a couple of hundred quid on a 1080 or more on a 1070.
Verdict
A new performance benchmark – but you pay for it.
GTX 1080 Ti Performance and overclocking
How we test
We’ve run all three of these cards on the open test bench that consists of the following components.
- Motherboard: Asus Z170-Deluxe
- Processor: Intel Core i5-6600K (Overclocked to 4.8GHz)
- RAM: Corsair Vengeance 2666MHz, 16GB DDR4
- Cooler: Corsair H60 liquid cooler
- PSU: Corsair CX750M
- SSD: Samsung 850 EVO
- OS: Windows 10 Pro 64-bit
I’ve run the numbers against the GTX 1080 and the Titan X, all with the very latest March 2017 drivers to give each GPU the fairest crack of the whip.
The games were run at their maximum settings with a healthy dose of anti-aliasing. Some might find the addition of AA controversial, but you’ll always get the best comparisons when pushing cards right to their very limits. I also ran each game at 5K, although here I’ll focus on the 4K results since this is what most people will be expecting.
GTX 1080 Ti – Performance and Benchmarks
(average fps)
To make the most of the 1080 Ti, we’ve added some benchmarks that represent a couple of more recent games to ensure it’s pushed to its limits. We start with a hit from last year.
Battlefield 1
DICE’s World War I epic is a great benchmark for high-end GPUs thanks to incredible detail, particle effects and GPU-sapping textures. I ran it at 4K and at 5K, although realistically most people will be maxing out at 4K for the time being.
Battlefield 1 doesn’t have a repeatable built-in benchmark, so the test is run during the on-rails section of the mission “Mud and Blood”.
The GTX 1080 Ti managed an average frame rate of 72fps, which is superb when you consider I’d turned everything to Ultra and whacked up anti-aliasing as well. Tweaking the graphics settings that aren’t as important to you will yield even better results.
Related: Best gaming PC specs to build yourself
Compared to the GTX 1080, which managed 55fps in the same test, the 1080 Ti is 24% faster.
Middle-earth: Shadow of Mordor
It’s almost time to retire Shadow of Mordor from our benchmarking suite. 2014’s adventure epic has an excellent built-in benchmark, but with sequel Shadow of War coming soon, the 1080 Ti benchmark is perhaps its last hurrah.
At 4K and Ultra settings, 86.6fps is another good result, with no noticeable dips in frame rate during the test. Surprisingly for such an old title, Shadow of Mordor was a huge outlier in my tests, generating a 1080 Ti score that was 44% faster than the regular GTX 1080.
Rise of the Tomb-Raider
One of the toughest tests at high resolutions, Tomb Raider’s beautiful benchmark was where the 1080 Ti came closest to nudging the sub-60fps barrier, with an average score of 61.5fps at maximum settings in 4K. This was 20% faster than the GTX 1080, which is a notable and noticeable improvement.
Drop a few graphics settings, including shadows and anti-aliasing, and you’ll have 70fps plus in Tomb Raider without any issues.
Hitman
Hitman released its final set of missions last year, but it still has a resource-hogging benchmark that really tests the mettle of GPUs.
In 4K, the GTX 1080 Ti managed an impressive 73.6fps at maximum settings, maintaining a relatively stable frame rate throughout the run. It wasn’t significantly faster than the GTX 1080, though, which put in an average frame rate of 59fps.
GTX 1080 Ti vs 1080
(average fps)
GTX 1080 Ti vs 980 Ti
I didn’t have a 980 Ti test with the latest drivers, but our benchmark results from last year point to 1080 Ti performance that’s between 30% and 50% faster at 4K in Hitman, Tomb Raider and Middle-earth: Shadow of Mordor.
GTX 1080 Ti vs Titan X
(average fps)
It was the comparison Nvidia made when it launched the 1080 Ti – and it is indeed faster for £300 less. If you own a Titan X, it isn’t exactly a tragedy; you’ve had world-beating performance for a full seven months, so think of it as a Titan Xclusive. Obviously, you’ve no need to upgrade.
GTX 1080 Ti – Overclocking, Heat and Power Consumption
The GTX 1080 Ti is a much more powerful beast than the GTX 1080, and considerably more so than the 980 Ti. Its TDP (thermal design power) is rated at 250W, the same as the 980 Ti and Titan X, while the GTX 1080 clocks in at just 180W. This could be all the difference for those with builds that have 500W power supplies.
Indeed, peak power draw for the entire TrustedReviews test system without overclocks was 412W, compared to the GTX 1080’s 340W in the same test. If you have a more powerful processor than the one we used, take a look at your power consumption readings to check you won’t over-stretch your PC’s power supply.
Running the Unigine Heaven benchmark on a loop in 4K, the GTX 1080 Ti was able to boost well beyond its stated 1,600MHz clock speed, happily sitting at around 1,800MHz at around 84oC.
Overclocking is as easy as ever, with the GTX 1080 Ti receptive to around 1,950MHz. I was quickly able to get a stable looping Unigine benchmark at 1,950MHz without the GPU throttling back. Temperatures maxed out at around 86oC on the automatic fan profile; I’d probably tweak it so it runs a little louder and cooler. Popping the fan speeds up to 70% reduced the temperature to a more comfortable 77oC.
It will be interesting to see whether third-party card manufacturers can encourage the 1080 Ti beyond 2GHz and what sort of cooling will be required in order to achieve such performance.
Memory overclocking was nice and easy: the 11GB of 11GHz GDDR5X memory was happy to receive a boost up to 11.8GHz without instability kicking in.
Nvidia says it has improved the fan design to run slightly quieter, but I didn’t hear a difference – and neither could my decibel meter. At 28,000rpm, both fans put out between 55 and 56dBA of noise, which is perfectly acceptable. In a compact, poorly insulated case with overclocks applied, it will likely become louder, but not significantly.
Should I buy the GTX 1080 Ti?
Buying the most expensive graphics card you can afford has long been a reviewer’s’ trope, but it’s almost always true. The 1080 Ti stretches this advice to its very limits with its super-high price, but with no competition from AMD at this end of the market, Nvidia has reached new performance territory.
When buying a 1080 Ti, don’t think about the Titan X – for what it offers, it’s rather overpriced at this point – but think more about the GTX 1080. With the 1080’s big price drop to £500, it’s become the new de facto choice for 4K gaming at High settings.
However, with 3GB more memory and at least 20% better performance, it’s hard not to recommend the 1080 Ti for those building the best system they possibly can for 4K and multi-screen gaming.
If you’re thinking about a direct upgrade from the 980 Ti, the performance increase is substantial; the 1080 Ti is up to twice as fast. If you have a 1080, your card has a lot of life left in it and it’s probably wise to stick with it for now.
Related: Best graphics card
If you don’t have a 4K monitor and don’t plan on buying one, the 1080 Ti is overkill – and even if you have a 1080 Ti but don’t always find yourself playing the very latest AAA games, you can save a couple of hundred quid on a 1080 or more on a 1070.
Verdict
A new performance benchmark – but you pay for it.
GTX 1080 and 1080ti — Zcash and Ethereum GTX 1080ti mining test results and optimal overclocking on Zcash | Page 60
Nikita48
Dancing with a tambourine
-
-
#1.181
storm067 said:
Yes, it doesn’t matter, everything worked for me with last year’s standards at 1080ti. Updated because 3000 series cards appeared on
Click to expand…
I understand, thanks for the advice!
Nikita48
Dancing with a tambourine
-
-
#1.182
storm067 said:
Yes, it doesn’t matter, everything worked for me with last year’s standards at 1080ti. Updated because 3000 series cards appeared on
Click to expand…
Tell me how much the ball should give the card? I noticed that with different overclocking, the shares \ min parameter is either more or less, does it somehow affect the profitability or look only at the hashrate?
Last edit:
storm067
Forum friend
-
-
#1. 183
Nikita48 said:
Tell me how much the ball should give the card? I noticed that with different overclocking, the shares \ min parameter is either more or less, does it somehow affect the profitability or look only at the hashrate?
Click to expand…
See how many hashes are fixed by the pool, if it converges with the miner approximately, then everything is fine. Watch at least 6 hours after the start of the miner, more is better, otherwise the statistics will not be informative. And I don’t advise mining on ezer with one 1080ti, although of course it’s up to you. nine0007
6 x gtx1650
Nikita48
Dancing with a tambourine
-
#1. 184
storm067 said:
See how many hashes are fixed by the pool, if it converges with the miner approximately, then everything is fine. Watch at least 6 hours after the start of the miner, more is better, otherwise the statistics will not be informative. And with one 1080ti, I don’t advise mining on ezer, although it’s certainly a matter of where is it better to mine? nine0007
Click to expand…
where is it better to mine in my case?
storm067
Forum friend
-
-
#1. 185
Nikita48 said:
and where is it better to mine in my case? nine0007
Click to expand…
binance (steals), nicehash best option IMHO
6 x gtx1650
BoostOrDie
Own person
-
-
#1.186
nine0002 Nik65 said:
At one time 17/18 — 1080 were the most left cards for mining .. and you want to reanimate them at the end of 21 .. the owners of this antiques were lucky with such buyers . . when you bought them, at least you understood that you were taking 1080! Not 1080ti?
Click to expand…
And what does ti have to do with it? I buy exactly 1080 (not ti), I know that ti gives 40-45 or something like that. And 1080 gives 33-35. But the price tag of 60-65k for 40mx is not profitable. 1080 Good cards, cold and with a margin of memory. + only 1 memory chip manufacturer, which removes the lottery factor when buying. Episode 10 is unkillable. nine0193 The most profitable cards are 1060 6gb, they give 24.5mx, they cost about 25-27tyr, but there is a lottery with memory. I had 12 1060. I bought them for 18-20 tyr, which was mega profitable at the beginning of 2021. Now I sell them for 28 and upgrade to 1080 to increase power. Fewer slots take up more power, so I will add power to the same farms with the same number of slots.
+ 1080 is colder than 1060 oddly enough. except for references. 1080 gamerocks aorus and so on heat up for 55s when like 1060s 58-63. nine0193 my current asus turbo heats up for 71s.
Buying a 30 series, which digs the same 30-35 mx, for the price of two 1080 is just stupid.
BoostOrDie
Own person
-
-
#1.187
I remembered the question I wanted to ask. What does setting P State to «Off» in Nvidia Inspector do? they say like the frequency so that it doesn’t jump .. but it doesn’t jump for me anyway …
Cybo64bit
Own person
-
-
#1. 188
Nikita48 said:
Hello everyone! There is a GTX 1080 Ti Gaming OC 3x card, no matter how I tried, I could only achieve 39mh on the air, the OhGodAnETHlargementPill tablet works strangely, it adds only 3-4 mh in the drain without a 36mh tablet, overclocking does not help, set different voltage and frequency limits looked in the gpu -z controller memory is loaded no more than 80%, I don’t know what is the reason, I disabled the parameter through NVidiaProfileInspector, set 4g, I tried all modes from 1-6 with the mt parameter, a slight increase in performance between them, I tried NiceHash QuickMiner on a high 44mx preset, which means the card can be configured, but how ? on NBMiner 39mx as in all other miners now on t-rex with the mt4 parameter. Maybe someone has some ideas I don’t know where to dig anymore? maybe there are owners of these cards, please help,
Click to expand. ..
run the afterburner with reset settings and then run the quickminer and see what settings it has set in the afterburner
pol70
Newbie
nine0008
-
-
#1.189
Hello forum users! Just registered, and, accordingly, just started mining. I ask for help, in the form of advice. I have 1080 and 1070. I caught up with 1070 to 25, and 1080 to 33 (on average). You can tell, only someone knows exactly what you are talking about. otherwise there are so many opinions that my head is spinning, the following: Approximately the average hashrate on the air in Chivios, and the average settings for the core and memory to get them at 1070 and 1080. I have up to 100 for the core, up to 1300 for memory.
ATP in advance!
Nikita48
Dancing with a tambourine
-
-
#1.190
Cybo64bit said:
run the afterburner with reset settings and then run the quickminer and see what settings it has set in the afterburner
Click to expand…
empirically found the best option for my card, it seems to me. the settings are as follows, at 962mv 1949 MHz for the core, -502 for memory, 80% pl? temperature 54-55 degrees coolers by 87% at 185w consumption 41. 7 mhsh on nicehash tell me if this is normal?
BoostOrDie
Own person
-
-
#1.191
The power went out today. Everything is in order with the farm, and the card that is in the Personal PC and mines all the time, has become artificial.
GTX 1080 Aorus.
System OK
Games OK.
In Furmark — OK
in Benchmarks — OK
I start the miner — already at the stage of forming a dag file — small stripes and colored squares appear for 5 seconds then a black screen. nine0193 If you remove the memory tweak from the miner, it works fine, but according to the hashrate, it «turns» into 1060. ..
What happened to it?
Cybo64bit
Own person
-
-
#1.192
BoostOrDie said:
The light went out today. Everything is in order with the farm, and the card that is in the Personal PC and mines all the time, has become artificial.
GTX 1080 Aorus.System OK
Games OK.
In Furmark — OK
in Benchmarks — OK
I start the miner — already at the stage of forming a dag file — small stripes and colored squares appear for 5 seconds then a black screen.
If you remove the memory tweak from the miner, it works fine, but according to the hashrate, it «turns» into 1060…
What happened to herClick to expand…
stripes and squares are bad symptoms. but just in case, try to outweigh the driver by removing it with ddu
Nikita48
Dancing with a tambourine
nine0009
#1.193
Hello everyone! please tell me, the problem is this when downvolting in the msy autoburner, the strange behavior of the card, when I set PL 80% in monitoring, it shows that the card operates at very low frequencies both in voltage and in the core, I attach the screen. 290 watts this happened after I started mining RVN now it shows normally for the first minutes then it starts to fail, while the temperatures and mining are normal stable tried to reinstall Windows \ drivers \ different miners tried \ atoberner reinstalled even flashed the bios, did not help. do you have any idea what it is connected with? the card is dying or what could it be? nine0007
Experienced
-
-
#1.194
Nikita48 said:
Hello everyone! please tell me, the problem is this when downvolting in the msy autoburner, the strange behavior of the card, when I set PL 80% in monitoring, it shows that the card operates at very low frequencies both in terms of voltage and core, I attach the screen. 290 wat this happened after I started mining RVN now it shows normally for the first minutes then it starts to fail, while the temperature and mining are normal stable tried to reinstall Windows \ drivers \ different miners tried \ atoberner reinstalled even the bios was flashed, did not help. do you have any idea what it is connected with? the card is dying or what could it be?
Click to expand…
Nothing is clear.
nine0002
Debian
Experienced
-
-
#1.195
Nikita48
Dancing with a tambourine
-
nine0012
-
#1. 196
vovyn said:
Nothing is clear.
Click to expand…
when the pl drops to 80, the card mines at very low frequencies, it reaches its frequency only at pl 120
Debian
Experienced
-
-
#1.197
Nikita48 said:
when the pl drops to 80, the card mines at very low frequencies, it reaches its frequency only at pl 120
Click to expand. ..
The smaller the PL, the lower the core frequency. Everything is correct. That’s the way it should be. What’s embarrassing?
nine0002
Nikita48
Dancing with a tambourine
-
-
#1.198
MegaVlad said:
The smaller the PL, the lower the core frequency. Everything is correct. That’s the way it should be. What’s embarrassing?
Click to expand…
what is on the air now the same monitoring works shows that the core is working at 1200 MHz and it used to work normally at pl80 1900MHz
Debian
Experienced
-
-
#1. 199
Nikita48 said:
what is on the air, now the monitoring also works, it shows that the core works at 1200 MHz, but before it worked at pl80 normally 1900MHz
Click to expand…
Reinstall the drivers, test the card on another PC, test it on Hive Rave. Most likely only then there will be a 100% diagnosis.
Here they can give you a lot of advice, but these will be hypotheses and assumptions only.
Nikita48
Dancing with a tambourine
-
-
-
#1.
nine0013