• Competitor rules

    Please remember that any mention of competitors, hinting at competitors or offering to provide details of competitors will result in an account suspension. The full rules can be found under the 'Terms and Rules' link in the bottom right corner of your screen. Just don't mention competitors in any way, shape or form and you'll be OK.

:confused: 3dMark results are lower when i O.C. Asus Turbo GTX 1080 Ti 11GB

Associate
Joined
12 Oct 2017
Posts
74
:confused:
Hi!
Few days ago i got this GPU, Asus Turbo GTX 1080 Ti 11GB (1481 Mhz core / 1377 Mhz mem) which it works now on this rig:
MB: Asus ROG Strix Z270E Gaming
Processor: Intel I7-7700K delided + relided w metal liquid
Memory: TEAM GROUP EDITION 16GB (2X8GB) DDR4 PC4-25600C14 3200MHZ DUAL CHANNEL KIT
HDD: Samsung SSD 970 Evo Plus 1TB
Cooler: water, custom loop (5 rads x 140)


The proc is rock-stable at 5 Ghz, and stable max 46 celsius in full load
The memory is rock-stable as XMP 2.0 in bios, 3200 Mhz

Now my confusing is that when i overclock this GPU the 3dMark results go lower for GPU, as you can see:

edit


https://docs.google.com/spreadsheets/d/1xbEPtwufS7T0mV1Mc9LYcY8gmebWtewBlGzGnUnD1MQ/edit?usp=sharing


I searced a lot on internet but did not find too much. Howerver, i found this 2 years old comparasion, what is saying exactly what i notice, that overclocking this GPU dont bring higher results in 3DMark:

https://www.3dmark.com/compare/spy/1303084/spy/1307626/spy/1310153/spy/1301697#

Is a 3dMark problem? Is a specific GPU problem? It matter? I mean should i ignore the 3dMark results because (maybe) in games it make a difference if the GPU is overclocked? I did not measure any results in games, yet.
Cheers!

l.e. i forgot to mention that the GPU is watercooled and the temp is maxim 33 celsius in full load.
 
Overclocking a 1080Ti makes a huge difference in 3Dmark (and games).

Could you link to your actual 3dmark result? Are those results on your datasheet showing graphics score or total score?

Next, what is your card actually boosting to? A stock speed of '1481mhz' means nothing - if your card was really stuck at 1481mhz, something would be dramatically wrong. Even at stock it should be boosting toward 1900mhz+ on water. So keep Afterburner open and have it monitor clockspeed. Check what the maximum and average mhz clockspeed is during the run at stock. Then see what happens to it when overclocked. Is adding to your offset actually increasing the mhz at peak usage during the benchmark? Keep Afterburner open on a second screen and monitor it closely and report back here :)


Thank you Maldoror!
The results refer strictly to GPU as the CPU results are ok (increase when the processor is at 5 GHz and decrease when turn off the processor overclock). I put the results only for GPU because i played a lot while OC (increasing-decreasing, GPU) and its much simple like that.
I will fallow your advice and i will come back with results.
Cheers!
 
I did what you adviced me, thank you!
The GPU is set now at +76 MHz for GPU Core and +459 MHz for GPU Mem.
I dont know where or how exactly to set the "default" for Nvidia 3d settings but i open the Nvdia controll panel and there i found something like "restore default" and did it... Also there is a refresh rate and i have to chose between 6o Hz and 100Hz. I left 60.
Then i got GPU-Z, see the screenshot, it says my card is overclocked and working on proper PCI-E speed, i think.
But then, during the 3dMark, the MSI Afterburner overlay it shows the base frequencys, not any fluctuation, it shows 1481 MHz for core and 5468 for mem. Should i set something in MSI Afterburner to show the actual frequency or is something not working properly?
Cheers!


We'll get to the bottom of this.
.....
So the first thing we need to establish - before worrying about overclocks - is what the card is actually doing and whether it's working correctly.

1) Go to Nvidia control panel and set 3d settings to 'default'. You may then need to set 'preferred refresh rate' (about 2/3rds of the way down) back to 'highest available' if you don't use G-Sync. But just to make sure you don't have some odd settings there.
2) Get GPU-Z and check the card is running in the right PCIE mode. Probably 16x 3.0 but to cause these problems it has to be something extreme, so 8x or suchlike is probably still okay. Regardless, post your result for that. The PCI-E mode is shown on the main GPU-Z screen ("Graphics card") in the middle right under 'Bus Interface'.
3) Set up MSI Afterburner (or EVGA Precision) and monitor gpu usuage and gpu clock speed. Run 3dmark and check that the card is under 99% usage load throughout the test and watch the clockspeed. We will see some fluctuations due to power limit but since it's watercooled and temps are fine (but check temps too, during the run), I would not expect it to drop below 1850-1900mhz during the test, or thereabouts.

Do these things and report back and we'll continue.
 
The temps are 29-30 celsius, 29 at the begining of test, 30 at the end of test.
I will try to use Asus Tweak instead MSI AB and come back with info.
Cheers!

l.e. how can i open and use GPU-Z during 3DMark test without intrerupting the benchmark? I dont have a secondary monitor / display
 
I know the temps are too low but i dont know if there the problem comes... who knows...
The GPU is watercooled. Even so, i know the temp are too low, as in my room is now 21.5 celsius, the water temp is now 28 celsius and MSI and GPU-Z shows, both of them, 22 celsius.... Defo something is wrong.

Try a different benchmark to double check
Firestrike or unigine etc

On a side note how are you measuring those temperatures as they are too low
I know Aisuite will be wrong if using that for the cpu
But 33c max on full load is also too low for a gpu
Around 30c is a typical idle for a gpu
45c to 50c under heavy load would be what I would expect
Unless your fans are at 100%
Or its Very cold in your room
 
Do you have a link for "GPU-Z render test" please?
I will do this then i will try a DDU driver clean. Till yesterday i had (NZXT) CAM software and yesterday i uninstall it, to be sure is not interfere with MSI AfterBurn, as CAM have as well, a tab for overclocking GPU... Who knows.... maybe that CAM software is to be blamed.... i have it because i have a NZXT Fan Hub in my rig.

Actually if u can post a screenshot of gpuz sensors tab open make sure all the boxes are clicked on twice to show max (says max in tiny letters) run the gpuz render test for 3 mins or so and let us see. Mainly the sensors for temps, power draw and clocks, rest not so much.
 
Yeah, seems that somehow the GPU is capped at the base frequency (1481 MHz). See the min and max recorded by GPU-Z after 3 min of GPU-Z render then full 3DMark.
min:
https://photos.app.goo.gl/LXVUpC6vhRjBHY98A

max:
https://photos.app.goo.gl/eBq4NHD8uroVijWr8

The CAM from NZXT is a software what manage the fans connected in NZXT Fan Hub. And this soft also have a overclocking tab, thats why i uninstall it.


Its built into gpuz. As said, on ur screenshot u posted, see the ? near PCI info box near right side, click the ? that opens a new window and click start render test. Then go to sensors tab on gpuz and watch the clocks value go up and the red bar raise up.

Never heard of that cam software, i wouldnt use it tbh, id just use gpuz or hwinfo or afterburner, not all at the same time of course lol.
 
In the meantime i did a DDU and now i am in the intalling Nvidia driver process. Should i chose Driver + GeForce Experience or just the driver?
The PSU is ok! Its a Seasonic 750 PLus Gold, and anyway... since i didnt hit any voltage settings while o.c. not the PSU should be blamed
Yes, i had Asus G-Tweak II installed, yesterday... trying to sort out this problem.
So, better to install just the driver or GeForce experience as well?
Thank you so much for your help!

l.e. i installed just the driver, not GeForce Experience
 
After DDU then clean install of Nvidia driver, then I ran the Unigine Heaven in "Extreme" preset for about 2min while monitor with GPU-Z.
The same, the GPU is stucked in base frequency. I mean its capped at base frequency, is not stucked.
 
Last edited:
I didnt expect so much patience and so much help from forum members to be honest!
THANK YOU, again!
Well...i took off the oc, then i uninstalled the MSI AB as well, now is not any oc soft installed as i am aware :D
Yes, i did DDU in safe mode then after restart i installed the driver again, exactly that one you said, without GeForce Experience.
Yes, the max VDDC during Unigine Heaven is 0.7930. The min VDDC (while browsing, not during Heaven) is 0.65
Have no idea why and also have no idea how to fix it...

l.e. the power plan of Windows is "High Performance"



The power is only 0.700v when running it? Thats idle power draw, its not going up. That would be my guess at the cause. Something is preventing it from drawing full power that it needs. The fan header on the gpu could be limiting the power draw if not attached, i dunno for sure but like the guy above said somewhere it could be that. That or maybe something wrong with the pci-e cables that you attached to the card. There in, arnt they?

You have taken off any overclock now havent you, on core or mem or volts?

If perfcap says idle then its in idle mode the gpu so wont boost.

Also just want to make sure of something, when you did DDU did you do it in safe mode or normal windows mode?
 
I will try a diff PCI slot, it will be a bit difficult, i dont know if the hoses let me but i will try. Also i will buy new cables between PSU and GPU.
The cables are properly attached. Its a 6 pin + 8 pin cables and they are properly attached to GPU and to PSU as well.
I bought the GPU from a friend of a friend and its useless to ask as i dont have any trust on what he can say about.... I bought it with original air cooling system and in same time i bought a waterblock from china (Byksky or something like that) because this was my intentions from the begining when i start to build this rig but in the past the prices of 1080 Ti was too high for me. With the original air cooling (what i keep it just 1 week till my waterblock arrived) the temps go too high (constant 80-84 in games, but im playing at 3440x1440) and i didnt try to oc, not even monitoring frequencies. I can put back the air to see whats happened.

Nice, we have something clear to work with now. The card is stuck in Idle power mode (and at the idle voltage).

This can be caused by an incorrect PCI-E slot mode setting; as mentioned above, try a different PCI-E slot for the card, if possible.

After a little google around, I did find several people with Asus cards who solved it with a vbios flash. This is also an option (and also a trick that helps 2080Tis for example; Gamer's Nexus had a video about a certain 2080 Ti bug that caused those cards to lock in idle power mode too, solved by a bios flash).

Double check all the external PSU plugs are in, btw.

Did the card come with a waterblock or an aircooler, when you bought it?

Nope, i dont have other cables, i will buy these days some. Yes, it have 6+8 connectors / cables, properly attached. Cheers!

We are a stubborn lot lol we don't stop until you get sorted

Not sure if it was asked as lot if info in the thread
But have you tried connecting different pci~e cables to the gpu?
I already assumed if it's got 2 connectors you connected both

I think its the latest Windows version but not sure: https://photos.app.goo.gl/G1mai2cG5DvTYCJU7
Just now i reinstalled the Asus GTweak and hit the OC button then "Apply".
Whats "funny" regarding this is that i start to play a bit with voltage scale and...... ive notice that while in GTweak i increase the voltage the GPU-Z show a decrease of voltage. I added 60 mv in Asus Tweak but GPU-Z shows a decrease from 0.793 to 0.765 or something. Probably that's why i got lower result with the GPU oc.

Are you on latest win 10, to check type winver in start search box and report back with a screenie unless u wana type it out. I read some guy fixed his issue by updating to a later version of win 10 at least creators edition.

Also some posts i found has said on asus tweak they clicked silent button and it wont boost any more, i think you said you had that on before. If you can, check see if silent button is on still or if you uninstalled it, reinstall and see if its still saying silent mode and try to switch it off then u can remove the software. I wonder if it might be saved as a registry entry. If so maybe a ccleaner run might help i dunno.
 
Last edited:
Is that when your benchmarking or doing gpuz or not stressing the card.

Also did you say you bought a waterblock before you got the card? Did you know you were defo getting the asus turbo 1080ti and so the waterblock is defo ok with it because ive read some waterblock fitting sites and it says the turbo card has a slightly different pcb design compared to other 1080ti cards and it was difficult to get a waterblock (didnt check the dates as might have been years ago or recent i dunno) for one that fit as intended. Just wondering about this after reading about that.

Do u know the exact name and model of the waterblock so we can google it. I mean i think u probably have done it all correctly getting the right block for the card etc, just thought nice to double check everything incase the waterblock maybe not doing something it suppose to be doing somehow somewhere.

Yes, i was trying to play with the voltage then use a benchmark and keep an eye on GPU-Z voltage.
Bought them in the same time, few hours after i bought the card i bought the waterblock. And its a proper one for this card and i fit it myself, using very good quality thermo paste and pads from overclockers.co.uk I bought it from china because i didnt find any here. Now all are focused on RTX's.
The waterblock is this: https://www.aliexpress.com/item/32836900125.html
The GPU is this: https://www.techpowerup.com/gpu-specs/asus-turbo-gtx-1080-ti.b4307



Interesting. For the card to hit 80-84 in games on its air cooler, it must have been doing more than 1481mhz/0.7v. Sure the air cooler isn't great on the Turbo but 84C is still fairly high and the temp reductions from Idle power modes are massive.

This reinforces my suspicion that the unplugged/0rpm fan is related to these problems (and/or Asus software associated with 0rpm/idle modes).

Edit - let's keep investigating and see what others think too but my inclination at this point would be to flash to a different vbios, which often fixes 0rpm problems.

Yes, more than sure you're wright, make sense!
Do you have any idea where i can find a vbios to flash it? It is difficult?

Its a shame that the card doesnt have a silence mode bios/switch on the pcb like some cards do as then its probs a no brainer that might be it lol.

Maybe have and i dont know....
 
Last edited:
Hi pals, i tried in this morning to post but i wasnt allowed because i reached 15 posts in 24H. So....
I reinstalled the Windows 10 last night then I flashed the vbios, put that one you advised me but unfortunately the problem remain the same, while im running 3DMark (now i dont know why unigine is not runnning anymore) the VDDC goes in max 0.7930 and GPU core maxim 1480.5 MHz and so on...
Adding to that, the Chrome browser wasnt working, even reinstalled.
Mine is ending in .22, the other one is ending in .01

https://photos.app.goo.gl/o4PNX79ZTTidTYjY8

So now i just flashed back my vbios and at least chrome seems to work properly.

Wednesday i will receive water cooling spare parts to be able to bypass the GPU with the hoses and then i will be able to try other PCI-E slot and also to put the original air cooling system and to go further in this game.
In the meantime, if anyone here wants to play, this is my vbios: https://drive.google.com/file/d/1nu-ua781xnU3fnGaP06omVyO7J5aB__U/view?usp=sharing
Im not able to cook such a pie, is too much for my currently knowledge.

Cheers!
 
Hi pals!
Today i got the fittings so ive been able to move the GPU out of that PCI-E slot.
First, after i bypass-ed the GPU watercooling was to mount back the original (air) cooling system. And with that one i mount back the GPU on the same PCI-E slot. Results: THE SAME (GPU dont boos, using Unigine Superposition benchmar and double check with GPU-Z recordings)
Then, ive moved the GPU on the next PCI-E slot (x8) and the result: THE SAME (GPU dont boos, using Unigine Superposition benchmar and double check with GPU-Z recordings)
Then, ive took out this amazing shaity GPU and ive put back my old GTX 970 4GB, on the X16 PCI-E slot. The result: the GPU boost to 1300-ish... from 1088 (for core) (using Unigine Superposition benchmar and double check with GPU-Z recordings)
SO... this GPU have a problem. The problem could be in his vbios or could be a physical one. If any of you is keen to find out i can send post it then of course i will pay the post back and the beers as well! Otherwise i will put it on ebay with mentions that is not boosting.

By the way: please, if you dont mind, of course, stay focus on what matter, not what my girlfriend can do / buy with that amount of money, if i like it, what i can do with that amount of money and so on....
Cheers, you are absolutely fabulous, all of you and i like you!
 
Leeds area, Pontefract more precise.


Yeah I would try the Asus 1080ti oc bios
If the op is anywhere near me would be happy to have a mess around with the card
And yeah I wondered whether would take a big hit on ebay and lose a lot of money
Probably paid about £70 for a byski block on top of what he paid for the card too
 
Yes, at some point i did it. I will reinstall now afterburner and try again. As i remember didnt help at all and as i remember as well each time when i tick the "power" slider, after i applied the modification, few seconds after that the windows freeeze and was necessary to reboot. Each time.
Can you give me a hint / help for flashing an xoc vbios, when you have few minutes, please!?
Cheers!
 
Back
Top Bottom