cancel
Showing results for 
Search instead for 
Did you mean: 

Bad 1080 ti overclocking results?

wonderin
Level 8
So recently I've decided to overclock my Rog Strix 1080 Ti OC a bit with gputweak.

78424

Every time I push the core voltage slider over the 10% state, it doesn't pass the 3D Mark Fire Attack stress test – it just stops.
Without changing the core voltage offset I can push the core clock only up to 2012 mhz (+70) without failing the 3dmark test. I guess i can get a bit more from my 1080ti oc with the right adjustments. The temperatures are fine (<65C with a custom fan curve).

What am I doing wrong? Or is it just the bad luck in the lottery?
179 Views
18 REPLIES 18

Silent_Scone
Super Moderator
Yes, it can be deceiving when chasing numbers if comparing on FM scoreboard. That run is actually with CAS16 4000MHz with very tight subs (Apex VI). You can be certain that also, a lot of the scores in the top 20 are using various LOD tweaks.
13900KS / 8000 CAS36 / ROG APEX Z790 / ROG TUF RTX 4090

question: should I change the advanced parameters before starting the oc scan?

upd

After moving voltage to 100% and then applying the oc scanner curve, it boost to 2012 still in Fire Strike benchmark, but only to 1974 in the same stress test. That's weird, isn't it?

https://www.3dmark.com/fs/18018897
https://www.3dmark.com/fsst/989549

i'm going to make several scanner oc profiles with different target voltage (100/50/20/0) and run benchmarks, i guess that's how i should do.

question: should I change the advanced parameters before starting the oc scan?

upd

After moving voltage to 100% and then applying the oc scanner curve, it boost to 2012 still in Fire Strike benchmark, but only to 1974 in the same stress test.

And I've noticed that it decreases base clock to 139 mhz for some reason after the scanner. That's weird, isn't it?

https://www.3dmark.com/fs/18018897
https://www.3dmark.com/fsst/989549

i'm going to make several scanner oc profiles with different target voltage (100/50/20/0) and run benchmarks, i guess that's how i should do.[ *

JustinThyme
Level 13
Leave it at stock for the OC scan. The point of the scan is to start at stock and see how far it will go without errors.



“Two things are infinite: the universe and human stupidity, I'm not sure about the former” ~ Albert Einstein

JustinThyme wrote:
Leave it at stock for the OC scan. The point of the scan is to start at stock and see how far it will go without errors.


i guess you are wrong here, if there is an option in the scanner for it
plus it gives a stable oc with any variation of parameters
with sll stock it makes a lower curve
anyway, my simple +70 manual oc gives a better benchmark results than the oc scanner one
it's pity that my oc card is ocing that low, but i guess there will be no reasonable performance boost *with more 30 mhz on the core
thanks*

wonderin wrote:
i guess you are wrong here, if there is an option in the scanner for it
plus it gives a stable oc with any variation of parameters
with sll stock it makes a lower curve
anyway, my simple +70 manual oc gives a better benchmark results than the oc scanner one
it's pity that my oc card is ocing that low, but i guess there will be no reasonable performance boost *with more 30 mhz on the core
thanks*


Justin is right to suggest leaving the card at stock for the OC Scan, although if you know a stable memory offset it's beneficial to leave this applied. The fact the scanner is producing lower results could indicate that your own OC stability is conditional. Simply soak test the system by using it for gaming for now.
13900KS / 8000 CAS36 / ROG APEX Z790 / ROG TUF RTX 4090

Silent Scone@ASUS wrote:
Justin is right to suggest leaving the card at stock for the OC Scan, although if you know a stable memory offset it's beneficial to leave this applied. The fact the scanner is producing lower results could indicate that your own OC stability is conditional. Simply soak test the system by using it for gaming for now.


Are you saying that I should leave the voltage and temp/power target at the default values in the OC Scan adjustments?
I'm just trying to understand the basics, and I don't get it why one should leave those by default for OC Scanner, but change them for manual overclock.

For some reason, first OC Scan attempt gave me the best curve so far and all other attempts were worse. It makes different curves from one attempt to another, and all of them are lower then the default one.
And why after OC Scan base clock drops down to 139 and begins to jump from 139 to 700 MHz and back on idle while it's being 1569 MHz constantly before? Just don't get it.

Anyway, thank you for your time and patience, guys.

And I'm not really sure if these small corrections to clock boost will change my overall gaming performance at all, so maybe i'm just chasing my tail. But I've got very excited about all this stuff in process. Just want to understand my GPU behavior better and to know if I'm so unlucky with the lottery 🙂

wonderin wrote:
Are you saying that I should leave the voltage and temp/power target at the default values in the OC Scan adjustments?
I'm just trying to understand the basics, and I don't get it why one should leave those by default for OC Scanner, but change them for manual overclock.

For some reason, first OC Scan attempt gave me the best curve so far and all other attempts were worse. It makes different curves from one attempt to another, and all of them are lower then the default one.
And why after OC Scan base clock drops down to 139 and begins to jump from 139 to 700 MHz and back on idle while it's being 1569 MHz constantly before? Just don't get it.

Anyway, thank you for your time and patience, guys.

And I'm not really sure if these small corrections to clock boost will change my overall gaming performance at all, so maybe i'm just chasing my tail. But I've got very excited about all this stuff in process. Just want to understand my GPU behavior better and to know if I'm so unlucky with the lottery 🙂


You can raise the power limit to max but Id leave the rest alone. My speedo in the car reads to 140 but car wont go that fast. if you are going to adjust everything else why bother with the scanner, just do manual. The scanner is just a quick and nasty to save the 100 times of adjusting and benching and adjusting and benching etc and what you get with it is pretty much good for everything and you don't have to worry about it dumping. If you get a crappy results that card lost the lottery. There is always room for improvement over that. OC SCan on my cards put +135 on the core where I can get core to +160 manually BUT and a HUGE BUT OC settings may not work for all titles or benchmarks. I can run my cards into power limit and bench timespy extreme all day long and part of tomorrow but then go to run Firestrike, not even extreme version, and it will lock on second graphics test and I have to back off from 1810 to 1795. One would think that if it runs TSX then FS would run no sweat, just isn't so though. This is why its good to create and save different profiles if you want Max OC for every title.



“Two things are infinite: the universe and human stupidity, I'm not sure about the former” ~ Albert Einstein

JustinThyme wrote:
You can raise the power limit to max but Id leave the rest alone. My speedo in the car reads to 140 but car wont go that fast. if you are going to adjust everything else why bother with the scanner, just do manual. The scanner is just a quick and nasty to save the 100 times of adjusting and benching and adjusting and benching etc and what you get with it is pretty much good for everything and you don't have to worry about it dumping. If you get a crappy results that card lost the lottery. There is always room for improvement over that. OC SCan on my cards put +135 on the core where I can get core to +160 manually BUT and a HUGE BUT OC settings may not work for all titles or benchmarks. I can run my cards into power limit and bench timespy extreme all day long and part of tomorrow but then go to run Firestrike, not even extreme version, and it will lock on second graphics test and I have to back off from 1810 to 1795. One would think that if it runs TSX then FS would run no sweat, just isn't so though. This is why its good to create and save different profiles if you want Max OC for every title.


it's always pity to lose the lottery when you've payed for the "ticket" almost 1000$ 1.5 year ago 🙂
thanks, i've noticed that ts might run like butter while fs would crash on the 2-10 screen.
I'm using my pc for gaming and adobe software so maybe that firestrile extreme is not an indicator for me. i was relying on FS according to the Guide.
anyway, I guess there is some kind of driver related bug/issue because i'm getting strange behavior after OC Scan: it lowers the base clock very much, and it gives different boosts everytime, and the gpu results are getting worse in FS even with better boosts.

And recently my GPU fans havent worked on some boots (reboot fixes the problem), that is very confusing 😞