2080 Ti Aorus Xtreme Wateforce - oc low

sewerino26 · 8 Sep 2019 at 20:21

Hi. I have Aorus Rtx 2080 Ti Waterforce Xtreme. So on stock is rock stable,no single crash on all games.Temps 58-60C,boost 1950-2010mhz.

I tried yesterday oc for Remedy Control ( for 1440P without DLSS ) <haha>. And i have poor result
I maxed power limit 110% ,fan pump 70% ( loud as hell ) and core +70.Temps was 50C
But on 2055-2070 it will crash no matter what.

Bios is F10. My question is why i have so such low result,is this related to ray tracing?
And why power limit is only 110% not higher? Thx

TheRealDeal · 8 Sep 2019 at 20:43

That's pretty normal. You buy the Auros Xtreme for high clocks out of the box which from my mates 2080ti/1080ti were around 2000mhz core out of the box on both. You then don't have much headroom to OC. With Turing most of the chips will max out around 2050mhz. That's what i seen from both of the Auros xtreme's i have played around with.

Edit just seen you have the water version. The chip is the limit so probably no advantage to having that version apart from it should keep temps lower with less noise. The card is probably needing more voltage/power which is limited in the bios i believe.

sewerino26 · 8 Sep 2019 at 20:48

So left in stock?

TheRealDeal · 8 Sep 2019 at 20:51

sewerino26 said:
So left in stock?

Yea it's plenty fast as is. Some win the lottery, some get normal and others just plain lose. You seem to have hit the normal kind of range which as i said is plenty fast.

sewerino26 · 8 Sep 2019 at 21:03

TheRealDeal said:
Yea it's plenty fast as is. Some win the lottery, some get normal and others just plain lose. You seem to have hit the normal kind of range which as i said is plenty fast.

oki thx brother

Poneros · 8 Sep 2019 at 21:31

Something else to keep in mind is that Control is the game that will drive the GPU harder than any other, even without RT enabled. What's stable OC otherwise is not going to apply here.

Grim5 · 9 Sep 2019 at 00:02

TheRealDeal said:
That's pretty normal. You buy the Auros Xtreme for high clocks out of the box which from my mates 2080ti/1080ti were around 2000mhz core out of the box on both. You then don't have much headroom to OC. With Turing most of the chips will max out around 2050mhz. That's what i seen from both of the Auros xtreme's i have played around with.

Edit just seen you have the water version. The chip is the limit so probably no advantage to having that version apart from it should keep temps lower with less noise. The card is probably needing more voltage/power which is limited in the bios i believe.

Maybe Gigabyte just have low binning. I've played with three ASUS Strix 2080ti's and all three have successfully overclocked to between 2130 and 2170mhz on air - with water they may have gone even higher.

TheRealDeal · 9 Sep 2019 at 00:19

Grim5 said:
Maybe Gigabyte just have low binning. I've played with three ASUS Strix 2080ti's and all three have successfully overclocked to between 2130 and 2170mhz on air - with water they may have gone even higher.

Just checked the timespy thread some go high some don't. It's all about luck of the draw really. In games the extra mhz is not really giving much more fps compared to 2050 mhz but handy for more points in benchmarks. Can you mod the bios for more power limit/volts or get it through different software. They only get 2085 in this review so not to far off.

https://www.overclock3d.net/reviews..._2080ti_aorus_extreme_waterforce_11g_review/3

sewerino26 · 9 Sep 2019 at 00:46

I tested only Control and on ( +core 70 ) 2055-2070 mhz its crashing after few minutes. When i am in stock no issues.
So far only in Control other benches i not tested so i dont know if only Control or somewhere else too.
I go back to stock. I dont tried +core 60 no sense

tyler_jrb · 9 Sep 2019 at 05:29

Most boost very close to their maximum clocks so 2-2.1ghz is very normal.

In terms of performance do not worry you’re talking 1-2% if that even.

You are better off overclocking the memory which yields better results.

HRL · 9 Sep 2019 at 13:11

Sounds like you’re hitting the power limit to me.

If you’re brave enough flash the BIOS to one with a significantly higher power limit and I bet you can increase the clocks some more.

Grim5 · 9 Sep 2019 at 14:20

HRL said:
Sounds like you’re hitting the power limit to me.

If you’re brave enough flash the BIOS to one with a significantly higher power limit and I bet you can increase the clocks some more.

If I’m looking at the data correctly the card should have a 366w bios - how much higher must he go? Im pushing 2150mhz On my 2080ti with just 320w to play with. Plus there aren’t that many options right - it’s a galaxy 380w file which is a small boost only or something like the kingpin bios which is like 520w - however that might be pushing the thermal limits of what that cards cooler can manage.

HRL · 9 Sep 2019 at 15:15

Grim5 said:
If I’m looking at the data correctly the card should have a 366w bios - how much higher must he go? Im pushing 2150mhz On my 2080ti with just 320w to play with. Plus there aren’t that many options right - it’s a galaxy 380w file which is a small boost only or something like the kingpin bios which is like 520w - however that might be pushing the thermal limits of what that cards cooler can manage.

Ah, if your figures are right then that’s a fair point. Have to say I didn’t look at what wattage his existing BIOS peaks at, just saw he was on water and getting very average clocks.

I’d be a bit pee’d as my Zotac AMP, non-extreme, gets higher clocks on air with the stock BIOS. :eek:

tyler_jrb · 9 Sep 2019 at 18:47

HRL said:
Ah, if your figures are right then that’s a fair point. Have to say I didn’t look at what wattage his existing BIOS peaks at, just saw he was on water and getting very average clocks.

I’d be a bit pee’d as my Zotac AMP, non-extreme, gets higher clocks on air with the stock BIOS.

Water never has done much for gpu’s really, especially when you have a poor/average clocker anyway. More so the noise reduction I guess and better stabilised clocks (custom loop etc).

366w is more than enough though. Increasing the power limit rarely makes a difference unless it is severely limited in the first place as there is very limited voltage adjustment.

Mine doesn’t go over 330w and runs at 2100mhz 24/7.

BigBANGtheory · 9 Sep 2019 at 19:17

tyler_jrb said:
Water never has done much for gpu’s really, especially when you have a poor/average clocker anyway. More so the noise reduction I guess and better stabilised clocks (custom loop etc).

Is that really the case? I thought with say the 1080Ti water allowed you to run the GPU consistently at the boost frequencies which would otherwise thermal throttle you back down. To me the 2000 series doesn't count its just a fill in product line with all manner of nonsense going on like power limits and add features over raw speed.

tyler_jrb · 9 Sep 2019 at 20:43

BigBANGtheory said:
Is that really the case? I thought with say the 1080Ti water allowed you to run the GPU consistently at the boost frequencies which would otherwise thermal throttle you back down. To me the 2000 series doesn't count its just a fill in product line with all manner of nonsense going on like power limits and add features over raw speed.

You are talking like 1–2% performance between air cooler vs water cooling.

I actually switched at one point from liquid cooled 1080Ti to air cooled and it ran within 50mhz of the liquid cooled card purely because of gpu boost 3.

BigBANGtheory · 9 Sep 2019 at 21:41

tyler_jrb said:
You are talking like 1–2% performance between air cooler vs water cooling.

I actually switched at one point from liquid cooled 1080Ti to air cooled and it ran within 50mhz of the liquid cooled card purely because of gpu boost 3.

Yeah I'm not sure that's true though, I think what you are talking about is peak perf where as what Im trying to say is it doesn't matter about peak its about the consistency i.e. at what frequency is sustainable

Grim5 · 9 Sep 2019 at 21:46

tyler_jrb said:
You are talking like 1–2% performance between air cooler vs water cooling.

I actually switched at one point from liquid cooled 1080Ti to air cooled and it ran within 50mhz of the liquid cooled card purely because of gpu boost 3.

Yes and no. 1-2% maybe if you run the air cooler at speed high enough to give you a headache.
You need to normalise the results - run both cards where they are say producing 40dba of sound only - now compare the performance and clocks.

I just don't think it's fair to compare one cooler that's running all out and another that's barely pushing a sweat, normalise the results so you are comparing apples to apples - this is how Tech Jesus at Gamers Nexus does all his testing.

tyler_jrb · 9 Sep 2019 at 21:47

Grim5 said:
Yes and no. 1-2% maybe if you run the air cooler at speed high enough to give you a headache.
You need to normalise the results - run both cards where they are say producing 50dba or 40dba of sound only - now compare the performance and clocks.

It was an Msi gaming x running at stock fan speed of around 60% 70c. Very very quiet card.

BigBANGtheory said:
Yeah I'm not sure that's true though, I think what you are talking about is peak perf where as what Im trying to say is it doesn't matter about peak its about the consistency i.e. at what frequency is sustainable

It sustained 2000mhz 24/7 even at 70c. And that was backed off a bit. Usually would do about 2020-2050mhz without much issue.

BigBANGtheory · 10 Sep 2019 at 21:30

Unless its a really good (expensive) air cool solution the 1080Ti is well known to thermal throttle down from its boost speeds, yes it can hit them but it can't sustain them under load without "cooling help". This is the value of water cooling for that range and others but not the 2000 series due to power limits.