das3erawsdas 2019 年 3 月 26 日 下午 12:53
2080 Ti - power limit and oc question
Hello. My question is. Card is Rtx 2080 Ti Aorus ( 1695mhz stock boost ). So when i max power limit 133% i have higher temps and i cant overclock high my card because it crashing in firestrike gpu test 1 usually after 1-2 hours on loop. I can ignore that because Heaven and games are stable.

But when i ticked 115% power limit i have lower temps and higher stability when oc for example +70 core. Is this normal?
最後修改者:das3erawsdas; 2019 年 3 月 26 日 下午 12:54
< >
目前顯示第 1-10 則留言,共 10
ugafan 2019 年 3 月 26 日 下午 1:08 
more power = higher temperature

your card is is shutting down to prevent damage from too much heat.

go with the more stable overclock.
das3erawsdas 2019 年 3 月 26 日 下午 1:20 
110 power?
ugafan 2019 年 3 月 26 日 下午 1:39 
how well an individual card overclocks is going to vary. settings that work for your card may not work for someone else's card. the only way to figure it out is to do the type of testing that you're doing.

some people go for an aggressive overclock to get maximum performance. others prefer a a smaller overclock to keep temperature and fan noise down. it's really up to you what you're going for.
Omega 2019 年 3 月 26 日 下午 11:00 
You can safely max the power limited. It will destroy your power efficiency but it will make sure the GPU isn't held back.
You need to understand how nvidia's boost is working. So when you set a lower power limit of say 115% and then you are able to overclock it +70 Mhz then this would be normal. Then later when you increase the card's power limit to say 133%, then the card will automatically self-overclock to the exact same speed as if you were at 115% power and manually setting it to +70 Mhz all on it's own. When you give the card more "headroom" by allowing it to consume more power with setting it to 133% then it self-overclocks as high as possible. And that's also why it's running hotter: With the extra power room it's self-overclocking higher and thus running hotter. What you need to be looking at is watch the actual core Mhz Speed and see what Mhz the core is running at when at 133% and again when at 115% with +70. It should be the same on both settings at least. That's why you can't manually overclock it when you set it to 133%: It's already doing it on it's own.
[☥] - CJ - 2019 年 3 月 27 日 下午 7:34 
Anyway...
Dont go too far with Core as you'll have a higher chance of the OC being unstable than Putting more into Memory. +70-85 Core should be enough while being stable.

My 1060 6GB is Stable at +95, but if i go in the area of +100 thats where it starts running into problems, while Memory is perfectly fine at +300, of which i'll be increasing until the Temps start being affected.

I changed the thermal on the card last week and it dropped up to 30C with now a MAX of 50-65C
depending on the game and if Vsync is enabled when it previously maxed out around 80C Before the change with or without Vsync.

So if you are comfortable with changing the paste on the card its definitely possible to largely decrease your temps.
最後修改者:[☥] - CJ -; 2019 年 3 月 27 日 下午 7:35
引用自 ☥ - CJ -
Anyway...
Dont go too far with Core as you'll have a higher chance of the OC being unstable than Putting more into Memory. +70-85 Core should be enough while being stable.

My 1060 6GB is Stable at +95, but if i go in the area of +100 thats where it starts running into problems, while Memory is perfectly fine at +300, of which i'll be increasing until the Temps start being affected.

I changed the thermal on the card last week and it dropped up to 30C with now a MAX of 50-65C
depending on the game and if Vsync is enabled when it previously maxed out around 80C Before the change with or without Vsync.

So if you are comfortable with changing the paste on the card its definitely possible to largely decrease your temps.
Your + values are completely arbitrary and different for every single user. What you should be telling us is what actual core Mhz is stable for your card. With how nvidia boost works, every card in every computer runs a different max boost speed depending on the user's power supply and how cool they can keep the cards. It may be 1800 mhz for one person and 2000 mhz for someone else. I have a second computer with a 3GB GTX 1060 that's stable @ 2100 Mhz core speed on stock air cooler, for example.
最後修改者:🦊Λℚ𝓤ΛƑΛᗯҜᔕ🦊; 2019 年 3 月 27 日 下午 8:03
Monk 2019 年 3 月 27 日 下午 8:38 
引用自 Aquafawks
引用自 ☥ - CJ -
Anyway...
Dont go too far with Core as you'll have a higher chance of the OC being unstable than Putting more into Memory. +70-85 Core should be enough while being stable.

My 1060 6GB is Stable at +95, but if i go in the area of +100 thats where it starts running into problems, while Memory is perfectly fine at +300, of which i'll be increasing until the Temps start being affected.

I changed the thermal on the card last week and it dropped up to 30C with now a MAX of 50-65C
depending on the game and if Vsync is enabled when it previously maxed out around 80C Before the change with or without Vsync.

So if you are comfortable with changing the paste on the card its definitely possible to largely decrease your temps.
Your + values are completely arbitrary and different for every single user. What you should be telling us is what actual core Mhz is stable for your card. With how nvidia boost works, every card in every computer runs a different max boost speed depending on the user's power supply and how cool they can keep the cards. It may be 1800 mhz for one person and 2000 mhz for someone else. I have a second computer with a 3GB GTX 1060 that's stable @ 2100 Mhz core speed on stock air cooler, for example.


THIS, this is what everyone needs to understand about modern nvidia cards, the added MHz you c an add to your core is meaningless without a total and temps.

Also, for a paste change to drop 30c.... Something wasn't put together correctly, you'd normally see closer to 5c, maybe 10.
引用自 ☥ - CJ -
My 1060 6GB is Stable at +95, but if i go in the area of +100 thats where it starts running into problems, while Memory is perfectly fine at +300, of which i'll be increasing until the Temps start being affected.
I decided to add more to your comment: So what was your maximum boost clock on your 1060 before you added +95 to it? 1800 Mhz? 1900? 2000? What is you're resulting clock after the +95? Are you running 1895 after the +95? 1995? 2095?
[☥] - CJ - 2019 年 3 月 27 日 下午 11:49 
I knew someone was going to mention it lol :)

Core 2050-2080, Memory 4303, so 8606 effective, Max Temp 64-65C(currently 63C in Shadow of War)

This is a 1st Gen EVGA 1060 SSC 6GB

And yeah, i was amazed at such a large temp drop, but all i did was redo the paste with Cooler Master paste. So it probably has to do with the way that i applied the paste and possibly having better contact than it originally did. The fan barely gets used now, usually staying around 700ish RPM's which is about 20% speed.

Max power target for my card is 120% btw
最後修改者:[☥] - CJ -; 2019 年 3 月 28 日 上午 12:31
< >
目前顯示第 1-10 則留言,共 10
每頁顯示: 1530 50

張貼日期: 2019 年 3 月 26 日 下午 12:53
回覆: 10