安裝 Steam
登入
|
語言
簡體中文
日本語（日文）
한국어（韓文）
ไทย（泰文）
Български（保加利亞文）
Čeština（捷克文）
Dansk（丹麥文）
Deutsch（德文）
English（英文）
Español - España（西班牙文 - 西班牙）
Español - Latinoamérica（西班牙文 - 拉丁美洲）
Ελληνικά（希臘文）
Français（法文）
Italiano（義大利文）
Bahasa Indonesia（印尼語）
Magyar（匈牙利文）
Nederlands（荷蘭文）
Norsk（挪威文）
Polski（波蘭文）
Português（葡萄牙文 - 葡萄牙）
Português - Brasil（葡萄牙文 - 巴西）
Română（羅馬尼亞文）
Русский（俄文）
Suomi（芬蘭文）
Svenska（瑞典文）
Türkçe（土耳其文）
tiếng Việt（越南文）
Українська（烏克蘭文）
回報翻譯問題
此主題的作者已將一篇留言標為問題解答。
討論區規矩與守則
檢舉此篇文章
I suggest to start with the hardware:
If dust has built up over time, clean your PC, aswell as the fans (shroud) and the heatsink of the GPU.
Disasemble the card, have a look at the thermal pads and replace them if necessary; if you want to replace them, either read on the www the recommended pad thickness or measure the distances yourself.
Likely the thermal paste on the GPU die has dried out. Replace it with MX-4 or something equivalent. I prefer MX-6.
Think about intake and exhaust fan configuration inside your PC aswell.
On the software side:
Reducing the Power Limit via MSI Afterburner can be an easily achievable, temporary solution until you changed the thermal paste.
Adjust the fan curve.
I highly recommend to undervolt the card.
If thermals are good, an undervolt 'sweetspot' (performance or efficiency) has been found and you still hit the power limit, you could flash the firmware with a vbios of a card with a higher power limit.
Like the OC version. They're the same card.
Under-volting the GPU even by a small amount and decrease the temps a lot. As others have said, make sure you have good airflow inside your case. Clean, cool air coming into the case and preferably blowing onto the graphics card, the GPU is the most heat sensitive component so it needs the clean air the most.
I would run a stress test on the GPU using a tool that allows you to compare with others online, 3D mark etc. This will give you a good idea if your card is behaving normally or not. 90C seems a bit hot for me, my card doesn't allow me to set the thermal limit past 83C.
People are suggesting using a water block, I'm going to strongly suggest you do not go down this route. If you are overclocking or have a load of money, then sure go for it. I initially setup a water loop for my CPU and GPU, with two 360mm radiators and very good pump. My CPU (12900k) ran a solid 15C cooler, my GPU (3080 TI) ran at the exact same temp as it did on stock air cooler. The reason the GPU ran at the same temp is because the GPU will max out until it hits the thermal limit, so I ran at the same temp but got better performance. Within 3D Mark I got 500 points more in the 4K GPU stress test, in game this translated to between 3-8 FPS.
In my opinion, this increase in performance was not worth the cost and hassle of water cooling. In the end I got rid of the water cooling kit and bought a Fractal Torrent case (see link), this can fit numerous sized fans to suit your needs, out of the box it has thee 140mm fans at the front and two 180mm fans at the bottom drawing cool air in and firing it directly at the graphics card.
https://www.fractal-design.com/products/cases/torrent/torrent/black-solid/
What is terrible on pretty much all GPUs is the way the maker has configured the fan curve. It waits too high of a temp to really start cooling the GPU with its fans. By the time these RPMs get high enough, the temp has risen way too high, on some GPUs, not always on all.
I realize they do this to keep the fan noise levels down. But you can keep your GPU under 75*C at all times without it being super noisy. Unless maybe its a high end card that uses a blower fan based cooler.
Another reason many people have issues with their GPU heating up is because most after-market GPUs don't blow the hot air out the rear slot, but instead blow the hot air simply away from the GPU surface, allowing the rest of the system to heat up. This requires good or otherwise better Case cooling regarding your intake and exhaust fans.
Customizing the GPU fan curve yourself is almost always a need if you're going to be using your PC in a manner where the GPU is going to be used heavily for many hours on-end.
Techpowerup shows the same results, actually.
https://preview.redd.it/8kal5vatkgj71.jpg?width=831&format=pjpg&auto=webp&s=4b99593c43a6b80076f7c238a25dfeb65a778c40
And anyone who knows nvidia, knows its performance is based on its boost, not its "normal top speed"
Boost will degrade based on temps, no matter what you set in control panel.
Why? Since all it does, is lock your GPU in it's "3D" state, pushing it to max clock and power for the card, never letting it idle. Which is generally, 1800 stock for many of the high end 30 series. 1800 is the "Stock" boost advertised, but most boost well into the 2000's.
Boost, is affect by temps, tested and true. Some even notice 15Hz drops even sooner per 1C
As suggested, the easier solution is increasing airflow. That applies not only to the inside of the PC; you need ample space around the case as well.
Make sure your PC case is clean and regularly remove dust out using compressed air.
If you do not have air conditioning, I do suggest installing it in your room. A hot summer day will easily compound any heating issues your PC may already have.
You also don't need to undervolt most GPUs. With NVIDIA GPUs you "could" slightly lower the max power limit in order to do that, of there was really a need to do such a thing.
Sound more to me like the Case is terrible and needs to be changed. And then have the Case configured better with regards to intake and exhaust case fans.
Then place/sit the PC Chassis in an area where it has proper breathing room around the front, top and rear.
My 3090 actually runs cooler than my 3060ti, even though I actually overclocked the former a little.