• Welcome to TechPowerUp Forums, Guest! Please check out our forum guidelines for info related to our community.

GPU Temp: 68°C, Hotspot Temp: 100°C (Sapphire Pulse RX 6700XT)

Joined
Feb 24, 2023
Messages
2,362 (4.55/day)
Location
Russian Wild West
System Name DLSS / YOLO-PC
Processor i5-12400F / 10600KF
Motherboard Gigabyte B760M DS3H / Z490 Vision D
Cooling Laminar RM1 / Gammaxx 400
Memory 32 GB DDR4-3200 / 16 GB DDR4-3333
Video Card(s) RX 6700 XT / RX 480 8 GB
Storage A couple SSDs, m.2 NVMe included / 240 GB CX1 + 1 TB WD HDD
Display(s) Compit HA2704 / Viewsonic VX3276-MHD-2
Case Matrexx 55 / Junkyard special
Audio Device(s) Want loud, use headphones. Want quiet, use satellites.
Power Supply Thermaltake 1000 W / FSP Epsilon 700 W / Corsair CX650M [backup]
Mouse Don't disturb, cheese eating in progress...
Keyboard Makes some noise. Probably onto something.
VR HMD I live in real reality and don't need a virtual one.
Software Windows 10 and 11
Temperatures below 110C on your hotspot are basically safe. Below 100C is the comfort zone. Below 90C = overclocking party invitation.
As far as you're within 110C and your junction-to-hotspot delta ain't all around the place (70C junction and 95C hotspot is sane; 70C junction and 110C hotpot is off) you don't need to worry about it much. Your GPU will die... about a dozen years from now. And what can I say, 12 y.o. GPU rarely do deliver playable FPS in recent titles.
 
Joined
Dec 19, 2023
Messages
444 (2.00/day)
Location
Gulf Breeze, FL
Processor Ryzen 9 7900 - Ryzen 7 8700G - Ryzen 7 7800X3D
Motherboard MSI B650 Project Zero - MSI B650 Project Zero - MSI B650 Project Zero
Cooling Custom water loop - Noctua NH - P1 - Custom water loop
Memory 64GB DDR5 - 64GB DDR5 - 32GB DDR5
Video Card(s) Integrated - 780M - Red Dragon RX 6800 XT
Display(s) Samsung Odyssey G7 32" WQHD 2560x1440 240Hz
Case JONSBO D300 - JONSBO TK-1 - Montech King 95
Power Supply Be Quiet 12M 850 - Be Quiet 12M 750 - Corsair 850
Mouse Logitech G403 + Logitech Trackman Marble
Keyboard Logitech G512
Software Win 11 - AIDA64 - HWInfo - Afterburner
My RX 6800 XT records hot spot at 74 c in Starfield...all default settings except the fan curve via Afterburner.
 
Joined
May 5, 2023
Messages
47 (0.10/day)
Temperatures below 110C on your hotspot are basically safe. Below 100C is the comfort zone. Below 90C = overclocking party invitation.
As far as you're within 110C and your junction-to-hotspot delta ain't all around the place (70C junction and 95C hotspot is sane; 70C junction and 110C hotpot is off) you don't need to worry about it much. Your GPU will die... about a dozen years from now. And what can I say, 12 y.o. GPU rarely do deliver playable FPS in recent titles.
Of course, I don't expect the GPU to work for life. If I can use a GPU for 5 years, that is enough time for me. Before switching to AMD GPU, the hotspot temperature of the GPUs I used (GTX 1070, RTX 2060 Super, RTX 3060Ti) was 75°C at most. But when I switched to the RX 6700XT, I was disappointed when I saw 100°C. When I sent it to technical service due to a heating problem, they sent me a brand new card. Of course, since the fan curve in the Sapphire Pulse 6000 series was problematic, I managed to keep the GPU temperature between 68-72°C and the hotspot temperature between 78-83°C by manually adjusting the fan.

I undervolted, set the MHz value to 2581 and the mV value to 1150, and now the GPU temperature is between 55-63°C and the hotspot temperature is between 74-78°C.

I'm happy with the GPU now. However, I feel like I will have problems with this issue again in the future.
 
Joined
Sep 3, 2019
Messages
3,051 (1.70/day)
Location
Thessaloniki, Greece
System Name PC on since Aug 2019, 1st CPU R5 3600 + ASUS ROG RX580 8GB >> MSI Gaming X RX5700XT (Jan 2020)
Processor Ryzen 9 5900X (July 2022), 160W PPT limit, 75C temp limit, CO -9~14
Motherboard Gigabyte X570 Aorus Pro (Rev1.0), BIOS F37h, AGESA V2 1.2.0.B
Cooling Arctic Liquid Freezer II 420mm Rev7 (Jan 2024) with off center mount for Ryzen, TIM: Kryonaut
Memory 2x16GB G.Skill Trident Z Neo GTZN (July 2022) 3600MHz 1.42V CL16-16-16-16-32-48 1T, tRFC:280, B-die
Video Card(s) Sapphire Nitro+ RX 7900XTX (Dec 2023) 314~465W (378W current) PowerLimit, 1060mV, Adrenalin v24.7.1
Storage Samsung NVMe: 980Pro 1TB(OS 2022), 970Pro 512GB(2019) / SATA-III: 850Pro 1TB(2015) 860Evo 1TB(2020)
Display(s) Dell Alienware AW3423DW 34" QD-OLED curved (1800R), 3440x1440 144Hz (max 175Hz) HDR1000, VRR on
Case None... naked on desk
Audio Device(s) Astro A50 headset
Power Supply Corsair HX750i, 80+ Platinum, 93% (250~700W), modular, single/dual rail (switch)
Mouse Logitech MX Master (Gen1)
Keyboard Logitech G15 (Gen2) w/ LCDSirReal applet
Software Windows 11 Home 64bit (v23H2, OSB 22631.3737)
Of course, I don't expect the GPU to work for life. If I can use a GPU for 5 years, that is enough time for me. Before switching to AMD GPU, the hotspot temperature of the GPUs I used (GTX 1070, RTX 2060 Super, RTX 3060Ti) was 75°C at most. But when I switched to the RX 6700XT, I was disappointed when I saw 100°C. When I sent it to technical service due to a heating problem, they sent me a brand new card. Of course, since the fan curve in the Sapphire Pulse 6000 series was problematic, I managed to keep the GPU temperature between 68-72°C and the hotspot temperature between 78-83°C by manually adjusting the fan.

I undervolted, set the MHz value to 2581 and the mV value to 1150, and now the GPU temperature is between 55-63°C and the hotspot temperature is between 74-78°C.

I'm happy with the GPU now. However, I feel like I will have problems with this issue again in the future.
You cant really compare hotspot temps between AMD/nVidia. Apples and oranges...
Both brands reveal exactly what they want users to see. So you cant be sure whats really going on inside the GPU die.
I think AMD has been more honest about it since Zen2/RDNA1 first came out in 2019.

Of course there is the cooler/contact implementation subject by vendors.
I think most of them are just seeking ok-ish cooling performance because of the 110C operating limit.
 
Joined
Jan 9, 2024
Messages
82 (0.41/day)
System Name Ryzen AM5 + Radeon 7000
Processor R7 7700X
Motherboard ASUS ROG STRIX B650E-E WIFI
Cooling AK500 ZERO DARK + 2x Arctic 120MM static opt. PWM fans + KRYOSHEET
Memory GSKILL Flare X5 32GB (2x16GB) DDR5-6000 CL36
Video Card(s) ASUS TUF Gaming RX 7700 XT OC 12GB
Storage WD Black SN750 500GB (OS Drive) + CRUCIAL P5 PLUS 2TB (Game Drive) + (20) Crucial MX500 250GB
Display(s) LG 34" 1440p ULTRAWIDE 160Hz (main) + 2X LENOVO 27" 1080p 75Hz (sides)
Case FRACTAL DESIGN Pop Air w/ 3x140MM + 2x120mm ARCTIC PWM fans
Power Supply DEEPCOOL PQ850M 850W Gold
Mouse Logitech G502 Gaming X Wired
Keyboard Microsoft ANB-00006 Wired Slim Keyboard
Software WINDOWS 11 PRO
Of course there is the cooler/contact implementation subject by vendors.
I think most of them are just seeking ok-ish cooling performance because of the 110C operating limit.
It is absolutely a die/coldplate contact issue.
I found a way to bring my 6700 XT to what it was supposed to be. Once I solved my contact issue it basically unlocked tons of OC and UV potential. The before and after was insanely different. I also don't cringe anymore when I peak at HWinfo64. Seeing 100-110°C and/or 15-30°C delta was sooo unsettling... undervolting was OK, but still was seeing high deltas at times and poor freq behavior, poor OC and UV performance. The stable ranges are WAY larger, I can hit much higher freq and lower voltages now. Even at stock settings after the fix I was seeing a performance increase due to higher and sustained clocks.
You will void the warranty though... personally, if I can't fix it myself I replace it. To each his own. Personally don't like RMA-ing, never had a good experience. I also work in the Tech industry and take electronics/computers/servers apart to service/refurbish regularly so this was right up my ally.
 
Joined
May 5, 2023
Messages
47 (0.10/day)
Unfortunately, after 8 months, the GPU's hotspot temperature rose again. On the first day of the GPU, the normal temperature was 70 and the hotspot temperature was 81. Of course, I used it with undervolt since the first day and while the normal temperature was around 62, the hotspot was in the range of 80-85. When I tried to test without undervolt last night, the normal temperature was in the 62-65 range, but the hotspot was close to 90 degrees again.

I think this will be an inevitable end for me. :fear:
 

#22

Joined
Apr 13, 2023
Messages
344 (0.73/day)
Location
Warszawa
Unfortunately, after 8 months, the GPU's hotspot temperature rose again. On the first day of the GPU, the normal temperature was 70 and the hotspot temperature was 81. Of course, I used it with undervolt since the first day and while the normal temperature was around 62, the hotspot was in the range of 80-85. When I tried to test without undervolt last night, the normal temperature was in the 62-65 range, but the hotspot was close to 90 degrees again.

I think this will be an inevitable end for me. :fear:

Keep in mind that hotspot temps temps and size of a difference tend to rise with growth of ambient temp. Maybe your case.
 
Joined
May 5, 2023
Messages
47 (0.10/day)
Keep in mind that hotspot temps temps and size of a difference tend to rise with growth of ambient temp. Maybe your case.
I have a mesh case. There are 2 Akasa Viper AK-FN063 fans on the front of the case as intake fans. There are 2 Arctic F120 exhaust fans on the top and 1 Arctic P120 exhaust fan at the back. My cafe is this;
1715931387713.png

The air temperature outside is in the range of 18-20 degrees.
 

#22

Joined
Apr 13, 2023
Messages
344 (0.73/day)
Location
Warszawa
I have a mesh case. There are 2 Akasa Viper AK-FN063 fans on the front of the case as intake fans. There are 2 Arctic F120 exhaust fans on the top and 1 Arctic P120 exhaust fan at the back. My cafe is this;

The air temperature outside is in the range of 18-20 degrees.

It doesn't matter what cooling config you have - you won't fight much temperature growth from higher ambient with airflow. Then it just cools your components with warmer air ;)

Ofc there's always option that thing responsible here is paste degradation, but if your ambient change recently, I would first pay attention to it when doing measurements.
 
Top