Radeon 5700xt Nitro+ extremely hot

[castbound]

Hello everyone.

First I would like to mention I am in no way knowledgeable enough about this topic, therefore I turn here for the advice.

I have recently bought a brand new PC and after the assembly my GPU runs extremely hot. I am talking 100+ degrees hot.
I tried to even adjust the fans to the absolute max, but it didn't seem to matter at all. System is otherwise stable,
I did not experience any crashes so far.

Initially I had the 8GB Sapphire Radeon RX 5700 Aktiv , but I sent it back because of the exact same issue, and I thought Nitro+ would fix this.

Apparently I was wrong.

My full setup
- AMD Ryzen 7 3800X
- Sapphire Radeon RX 5700 XT Nitro+
- 512GB Samsung 970 Pro
- 850 Watt Corsair TX-M Series TX850M Modular 80+ Gold
- 32GB Corsair Vengeance RGB PRO weiß DDR4-3200
- Corsair Hydro Series H100i RGB PLATINUM SE White
- MSI MPG X570 GAMING PLUS AMD X570
- Corsair iCUE 465X RGB Midi-Tower ATX

I live in in Germany, if anyone is wondering for the temperature, if anyone is wondering about the ambient temperature.

Above is the picture of the current situation. There are 5 fans in total. 3 in the front, 1 at the back and 1 at the top.
There is space for one more at the top as an exhaust, but that's about it.

I absolutely have no clue why it is running so damn hot
Help, please.

 

[mrthanhnguyen]

blow hot air from radiator into the case will make the gpu a few more degree hotter. Did you try to undervolt it? I got my 580 and I undervolt it and the temp went down like 5c.

 

[Caring1]

Download and run HWMonitor (if you don't already have it) and reply with a screenshot.

HWMONITOR | Softwares | CPUID

HWMonitor for Windows® x86/x64 is a hardware monitoring program that reads PC systems main health sensors : voltages, temperatures, powers, currents, fans speed, utilizations, clock speeds ... The program handles : CPU and GPU-level hardware monitoring...

www.cpuid.com

If the temperature is over 100 (I assume Celsius) the system would be throttling or shutting down.

 

[castbound]

Download and run HWMonitor (if you don't already have it) and reply with a screenshot.

HWMONITOR | Softwares | CPUID

HWMonitor for Windows® x86/x64 is a hardware monitoring program that reads PC systems main health sensors : voltages, temperatures, powers, currents, fans speed, utilizations, clock speeds ... The program handles : CPU and GPU-level hardware monitoring...

www.cpuid.com

If the temperature is over 100 (I assume Celsius) the system would be throttling or shutting down.

I just downloaded it and strangely it is not showing temperatures of GPU, or I am blind (please prove me blind)

blow hot air from radiator into the case will make the gpu a few more degree hotter. Did you try to undervolt it? I got my 580 and I undervolt it and the temp went down like 5c.

I played around with it but without much success. Temperatures didn't really drop or maybe by a degree or two.
Have any suggestions?

Download and run HWMonitor (if you don't already have it) and reply with a screenshot.

HWMONITOR | Softwares | CPUID

HWMonitor for Windows® x86/x64 is a hardware monitoring program that reads PC systems main health sensors : voltages, temperatures, powers, currents, fans speed, utilizations, clock speeds ... The program handles : CPU and GPU-level hardware monitoring...

www.cpuid.com

If the temperature is over 100 (I assume Celsius) the system would be throttling or shutting down.

GPUZ is reporting it the same as AMD Software tho. As you can see, it's around 100.

 

[mbeeston]

are the screws on the back of the card loose?

 

[castbound]

mare the screws on the back of the card loose?

Which screws exactly do you mean? Those holding it to the case?
Could that really play a role?

 

[mbeeston]

Which screws exactly do you mean? Those holding it to the case?
Could that really play a role?

no, the ones holding on the back plate and cooler, the ones on the back of the card.

 

[Caring1]

HWM probably needs to be updated to see the temps, nothing you can do about that.
GPU-z temps seem fine for memory and VRM, the 100C may indicate a dodgy application of thermal paste on the Die, from the factory, or a loose cooler mounting.

 

[castbound]

no, the ones holding on the back plate and cooler, the ones on the back of the card.

Should I attempt to do something? Is it safe?

HWM probably needs to be updated to see the temps, nothing you can do about that.
GPU-z temps seem fine for memory and VRM, the 100C may indicate a dodgy application of thermal paste on the Die, from the factory, or a loose cooler mounting.

In both cases I would have to return it, right?
It is extremely strange tho, that this is in fact already second GPU I am testing. Would be quite a coincidence to get 2 faulty in a row.

Could it somehow be my case maybe? Maybe not the optimal airflow, dunno.

 

[chr0nos]

what app are u using to stress the GPU?

Edit: I have the same card, this is what it should look like, also something wrong with UVD clocks not showing.

After 5 mins stress test

 

[castbound]

what app are u using to stress the GPU?

i used Heaven Benchmark.
But all of the above is reproducible in games too, like ancestors legacy or shadow of war.

what app are u using to stress the GPU?

Edit: I have the same card, this is what it should look like, also something wrong with UVD clocks not showing.
View attachment 151080 View attachment 151081

yes that looks reasonable on your side.
i just have no clue what to do honestly.
i mean, 100 degrees is outrageous.

 

[chr0nos]

U need to get a full sensor screenshot of the gpuz sensor tab under load to get a better idea

It may be GPU voltage crazy high also power draw

just lightly turning the screws to make sure they're tight shouldn't hurt anything

That could work too

 

[mbeeston]

Should I attempt to do something? Is it safe?

just lightly turning the screws to make sure they're tight shouldn't hurt anything

 

[castbound]

U need to get a full sensor screenshot of the gpuz sensor tab under load to get a better idea

It may be GPU voltage crazy high also power draw

Sorry for getting back late, timezones I guess

Should I maybe undervolt it or something?

just lightly turning the screws to make sure they're tight shouldn't hurt anything

I tried them all and they seemed to be fine. I of course don't have a perfect screwdriver for it as it jumped out, so I can't be 100% sure.

I as well tried Sapphire Trixx fan check and it all lid up greed.

 

[oxrufiioxo]

I would just RMA it again. I'm guessing whoever you purchased this through got a bad batch. I wouldn't tinker with a card I could still return for replacement.

Your case isn't the greatest for airflow but at worst you should be seeing in the 80s unless your ambient is like 35c.

 

[Lindatje]

No undervolt is not the problem. My Sapphire RX Vega 64 nitro+ is over 300watts, even this card is has a temp of 76, standard.

The 5700XT uses 200/210 watts. I think the software is reading it wrong.

 

[castbound]

I would just RMA it again. I'm guessing whoever you purchased this through got a bad batch. I wouldn't tinker with a card I could still return for replacement.

Your case isn't the greatest for airflow but at worst you should be seeing in the 80s unless your ambient is like 35c.

I will probably do just that. Just the shipping costs are ridiculous here as I have to pay for it.
Furthermore i am seriously considering to switch the case, even though it's a lot of work again, but screw it.
Any suggestions in the same price range? At this point I don't care much about RGB, if airflow will be much better.

Maybe I'll just order the same from Amazon. Or would someone suggest some different card all together?

No undervolt is not the problem. My Sapphire RX Vega 64 nitro+ is over 300watts, even this card is has a temp of 76, standard.

The 5700XT uses 200/210 watts. I think the software is reading it wrong.

How can I verify the actual temperatures then? I mean, even AMD Software is reporting it the same.

 

[oxrufiioxo]

The msi gaming X is also a really good card but you shouldn't be having these issues with your model either as it's one of the best as well. The Red Devil is also pretty good but I shy away from recommending it because I got 2 for builds with dead ports.

My guess like others have said are either a bad sensor or poor thermal paste application your vrm and memory temps are great but your edge temps are terrible... voltage seems within normal range so you can rule out dodgy bios.

Speaking of bios have you tried the other one just to make sure you get similar readings?

 

[castbound]

The msi gaming X is also a really good card but you shouldn't be having these issues with your model either as it's one of the best as well. The Red Devil is also pretty good but I shy away from recommending it because I got 2 for builds with dead ports.

My guess like others have said are either a bad sensor or poor thermal paste application your vrm and memory temps are great but your edge temps are terrible... voltage seems within normal range so you can rule out dodgy bios.

Speaking of bios have you tried the other one just to make sure you get similar readings?

Well that means I am returning this one as well. Holy shit
It's just insane to get 2 cards like this in a row.

 

[MadDog4477]

Hi castbound

hab das gleiche System...meine GPU wird 96 Grad warm...würd gern mal mit dir reden hast du Teamspeak?

Gruß

MadDog

 

[Regeneration]

Fans speed is abnormal in the GPU-Z screenshot. 100c and only 60 percent.

Are you using MSI Afterburner? if so, set low-level hardware access interface to kernel mode.

Try to set custom fan curve.

 

[chr0nos]

looks like u got a bad card after all, all readings look "normal" , even GPU voltage is lower than mine. I find odd that your UVD clocks are not showing.

Fans speed is abnormal in the GPU-Z screenshot. 100c and only 60 percent.

Are you using MSI Afterburner? if so, set low-level hardware access interface to kernel mode.

Try to set custom fan curve.

Fan speed is fine, i get to 50% max most of the time, target temp of GPU is 80°C I think (Not Hotspot)

EDIT: It may be the BIOS is set to quiet mode

 

[MadDog4477]

Hi its my card

 

[oxrufiioxo]

Hi its my card

Your temps are a little on the high side depending on ambient but still within norms.

Your bios seems to be on the quiet settings as well.

 

[castbound]

Hi castbound

hab das gleiche System...meine GPU wird 96 Grad warm...würd gern mal mit dir reden hast du Teamspeak?

Gruß

MadDog

Hi MadDog4477, wir können sprechen, aber mein Deutsch ist noch nicht so gut. Ich lebe jetzt seit ~6 Monaten hier.

Hi its my card

Running as well close to 100? Hmm..

Fans speed is abnormal in the GPU-Z screenshot. 100c and only 60 percent.

Are you using MSI Afterburner? if so, set low-level hardware access interface to kernel mode.

Try to set custom fan curve.

Well I tried to adjust them in AMD Software. Turning it up to 100% doesn't lower temperature that much.. maybe 1-2 degrees, which is strange to me.

Furthermore, just came across this article saying 110 degrees is expected and within specs :O Was I stressing for no reason?

Your temps are a little on the high side depending on ambient but still within norms.

Your bios seems to be on the quiet settings as well.

I will check that. No clue where, but will try to google