RX 7900 XTX reference at possible vapor chamber design problem + very high hot spot (110c)

[eidairaman1]

Look 1 Person Provided a Helpful Link. If you have a problem Contact the Support Department.

AMD Responds To RX 7900 XTX 110C Hotspot Issues

Contact support.

www.tomshardware.com

Thanks @LFaWolf for providing the Important info needed to help users resolve this problem

I vote the thread be locked

 

[OneMoar]

agree lock this thread most of the people commenting are going in circles and not reading though it and thus repeating the same crap

 

[shovenose]

You got an MBA 7900XT? Cooler looks a bit different, one screw hole less around GPU.

Yes I do

 

[Steevo]

A lot of cry babies in this thread that may or may not have the hardware. Also, someone calling something a personal attack when it’s clearly not.

On topic, looks like the cooler migrates some of the fluid out of circulation when moved around, strange that W1zz didn’t see the high temps with horizontal mounting (based on the case) but maybe he is using a riser cable and vertical position.

The silly guy in the video should have turned it horizontal with the fans up, gravity would have pushed the liquid onto the area where the dies are and temps would be lower again I bet.

 

[Vlooi]

I observed a very interesting thing that happened yesterday out of the blue, and re-tested it again this morning.

A few notes first:

1. Ambient temps about the same.
2. Settings on Adrenaline the same.
3. Same games (Cyberpunk, HZD, Witcher3 Next Gen). Raytracing on for Cyberpunk and Witcher.

My 7900XTX maxed at 100 degrees hotspot after more than an hour gameplay, core at 67, fans RPM at 2900. HDR on in Windows 11 as always.

At some stage I switched HDR off in Windows, and immediately went back into the games to test for more than an hour again. My hotspot did not go above 86 degrees celcius, core still at 67 degrees max, fans now at average 1920 RPM. Same scenes, etc.

This morning re-activated HDR, went back in same games, same scenes, same settings in game and Adrenaline, and hotspot maxed at 86 degrees, core still at 67. Same when I de-activated HDR.

Additional info:

1. GPU Fans average 1920 rpm.
2. Core speed averages above 2700 Mhz.
3. GPU utilization 100%.
4. GPU board power maxed at 373W during gameplay.
5. VRam at 2535.

Could this be related to the displayport, maybe the cable, maybe a combination? Its just too coincidental. Why would my hotspot drop average 14 DEGREES CELCIUS because I flipped a switch that relates to display and the cable? Did some component "unload" during the on and off switching of HDR? The card's temps is now rock stable.

May there be some odd thing with the new displayport 2.1?

Thought I would just throw some thoughts and finding around on the forum, because this to me is very interesting. Is there a problematic BIOS setting in some cards related to this, influencing the algorithm?

Regards

 

[TumbleGeorge]

I observed a very interesting thing that happened yesterday out of the blue, and re-tested it again this morning.

A few notes first:

1. Ambient temps about the same.
2. Settings on Adrenaline the same.
3. Same games (Cyberpunk, HZD, Witcher3 Next Gen). Raytracing on for Cyberpunk and Witcher.

My 7900XTX maxed at 100 degrees hotspot after more than an hour gameplay, core at 67, fans RPM at 2900. HDR on in Windows 11 as always.

At some stage I switched HDR off in Windows, and immediately went back into the games to test for more than an hour again. My hotspot did not go above 86 degrees celcius, core still at 67 degrees max, fans now at average 1920 RPM. Same scenes, etc.

This morning re-activated HDR, went back in same games, same scenes, same settings in game and Adrenaline, and hotspot maxed at 86 degrees, core still at 67. Same when I de-activated HDR.

Additional info:

1. GPU Fans average 1920 rpm.
2. Core speed averages above 2700 Mhz.
3. GPU utilization 100%.
4. GPU board power maxed at 373W during gameplay.
5. VRam at 2535.

Could this be related to the displayport, maybe the cable, maybe a combination? Its just too coincidental. Why would my hotspot drop average 14 DEGREES CELCIUS because I flipped a switch that relates to display and the cable? Did some component "unload" during the on and off switching of HDR? The card's temps is now rock stable.

May there be some odd thing with the new displayport 2.1?

Thought I would just throw some thoughts and finding around on the forum, because this to me is very interesting. Is there a problematic BIOS setting in some cards related to this, influencing the algorithm?

Regards

Still looks like a driver issue. Apparently the driver fails to detect the load correctly initially. After the manipulation of the button, it apparently leads to reloading of the driver or part of it and it finally "bites", how to work properly.

 

[nguyen]

Well it's winter in the northern hemisphere, AMD would need to solve the issue before summer, or just stop making MBA cards and let AIBs charge whatever they want.

 

[Vlooi]

Well it's winter in the northern hemisphere, AMD would need to solve the issue before summer, or just stop making MBA cards and let AIBs charge whatever they want.

For interest sake I live in very hot part of the world, north-eastern South Africa, right next to the Kruger Park, no air-conditioning inhouse. Thus with the temps I am getting now with the reference cooler you guys in the North should be A-OK in summer. (Ambient in summer varies between 29 to 35 celcius inhouse, outside up to 45 celcius with high humidity). I am actually quite surprised at how well the cooler works under these circumstances.

 

[Theswweet]

I observed a very interesting thing that happened yesterday out of the blue, and re-tested it again this morning.

A few notes first:

1. Ambient temps about the same.
2. Settings on Adrenaline the same.
3. Same games (Cyberpunk, HZD, Witcher3 Next Gen). Raytracing on for Cyberpunk and Witcher.

My 7900XTX maxed at 100 degrees hotspot after more than an hour gameplay, core at 67, fans RPM at 2900. HDR on in Windows 11 as always.

At some stage I switched HDR off in Windows, and immediately went back into the games to test for more than an hour again. My hotspot did not go above 86 degrees celcius, core still at 67 degrees max, fans now at average 1920 RPM. Same scenes, etc.

This morning re-activated HDR, went back in same games, same scenes, same settings in game and Adrenaline, and hotspot maxed at 86 degrees, core still at 67. Same when I de-activated HDR.

Additional info:

1. GPU Fans average 1920 rpm.
2. Core speed averages above 2700 Mhz.
3. GPU utilization 100%.
4. GPU board power maxed at 373W during gameplay.
5. VRam at 2535.

Could this be related to the displayport, maybe the cable, maybe a combination? Its just too coincidental. Why would my hotspot drop average 14 DEGREES CELCIUS because I flipped a switch that relates to display and the cable? Did some component "unload" during the on and off switching of HDR? The card's temps is now rock stable.

May there be some odd thing with the new displayport 2.1?

Thought I would just throw some thoughts and finding around on the forum, because this to me is very interesting. Is there a problematic BIOS setting in some cards related to this, influencing the algorithm?

Regards

While there definitely appears to be hardware issues on at least some units, there's definitely driver issues as well; games will just randomly not perform as well out of the blue and require a restart sometimes. Beyond the multi-monitor idle power draw, we've had folks like LTT notice performance differences just depending on the monitor you used... here's hoping we quickly get some driver improvements this month.

 

[JustBenching]

The RTX 4090 model also encountered this burning power input problem. I don't know if they fixed it.

Τhere was no burning problem. Users didnt plug the cable properly. It's a user error, nothing to do with nvidia

 

[Chomiq]

I never had any issues with my reference 6900XT, but I'm definitely glad I went with a Red Devil for my 7900XTX.

Have you looked into RMA options from Powercolor? Because in Poland they don't do any RMA services for end users, they tell them to go back to the store.

 

[Bomby569]

Have looked into RMA options from Powercolor? Because in Poland they don't do any RMA services for end users, they tell them to go back to the store.

don't ever RMA in Europe stuff inside the warranty, you will lose all your legislation protection, they will no longer have to answer you in 30 days, they can do whatever they want. That only applies to retailers. Trust me i felt for this once. That's for the AmeriNoRights people

 

[wolf]

I vote keep the thread going, for the most part it's an interesting discussion that can evolve as the situation does.

I do believe in coming days more reputable testing will be done and shed more light on it, and technical discussion should always be welcome on the forum. See a shitpost? Report it.

 

[Falck]

Just tried the HDR on/off method as @Vlooi experienced, no luck

Tried ONLY HDMI connected, no luck.

What helped was tilting the case before HS gets heated.

 

[Vya Domus]

My 3090RTX is colder when consuming 470W.

I think you might need one or two courses in physics, temperature =/= hot or cold. Your almost 500W RTX 3090 is in no way or shape "cold" buddy.

You got an MBA 7900XT? Cooler looks a bit different, one screw hole less around GPU.

Yeah the shape of the vapor chamber looks different too, so it's probably not impacted (though, if it is, maybe not enough folks have bought one to notice?

I said this in another thread but if the vapor chamber really was the problem then the entire GPU would run hotter, it wouldn't affect just the hot spot temperatures. Meaning there wouldn't be a 30C+ gap between those temperatures, which is the actual problem here.

 

[Theswweet]

Have you looked into RMA options from Powercolor? Because in Poland they don't do any RMA services for end users, they tell them to go back to the store.

Ok? I got the protection plan for my card at Microcenter, so it won't be in my system any longer than 2 years regardless.

 

[Vlooi]

I think you might need one or two courses in physics, temperature =/= hot or cold. Your almost 500W RTX 3090 is in no way or shape "cold" buddy.




I said this in another thread but if the vapor chamber really was the problem then the entire GPU would run hotter, it wouldn't affect just the hot spot temperatures. Meaning there wouldn't be a 30C+ gap between those temperatures, which is the actual problem here.

Well stated. One tends to lose perspective sometimes, and don't look at things more broadly. Sounds very logical.

Just tried the HDR on/off method as @Vlooi experienced, no luck

Tried ONLY HDMI connected, no luck.

What helped was tilting the case before HS gets heated.

Falck,

I have an aftermarket DP cable, I don't use the one that came with the monitor. Have you tried another cable? It may be simple suggestion, but who knows? I do take note of the tilting of the case which decreases your temps though.

 

[Vayra86]

Does anyone work for free for public benefit by his own choice?

Many do.

to be fair that was a new product where the company doing the standard is also at fault, this is a vapor chamber, this is old tech, this should not have happened.

There isn't a difference here.

Nvidia used an old principle of dividing power across strands of copper/metal to direct power to the GPU, and this doesn't work proper. They deployed an adapter (simple stuff) to cover the transition from old to new. The standard isn't at fault. The adapter Nvidia built, is.

AMD is using a new technology too, they're stacking chips beside the GPU. They've applied an old principle of a vapor chamber (simple stuff) over it and this doesn't work proper.

In both cases its the application of existing technologies and techniques, the quality of implementation is shite in both camps. Both approaches scream cost reduction over common sense and QC; and if you compare the 12VHPWR spec to the pcie spec, this also screams cost reduction over foolproof and safety. Tolerances are lower, but we said 'this is fine'.

And now they both burn for it. Well played, I hope the reality check holds for longer than this gen.

 

[AusWolf]

I don't think there should be any facts that should be ignored. The fact that companies react when there are irregularities is not by their will, but by force of the circumstances, and there are probably laws that they have to comply with.

They could just issue a statement saying that "it's within spec", as turbo clocks are always an "up to" value. Let's see if that happens.

 

[Falck]

All i know is if AMD, or suppliers, don't take responsibility I'm most likely to returning the card. I have a 60-day return window so I have until Feb. 15. to make up my mind.

 

[AusWolf]

I don't see anything new, I have tried 4 6800XTs from different manufacturers and 2 6900XTs, all of them were trash. The hotspots were always reaching 100C on them, having my case side panel off/on, no change.

Winter here and having 20C+ in the room, I don't know the exact number. So during summer, when 35C is nothing unusual here, I would be reaching 110C+ on all of them and would throttle.

AMD GPUS are trash. My 3090RTX is colder when consuming 470W. Pathetic. This has been the continuous problem since 5700XT.

Are AMD GPUs trash because they give you actual hotspot temps instead of estimates (yes, up until Ampere, hotspot is just a software estimate on GeForce cards), or is it because they don't run your games?

Guess what, my reference 6750 XT runs with a 105-107 °C hotspot, but achieves 99.7% in the 3DMark Time Spy stress test every single time.

Guys, stop judging a graphics card by its temperature alone! It's not 2008 anymore.

All i know is if AMD, or suppliers, don't take responsibility I'm most likely to returning the card. I have a 60-day return window so I have until Feb. 15. to make up my mind.

Have you tried a 3DMark stress test yet?

 

[delshay]

I said this in another thread but if the vapor chamber really was the problem then the entire GPU would run hotter, it wouldn't affect just the hot spot temperatures. Meaning there wouldn't be a 30C+ gap between those temperatures, which is the actual problem here.

That's a very valid point. It could be an component that's related to the thermal read-out. .Let me watch that video again.

 

[Vya Domus]

Just because something is consuming more power, doesn't mean, it is automatically warmer, you know there are more things into, aka the die area, cooling solution etc.

Except that in this case they are, ICs convert over >99% of power into heat, the die sizes are comparable, which means that yes, they're just as warm. If something has the same power output, same area, same heat, it will also have the same thermal density, irrespective of what a temperature sensor reports, which might I add, Nvidia and AMD are at complete liberty to choose what exactly their driver reports and how that is calculated since there is no correct or incorrect way to calculate temperature in an IC, so 100C in an Nvidia card might and most likely means something completly differnt from what 100C means in an AMD card. Long story short, that number means absolutely nothing with regards to how cold or not a GPU really is. Of course someone with a non-empty head would be well aware of that.

Stop being in denial about 8th grade physics.

 

[AusWolf]

I observed a very interesting thing that happened yesterday out of the blue, and re-tested it again this morning.

A few notes first:

1. Ambient temps about the same.
2. Settings on Adrenaline the same.
3. Same games (Cyberpunk, HZD, Witcher3 Next Gen). Raytracing on for Cyberpunk and Witcher.

My 7900XTX maxed at 100 degrees hotspot after more than an hour gameplay, core at 67, fans RPM at 2900. HDR on in Windows 11 as always.

At some stage I switched HDR off in Windows, and immediately went back into the games to test for more than an hour again. My hotspot did not go above 86 degrees celcius, core still at 67 degrees max, fans now at average 1920 RPM. Same scenes, etc.

This morning re-activated HDR, went back in same games, same scenes, same settings in game and Adrenaline, and hotspot maxed at 86 degrees, core still at 67. Same when I de-activated HDR.

Additional info:

1. GPU Fans average 1920 rpm.
2. Core speed averages above 2700 Mhz.
3. GPU utilization 100%.
4. GPU board power maxed at 373W during gameplay.
5. VRam at 2535.

Could this be related to the displayport, maybe the cable, maybe a combination? Its just too coincidental. Why would my hotspot drop average 14 DEGREES CELCIUS because I flipped a switch that relates to display and the cable? Did some component "unload" during the on and off switching of HDR? The card's temps is now rock stable.

May there be some odd thing with the new displayport 2.1?

Thought I would just throw some thoughts and finding around on the forum, because this to me is very interesting. Is there a problematic BIOS setting in some cards related to this, influencing the algorithm?

Regards

Very interesting!

Could hotspot be measured at the GPU's display engine when it runs hot, instead of its parts responsible for 3D?

Except that in this case they are, ICs convert over >99% of power into heat, the die sizes are comparable, which means that yes, they're just as warm. Same power output, same area, same heat, irrespective of what a temperature sensor reports, which might I add, Nvidia and AMD are at complete liberty to choose what exactly their driver reports and how that is calculated since there is no correct or incorrect way to calculate temperature in an IC. Long story short, that number means absolutely nothing with regards to how cold or not a GPU really is. Of course someone with a non-empty head would be well aware of that.

Stop being in denial about 8th grade physics.

Not to mention that how hot a chip is designed to run isn't related to its use value. Someone try running a Pentium 3 at temperatures modern CPUs run at and see what happens!

 

[SaLaDiN666]

Are AMD GPUs trash because they give you actual hotspot temps instead of estimates (yes, up until Ampere, hotspot is just a software estimate on GeForce cards), or is it because they don't run your games?

Guess what, my reference 6750 XT runs with a 105-107 °C hotspot, but achieves 99.7% in the 3DMark Time Spy stress test every single time.

Guys, stop judging a graphics card by its temperature alone! It's not 2008 anymore.


Have you tried a 3DMark stress test yet?

I will tell you something shocking, the hotspot is not actually the hottest part of the chip, because it is also an estimate according to the simulations so the sensor is placed there. The entire die is not fully covered by temp measuring sensors so if someone, for example, manages to leave the part of the die without paste or not touching the cooler, it can simply get damaged by high temps depending on the sensor grid and solution.

Secondly, the safe operating temperatures of GPUs are specifically made with hotspots in mind and local temps up to 120C are safe depending on the process, architecture, area, PCB, material and longevity of the exposure to such temp. It was AMD's decision to OC them out of the box and so much they are risking damaging the die if the exposure to such temps would be prolonged so that's why they actually measure the hotspot, it is not because they like you or technically superior.