• Welcome to TechPowerUp Forums, Guest! Please check out our forum guidelines for info related to our community.
  • The forums have been upgraded with support for dark mode. By default it will follow the setting on your system/browser. You may override it by scrolling to the end of the page and clicking the gears icon.

GPU Die Defects Found in PowerColor RX 9070 XT Graphics Card

Seems like a propaganda piece, what’s their sample size? Where’s the source link? (found it, why is it in tiny gray letters, in a minuscule font at the end of the page??)

We’re talking about 100ths of a mm ffs, the majority of surfaces are not flat, and the sole purpose of TIM is to account and or make up for that fact.
I mean every time I ask for those figures on the 12VHPWR connector, it didn't seem to matter. Doubt it does here, as such. /s
 
New here but use Techpowerup.com everyday. I bought a PowerColor RX 9070 XT Red Devil on launch day. I got the card installed and tested some benchmarks and noticed the card would not hit the rated speeds at all.

I then tested a few games and found that my GPU temp would get in the 80's C while my hotspot was hitting 110 C I have seen it get up to 113 114 C.

I contacted PowerColor about the issue and I'm in talks with them.

I tried changing the thermal martial, I know that use PTM7950 but I put some Kryonaut extreme and tested with the same results, I thin swapped back to my own honeywell PTM7950 just to test if maybe they used a lesser quality PTM and still got the same results.

After reading this article, Now I'm thinking I may have gotten one with this issue.

I tested this on two workstation as well both in Thermaltake tower 900 cases with custom loops on the cpu side.

1st
Asus Crosshair VIII hero wifi (x570)
Ryzen 3950x
32GB 2x16 Gskill 3600M/T system memory
Seasonic Prime TX-1000w
2x WD Black SN750 2TB
ASUS STRIX RX 5600 XT (was pulled out for testing)
Powercolor RX 9070 XT Red Devil (First time purchasing PowerColor)
2 year old windows 11 Pro install with 23h2

2nd
Asrock x870e Taichi with bios 3.20
Ryzen 9950X3D
48GB 2x24 Gskill 6400M/T system memory
Seasonic Prime TX-1300w
2x WD Black SN850x 4TB
Powercolor RX 9070 XT Red Devil (First time purchasing PowerColor)
Brand new Windows 11 Pro install with 24h2
 

Attachments

  • unnamed.jpg
    unnamed.jpg
    440.3 KB · Views: 59
  • unnamed (1).jpg
    unnamed (1).jpg
    428.2 KB · Views: 56
Did it fly off the grinder and someone put it back on? The edges of the pits look smoothed. I don't think they'd look like that if the damage happened after grinding.
 
New here but use Techpowerup.com everyday. I bought a PowerColor RX 9070 XT Red Devil on launch day. I got the card installed and tested some benchmarks and noticed the card would not hit the rated speeds at all.

I then tested a few games and found that my GPU temp would get in the 80's C while my hotspot was hitting 110 C I have seen it get up to 113 114 C.

I contacted PowerColor about the issue and I'm in talks with them.

I tried changing the thermal martial, I know that use PTM7950 but I put some Kryonaut extreme and tested with the same results, I thin swapped back to my own honeywell PTM7950 just to test if maybe they used a lesser quality PTM and still got the same results.

After reading this article, Now I'm thinking I may have gotten one with this issue.

I tested this on two workstation as well both in Thermaltake tower 900 cases with custom loops on the cpu side.

1st
Asus Crosshair VIII hero wifi (x570)
Ryzen 3950x
32GB 2x16 Gskill 3600M/T system memory
Seasonic Prime TX-1000w
2x WD Black SN750 2TB
ASUS STRIX RX 5600 XT (was pulled out for testing)
Powercolor RX 9070 XT Red Devil (First time purchasing PowerColor)
2 year old windows 11 Pro install with 23h2

2nd
Asrock x870e Taichi with bios 3.20
Ryzen 9950X3D
48GB 2x24 Gskill 6400M/T system memory
Seasonic Prime TX-1300w
2x WD Black SN850x 4TB
Powercolor RX 9070 XT Red Devil (First time purchasing PowerColor)
Brand new Windows 11 Pro install with 24h2
Some of the pits seems to be observable with the naked eye. Did you notice any inperfection on the surface of your die when you changed the PTM?
 
Some of the pits seems to be observable with the naked eye. Did you notice any inperfection on the surface of your die when you changed the PTM?
I didn't but I only gave it a glance over. I can take it back apart after work and check as I plan on trying a different Thermal paste just in case.
As far as I could tell it was smooth but i didn't touch the die with a bare finger.
 
How would you know if Nvidia removed the hotspot sensor?
Conspiracy successfully planted.
Right?!? Weird they would have done that this go'round...
 
My 9070 XT Hellhound is working soundly, doesn't get past 60c on the core and I've locked the fan speed to 48% max. Bummer for whoever had that chip in the image. It looks like it was physically damaged, don't know how that particular one could get past QC.
 
I have seen these pitting defects on several heatspreaders / CPU cooler plates so far.
And if i remember correctly even on some of my older Socket A CPUs, which were all direct die cooled by default.
 
New here but use Techpowerup.com everyday. I bought a PowerColor RX 9070 XT Red Devil on launch day. I got the card installed and tested some benchmarks and noticed the card would not hit the rated speeds at all.

I then tested a few games and found that my GPU temp would get in the 80's C while my hotspot was hitting 110 C I have seen it get up to 113 114 C.

I contacted PowerColor about the issue and I'm in talks with them.

I tried changing the thermal martial, I know that use PTM7950 but I put some Kryonaut extreme and tested with the same results, I thin swapped back to my own honeywell PTM7950 just to test if maybe they used a lesser quality PTM and still got the same results.

After reading this article, Now I'm thinking I may have gotten one with this issue.

I tested this on two workstation as well both in Thermaltake tower 900 cases with custom loops on the cpu side.

1st
Asus Crosshair VIII hero wifi (x570)
Ryzen 3950x
32GB 2x16 Gskill 3600M/T system memory
Seasonic Prime TX-1000w
2x WD Black SN750 2TB
ASUS STRIX RX 5600 XT (was pulled out for testing)
Powercolor RX 9070 XT Red Devil (First time purchasing PowerColor)
2 year old windows 11 Pro install with 23h2

2nd
Asrock x870e Taichi with bios 3.20
Ryzen 9950X3D
48GB 2x24 Gskill 6400M/T system memory
Seasonic Prime TX-1300w
2x WD Black SN850x 4TB
Powercolor RX 9070 XT Red Devil (First time purchasing PowerColor)
Brand new Windows 11 Pro install with 24h2
110 °C hotspot with 100% fan speed is definitely not right. I hope you can RMA it and report back what they said. :)
 
110 °C hotspot with 100% fan speed is definitely not right. I hope you can RMA it and report back what they said. :)
I took the GPU apart again yesterday night and check cooler and die contact and it does seem that parts of my die are not making good contact. I did notice the thermal pads on the memory seem a bit thick. I just got my Thermal putty in and will be testing this out on the memory this afternoon.
 
I didn't but I only gave it a glance over. I can take it back apart after work and check as I plan on trying a different Thermal paste just in case.
As far as I could tell it was smooth but i didn't touch the die with a bare finger.
I checked my die last night and I can see no pitting on the die but that is with out a microscope. I'm going to try a couple other things as PowerColor seems to no be responding to my emails. I got a couple with the last one asking me for pictures of my rig.
 

Attachments

  • Screenshot 2025-04-04 093120.png
    Screenshot 2025-04-04 093120.png
    61.2 KB · Views: 47
  • Screenshot 2025-04-04 093233.png
    Screenshot 2025-04-04 093233.png
    69.6 KB · Views: 46
  • Screenshot 2025-04-04 093249.png
    Screenshot 2025-04-04 093249.png
    225 KB · Views: 41
Reporting back. I swapped out the thermal pads with thermal putty and this seems to have helped some with my temp issue. My memory temp definitely has gone down to around 80-82c. I still was getting the 90-100c hotspot but fixed this by setting a 185-200fps cap with Chill and now my core and hotspot temps never go over 50c.

As soon as the Asrock 9070 xt taichi is back in stock I will be picking one up along with a alphacool core gpu block so hopefully it doesn't act this way.
 
Reporting back. I swapped out the thermal pads with thermal putty and this seems to have helped some with my temp issue. My memory temp definitely has gone down to around 80-82c. I still was getting the 90-100c hotspot but fixed this by setting a 185-200fps cap with Chill and now my core and hotspot temps never go over 50c.

As soon as the Asrock 9070 xt taichi is back in stock I will be picking one up along with a alphacool core gpu block so hopefully it doesn't act this way.
So the only reason you're not hitting 100 ˚C hotspot on the GPU is because you're not loading your card 100%? That's not good for a Red Devil. It cries for an RMA.
 
So the only reason you're not hitting 100 ˚C hotspot on the GPU is because you're not loading your card 100%? That's not good for a Red Devil. It cries for an RMA.
100% does, I haven't heard from Powercolor since March 21. so far not super happy with their support. Use to EVGA/XFX support and this is my first PowerColor product ever. Definitely leaving a bad taste. I will update if I hear from them.
 
Reporting back. I swapped out the thermal pads with thermal putty and this seems to have helped some with my temp issue. My memory temp definitely has gone down to around 80-82c. I still was getting the 90-100c hotspot but fixed this by setting a 185-200fps cap with Chill and now my core and hotspot temps never go over 50c.

As soon as the Asrock 9070 xt taichi is back in stock I will be picking one up along with a alphacool core gpu block so hopefully it doesn't act this way.
I'm thinking I'll get whatever 9070 XT I can get near the MSRP & then Alphacool waterblock it too so hopefully I won't hit this problem!
 
Update on the 9070 xt high hotspot temps. I still have not heard back from PowerColor so this will be my last product of theirs purchased. As for the issue, I swapped my GPU from my asus crosshair x570 hero with a 3950x to my Asrock x870e Taichi with my 9950x3d with the power supply also being different. Seasonic PX-1000w vs Seasonic PC-1300w. After testing the temps are way down, I even set the undervolt back to default along with all other settings. Hotspot never goes above 75c. I even can set the undervolt to -100mv and the power to +10% with a core +250 and the card will hit 3200Mhz easy with the temps never going above 56c core and 75c hotspot. what I have noticed is on the old setup I would see 3100Mhz but the power draw would hit 400W + now only seeing 290-320W. I also do not have to limit my frames as I did on the x570 with 3950x.
 
Update on the 9070 xt high hotspot temps. I still have not heard back from PowerColor so this will be my last product of theirs purchased. As for the issue, I swapped my GPU from my asus crosshair x570 hero with a 3950x to my Asrock x870e Taichi with my 9950x3d with the power supply also being different. Seasonic PX-1000w vs Seasonic PC-1300w. After testing the temps are way down, I even set the undervolt back to default along with all other settings. Hotspot never goes above 75c. I even can set the undervolt to -100mv and the power to +10% with a core +250 and the card will hit 3200Mhz easy with the temps never going above 56c core and 75c hotspot. what I have noticed is on the old setup I would see 3100Mhz but the power draw would hit 400W + now only seeing 290-320W. I also do not have to limit my frames as I did on the x570 with 3950x.
So the card is fine, there was only something wrong with your last setup? (I can't imagine what, though)
 
So the card is fine, there was only something wrong with your last setup? (I can't imagine what, though)
That is my? I put my 5600 xt back in and all is fine. That rig has windows 11 23H2 and the New one has 24H2 but from my observations its like something was sending it more power. All test where done in Nomad and Rage 2 in the same spots. Just stumped on what could have been causing it now. I always DDU drivers upon new drivers or updating hardware as well.
 
That is my? I put my 5600 xt back in and all is fine. That rig has windows 11 23H2 and the New one has 24H2 but from my observations its like something was sending it more power. All test where done in Nomad and Rage 2 in the same spots. Just stumped on what could have been causing it now. I always DDU drivers upon new drivers or updating hardware as well.
Is it in the same case? If yes, maybe a fan header didn't work on the first motherboard (maybe fans had different tunings/%). Other than pcie 4 -> pcie 5 there shouldn't be a difference. weird
 
Is it in the same case? If yes, maybe a fan header didn't work on the first motherboard (maybe fans had different tunings/%). Other than pcie 4 -> pcie 5 there shouldn't be a difference. weird
Same Case and I use fan control for all my fans. Rad fans run based on CPU and GPU temp (the old 5600 xt was watercooled but I converted it back to stock for my son) Front case fans are set to a static 55%. I'm thinking it was the power supply maybe.
 
That is my? I put my 5600 xt back in and all is fine. That rig has windows 11 23H2 and the New one has 24H2 but from my observations its like something was sending it more power. All test where done in Nomad and Rage 2 in the same spots. Just stumped on what could have been causing it now. I always DDU drivers upon new drivers or updating hardware as well.
That's weird.

Personally, I only DDU if I'm getting issues after a driver update (sometimes it introduces problems rather than solves them), but I doubt your problem is related to that.
 
That's weird.

Personally, I only DDU if I'm getting issues after a driver update (sometimes it introduces problems rather than solves them), but I doubt your problem is related to that.
I'm going to do some more testing this weekend when I get the rest of my watercooling parts in to finish the CPU loop. Just have to wait on a GPU waterblock but at least Alphacool has stated they are making a block for the red devil and other cards. May also pick up the Gigabyte aorus 9070 xt.
 
I'm going to do some more testing this weekend when I get the rest of my watercooling parts in to finish the CPU loop. Just have to wait on a GPU waterblock but at least Alphacool has stated they are making a block for the red devil and other cards. May also pick up the Gigabyte aorus 9070 xt.
As an IT support guy, I've seen inexplicable things as yours. All in all, if it's good, then it's good. Gremlins exist, I tell you.
 
As an IT support guy, I've seen inexplicable things as yours. All in all, if it's good, then it's good. Gremlins exist, I tell you.
Same here, Been in IT for the better part of 20 years. I have been building rig since the Althon 64 single jingle days. This is the first time I have seen this but I have seen some odd ball things in my time.

I did play several games last night to test and most all of them at 1440p max settings I'm seeing 170-260 fps on this rig and Rage 2 (I know its older) was getting close to 300 and after 2 hours the temps on the hotspot never got over 75, I also noticed the fans are not ramping up like there were and the system is much much quitter because of this.
 
Some gpu's don't like being vertical like in your Tower 900, was the hotspot high in vertical orientation? As a guy not in IT and building PC's since Pentium days, that's the first thing I would test before taking a card apart just to get my warranty denied.
 
Last edited:
Back
Top