• 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.

White Taichi 7900XTX

Joined
Feb 15, 2024
Messages
347 (0.76/day)
Location
Bulgaria
System Name Mid-Range
Processor Ryzen 7 5800X3D
Motherboard ASUS TUF Gaming B550 Plus WiFi II
Cooling AiO LS 520 White edition(MX-6 thermal paste applied)
Memory 2x16GB DDR4 3600Mhz CL18 Corsair Vengeance RGB PRO SL
Video Card(s) ASRock RX 7900XTX Taichi White
Storage Kingston FURY Renegade 2TB /TeamGroup Cardea A440 1TB/ 256GB SATA SSD / 500GB HDD
Display(s) MSI G272QPF 27 inch 2560x1440p 170hz + Technika 1080p 24'' 60hz
Case NZXT H6 Flow ARGB + 2x140mm Lian Li Uni infinity mirror
Audio Device(s) RGB Kogaion Speakers/Naraka bladepoint MR720 headset
Power Supply Lian Li EG1000 Gold White (L-shaped) 1000W
Mouse A4 Tech Bloody Naraka W95 MAX
Keyboard A4 Tech Bloody Naraka S98 Mechanical
VR HMD no
Software Windows 10 Home SL
So I got this card for a week now and it came out of the box with Performance BIOS switched on. It works great but I noticed that it jumped 7-8C on hotspot since I use it(I went from win10 to win11 too if that has something to do with this , I don't know) I saw 94-95C hotspot with 50-55C die temps in most extreme scenarios while when It was freshly bought They were like 78-82-85C hotspot at the maximum with the same operating DIE temps like it is now(specifically when playing Squad and looking at burning objects with fire - the hotspot jumps a lot) and this is on performance mode sucking 405watts on 99-100% utilization. Yes I know that 94C is completely in the safe range for these cards which are fine to operate until 110C hotspot but this temperature increase only in a week somehow bothered me. So I went to Quiet bios mode with the switch on the card and what surprised me is that nothing changed as of FPS but temepratures went 10C down and now it's max drawing 347-355Watts at max utilization and runs much more cooler. I also did a benchmark on ''Monster Hunters wild" and it gave me better result with 1k points more than when it was in Performance mode. Did I NOT won the lottery with this sillicone if it performs worse or the same on P mode while only getting worse temps with it ?
 
Sorry if this sounds harsh, but.......well DUH......

less watts = less heat !

Surely you realize that this is NOT a big secret or some hidden AMD magic, yes ?

So in YOUR particular case, with YOUR particular rig and YOUR particular "tests", you somehow managed to get similar performance with less powah....

That's usually NOT the case, but oh well, maybe you did get lucky after all...MOST people don't.. :)

Good luck either way !
 
Sorry if this sounds harsh, but.......well DUH......

less watts = less heat !

Surely you realize that this is NOT a big secret or some hidden AMD magic, yes ?

So in YOUR particular case, with YOUR particular rig and YOUR particular "tests", you somehow managed to get similar performance with less powah....

That's usually NOT the case, but oh well, maybe you did get lucky after all...MOST people don't.. :)

Good luck either way !
About the heat - sure it's not a secret or smh...Wonder why are the frames still same with 50w less, shouldn't be the opposite ? ESpecially in benchmarks?
 
About the heat - sure it's not a secret or smh...Wonder why are the frames still same with 50w less, shouldn't be the opposite ? ESpecially in benchmarks?
I believe that because of the elevated hotspot temp the GPU is not “feeling” free to sustain the highest possible frequency.

From what I’ve seen on mine past 360-370W the performance gains are negligible. I’ve test it up to 450W.
Some undervolting can help a bit but still after ~370W is not worth it considering temps and heat dissipation.
I’ve set the Vcore at 1070mV (1150 default). At some (less demanding) games can even sustain 1040~1050mV but it’s not very convenient to set it for every game every time. So I’ve settled for 1070mV.

For the record on my Nitro+ the GPU Hotspot peaks at low 70s and avgs just under 70C at 55~60% fan PWM (1900-2100rpm). Thankfully its fans are not loud at this speed. Case fans are more audible when playing games + I’m always wearing over-ear headphones.

EDIT:
Oh yes…forgot to ask.
Did your ambient temp increase during the last week?
 
Last edited:
I believe that because of the elevated hotspot temp the GPU is not “feeling” free to sustain the highest possible frequency.

From what I’ve seen on mine past 360-370W the performance gains are negligible. I’ve test it up to 450W.
Some undervolting can help a bit but still after ~370W is not worth it considering temps and heat dissipation.
I’ve set the Vcore at 1070mV (1150 default). At some (less demanding) games can even sustain 1040~1050mV but it’s not very convenient to set it for every game every time. So I’ve settled for 1070mV.

For the record on my Nitro+ the GPU Hotspot peaks at low 70s and avgs just under 70C at 55~60% fan PWM (1900-2100rpm). Thankfully its fans are not loud at this speed. Case fans are more audible when playing games + I’m always wearing over-ear headphones.

EDIT:
Oh yes…forgot to ask.
Did your ambient temp increase during the last week?
Аctually now in my country is a spring season and temperatures started to rising a little bit but not so high to make ambient room temperature and my gpu hotspot to jump that much.I mean even if it's cold/hot outside I always keep my AC turned on the opposite and keep normal ambient temperature of 20-22C during winter and 16-18C during summer .I have custom fan curve on the gpu from AMD software(it's like 30C-40% , 40C-50%,50C-60% and so on),also the NZXT H6 flow case is pretty good with solid airflow and it's big tank(for example when I remove my rear glass panel when troubleshooting my hand is starting to ''freeze'' and gets really cold from the front case fans on that case(3x120mm nzxt core fans) no matter if it's summer or winter.My case is on the table and when the glass is removed it blows directly on my right hand where the mouse is.
Does ambient temeprature has something to do with hotspot temps? Because my die temperature is the same as before, it's just the hotspot temps that rised
 
Last edited:
So I just made another benchmark on Quiet mode bios switch and it hits 3000mhz gpu with 347Watts TDP and 34439 points on ultra + frame generation enabled in 2k resolution. Tried 4 different benchmarks here are the results on Monster hunters wild 2k resolution , ultra 1440p:


34439 Quiet FSR on
37747 Quiet FSR off

34609 Performance FSR on
37747 Performance BIOS FSR OFF

They all hitting 3000mhz except the quet mode gives me 10-12C better hotspot temperatures with less power draw
 
Last edited:
So I just made another benchmark on Quiet mode bios switch and it hits 3000mhz gpu with 347Watts TDP and 34439 points on ultra + frame generation enabled in 2k resolution. Tried 4 different benchmarks here are the results on Monster hunters wild 2k resolution , ultra 1440p:


34439 Quiet FSR on
37747 Quiet FSR off

34609 Performance FSR on
37747 Performance BIOS FSR OFF

They all hitting 3000mhz except the quet mode gives me 10-12C better hotspot temperatures with less power draw
Any undervolting so far?
Even 40~50mV could make a difference while keeping the same power budget. Or with lower power for keeping same performance.

Mine too has dual bios and I keep it at the perf one.
Performance VBIOS has a range of 366~467W (-10/+15% from adrenaline). I keep it 370W currently with -80mV, 2800MHz max shader boost and a custom fan curve to my liking (noise and temp wise).

Another way to save power and heat is to insert a max FPS cap from drivers. If you’re not after highest possible FPS. Works well if your monitor supports VRR.

EDIT:
typo
 
Last edited:
Any undervolting so far?
Even 40~50mV could make a difference while keeping the same power budget. Or with lower power for keeping same performance.

Mine too has dual bios and I keep it at the perf one.
Performance VBIOS has a range of 366~467W (-10/+15% from adrenaline). I keep it 370W currently with -80mV, 2800MHz max shader boost and a custom fan curve to my liking (noise and temp wise).

Another way to save power and heat is to insert a max FPS cap from drivers. If you’re not after highest possible FPS. Works well if your monitor supports VRR.

EDIT:
typo
Honestly I don't bother to change such settings as mV as I don't understand really good how to do it properly, I only change the bios settings OR when I am on performance vbios I can touch the max power draw +/- 10 from Adrenaline software
In most demanding games I put Vsync with AMD Freesync but even If I do that it doesn't really matter because I can't reach the 165fps in Squad and HLL
 
From TPU review:
1745480351374.png


6C difference in hotspot temp between two bioses. Do you have the card physically supported in any way? It weighs almost 2 kg, could be that sagging is affecting contact.

Also, maybe it's pump out?
 
Last edited:
From TPU review:
View attachment 396600

6C difference in hotspot temp between two bioses. Do you have the card physically supported in any way? It weighs almost 2 kg, could be that sagging is affecting contact.

Also, maybe it's pump out?
Yes , this card comes with anti-sag gpu stand and I installed it - gpu sag shouldn't be a problem at all , it's brand new so it's in warranty for 3 years since 1-2 weeks ago, I don't think I should really open it for repasting even if that's the case , but I doubt it's pump out effect because I had ASRock Steel Legend RX 7700XT before this 7900xtx with exactly the same problem with 40-45C delta(except the 7700XT never reached 95C hotspot but it's normal since it's much lower TDP). I contacted the reseller that I bought the card from and they told me that if the card is open(no matter if I do it or send it for maintance to them) the warranty is voided. I can send it for RMA to ASRock and they will send me new one but again - the problem will be the same after 1-2 weeks of using it because in most of the cases it seems like that the problem is with the factory TIM or factory TIM appliance.
 

Attachments

  • 490979422_664959312918739_3638463159326434613_n.jpg
    490979422_664959312918739_3638463159326434613_n.jpg
    313.9 KB · Views: 25
  • 490983101_981445724202671_3422012545532015862_n.jpg
    490983101_981445724202671_3422012545532015862_n.jpg
    258 KB · Views: 22
Last edited:
Yes , this card comes with anti-sag gpu stand and I installed it - gpu sag shouldn't be a problem at all , it's brand new so it's in warranty for 3 years since 1-2 weeks ago, I don't think I should really open it for repasting even if that's the case , but I doubt it's pump out effect because I had ASRock Steel Legend RX 7700XT before this 7900xtx with exactly the same problem with 40-45C delta(except the 7700XT never reached 95C hotspot but it's normal since it's much lower TDP). I contacted the reseller that I bought the card from and they told me that if the card is open(no matter if I do it or send it for maintance to them) the warranty is voided. I can send it for RMA to ASRock and they will send me new one but again - the problem will be the same after 1-2 weeks of using it because in most of the cases it seems like that the problem is with the factory TIM or factory TIM appliance.
Yeah repasting isn't an option then and it's up to you if you wanna deal with the RMA. Still sucks to see that big temperature jump within a week. No wonder some manufacturers opted for PTM with 9070 series.
 
Yeah repasting isn't an option then and it's up to you if you wanna deal with the RMA. Still sucks to see that big temperature jump within a week. No wonder some manufacturers opted for PTM with 9070 series.
The problem is that with that 7700XT Steel legend again from ASRock I did voided my warranty so I can fix that hotspot and I tought that It will be just pumped out thermal compound but that wasn't the case( the paste was actually in pretty good condition when I opened it) after repasting with MX-6 and I can say that I did pretty good job on repasting back then and nothing has changed except lowering 1-2-3C on gpu temperature - hotspot remain the same... I only voided my 2 years remaining warranty for nothing...And that's why I am scared now to bother with the warranty , not because I will void the warranty but because I will void the warranty and the problem may still exists after this...except If I find some way to remove the sticker without damaging it with hair dryer and thin knife or smh... I don't really care about warranty since I know this GPU is brand new and it's not some second hand crap but only if there will be result from repasting it. After all I paid ~900$ for it and it's a lot of money in my country. The minimum monthly salary here is 600$ and if it seems a bit low it's actually enough for a single person to live good and buy everything he need if you don't pay rent, and we rarely pay rent as we own our own houses.

I actually don't want to RMA this gpu chip because it seems like I've hit the silicon lottery in some ways
 
Last edited:
Honestly I don't bother to change such settings as mV as I don't understand really good how to do it properly, I only change the bios settings OR when I am on performance vbios I can touch the max power draw +/- 10 from Adrenaline software
In most demanding games I put Vsync with AMD Freesync but even If I do that it doesn't really matter because I can't reach the 165fps in Squad and HLL
Its the most easiest process to undervolt a 7900XTX and can only do you good for thermals and heat output.
Just a couple of sliders to UV and shader speed (MHz) cap.
Not necessary to start with -100mV. Start with -30~40mV and work your way up.
Default max voltage is 1150mV.
Setting a GPU (shader) clock cap makes a bit easier to UV. Cuts off the rapid bursts to 3.0~3.2GHz which is most likely unnecessary, unless you are after benchmark numbers.
You can't do any damage while undervolting or capping speed. Extensive UV can lead only to instability because of insufficient voltage for a given speed. If unstable you give back 10~20mV and thats it.

1745492657348.png

Vsync from drivers and in game is not needed. Keep that off all times.
You only need VRR (adaptive sync) on from adrenalin display settings so your display can maintain its speed with FPS.

1745492906937.png

And DRR from windows advanced display settings.

2025-04-24 (6).png
 
If that's the case:

then I am cooked... But why this would be the problem if I had no high temperatures in the first week of using the GPU ?
 
If that's the case:

then I am cooked... But why this would be the problem if I had no high temperatures in the first week of using the GPU ?
I think that was an issue with the reference AMD, and AIB cards that copy that same design.
 
Update: Now the GPU die temperature raised at 63C from 55C max( Squad training zone with blowed up vehicles producing flames and smokes). Just to note - I don't experience such temperatures while real world gaming, only in the most GPU loading scenes.
 
Last edited:
Back
Top