[kfxsti]

Probably a dumb question. But I got my Sapphire Nitro+ 7900xtx installed a few days ago, and I am running the Ryzen 5900x CPU and only 2 nvme drives . Have not had any issues in Warzone 2.0 at all. But on a whim I decided to check the PSU reqs and saw the 1000w psu req being tossed around online. Am I safe with my Cosair RM850x ? or should I pull the trigger on a 1000w psu?

 

[kfxsti]

3440x1440p with frames capped at 165fps. And I just let the card do its thing lol

 

[kfxsti]

You're fine as long as the RM850x is good quality. No idea but in terms of raw power (70a on a 12v rail?) you're fine.

The transients are typically (from software) around 550w but according to Gamers Nexus' they can go up to 700+w but hard to recreate. Most instances of transients were below 700w. That's when a poor PSU will likely trip.

Spoiler: Source

I never had an issue with my XFX Merc 6900xt and same CPU and PSU, so hopefully I should be fine . "Hopefully" lol

 

[kfxsti]

Awesome, Thanks everyone !

 

[kfxsti]

Finally got the time to toss my Alphacool Eiswolf Block AIO on my Nitro +. Temps have not made it over 40c after playing a few hours of Warzone 2 . Haven't gotten to the Overclocking and benching yet, but its coming lol

 

[kfxsti]

So after doing some tinkering. I have noticed that any adjustments made besides undervolting will cause my hotspot temp to shoot up to around 80c but my core temp is showing 40-43c (running superposition). Under normal gaming the core does it's hovering around 40-45c but the hotspot is between 59-70c.
I'm wondering if I need to pull the waterblock back off and do a repaste. The block came with Alphacool Apex as the TIM. And damn that was seriously thick paste to try and spread around.
I have a fresh tube of Kryo Extreme I think I'm going to crack open and try on it.
But I'm also wondering if having this thing mounted vertically could be part of it.

 

[kfxsti]

Hello guys and gals!
Long time lurker here.
YesterdayI upgraded my gaming rig from a Vega 64 to a 7900XT! Between 2.5 (Marvel’s Guardians of the Galaxy) and 4 (Kena Bridge of Spirits) times faster than the old card.
Here it is!
View attachment 2618700
As you can see, it’s a pretty cramped case, and my loop is very old-school (I love this style!).
I didn’t have any time for some good tests, but my card is pretty hot running, at least compared to my Vega. 47°C core and 65ºC hotspot, when my Vega was 40 and 48ºC max…
Could it be the a bad application of the Alphacool Subzero that came with the WB? As @kfxsti said, that’s some seriously thick paste. Or two 240mm rads are too small? Or my pump is weak? Or it’s just that 5nm is hot because it’s hugely denser than the old 14nm of my Vega?
Not that I’m complaining too much, since it’s at least 20ºC better than the air-cooler reference.

And yes, in the future I’m upgrading the case to something bigger, since it’s a PITA to work in it sometimes (and thus switching to a tube pump/res combo with a D5 pump).


P.S. My reference still had the protective film on the main VRM thermal pad
View attachment 2618701

I ordered a fresh tube of Subzero. Just to rule out it being thickly old LOL. I'll let you know how it works out in the coming days.

 

[kfxsti]

Wellp. I was going to put off doing the re-pasting. But definitely will be making that happen hopefully tomorrow. A 2 hour gaming session of Warzone 2 --

 

[kfxsti]

Welp.. I tried a fresh tube of Subzero, and a fresh tube of Kryonaut Extreme .
Tried pea size method, X method, spread method.
Gpu temp MAX has been 52c and hotspot 74c. I guess I'm just going to have to settle for 74c as a hotspot. Lol

 

[kfxsti]

Nearly god as Liqid Metal.
View attachment 2620261

I wonder if this would help my hotspot temp. And I wonder how it would do under my Arctic Liquid 420 AIO on my 7900x lol
My reg temp is great. Its the hotspot that likes to get a bit fancy . lol * Edit *
Did a bit of research on this, and it looks to have helped a few peoples hotspot temps on different cards.
@RedF so 25x25 is the size needed for a 7900xtx ?

 

[kfxsti]

Hey all. I have a few questions about my Red Devil 7900 XTX. I have finally gotten around to playing with undervolting this thing after tearing it down cleaning and applying some thermalgrizzly Phasesheet PTM (I have normal PTM7950 too , just wanted to try this out lol)
Here are my current settings

(Hopefully this image scales correctly 😬)
After about 4 hours of some Warzone-Resurgence on Rebirth Island. Here is my peak temps, core clock, and the item that concerns me the Power W which is 577 😬

Is this normal behavior for undervolting ? For the Power W to be that high ? I know there are spikes and it's not a constant 577w , 300-380w is what I was seeing on average. But jeez 577w lol
Just wanting to make sure I'm not harming my card in anyway , if that makes sense ?
I'm loving my temps on the card now though. And the GPU fans are not bothersome noise wise with my fan curve in the Flux Pro case. I added two fans in the very bottom and have 3 under the GPU.
3440x1440 @240hz. I can't remember the in-game settings but I have the frames capped at 240fps in Adrenalin.
Sorry for the long post , just kinda shocked the crap out of me tonight with that massive power draw lol

 

[kfxsti]

Should be fine mate. Looks like it's likely just a temporary power spike and/or rebound effect from instances quickly transitioning from very high --> very low power requirements. Given the majority of your power use is in stock ranges I don't have any concerns.

I can't comment on the accuracy of Adrenaline but even so I'd recommend using HWiNFO 64 to monitor metrics as you're able to view a wider range of metrics in each category - im this case it would show items like total board power, graphics power, etc. Able to get a lot more insight into what's happening.

I'm no expert so take my advice with a grain of salt. If your PSU isn't sufficient, faulty, or poor quality then that could also be the cause. Poor regulation or not enough capacity can lead to spiking/dipping.

Got it , and thanks for the help of easing my mind and worry lol
My current PSU is the Thermaltake Toughpower GF A3 Snow Edition 1050W 80+ Gold. Had a chance to grab it at a great price so I got it when I was having a issue with Warzone shutting my rig off randomly mid games . I got tired of troubleshooting the issue, so jumped ship from a 850w PSU, x670e Mobo and 7950x3d to the 1050w , x870e and 9900x3d. And have had 0 power off issues.
I'll give hwinfo a look over tomorrow. MSI Afterburner is what I have been using to track the speeds, temps, voltages etc
Thanks again 😁

 

[kfxsti]

No problems mate happy to help!

I had another look after my original comment and your temps are well controlled, plus the timing of the power spikes correlates with big frequency changes so everything checks out that I can see. Can't see voltage in there but I'd wager it also lines up.

Your PSU is definitely more than fine. I'm running a 80+ gold 1000W Corsair rmx with a 12600k.

The only other thing I could think of which might cause this would be a faulty motherboard or pcie slot - but I reckon you'd have other symptoms if that were the case.

Highly recommend HWiNFO, you can monitor graphically with it if you're that way inclined. I'd put money on you not going back after using it haha.

Enjoy the rig!

Wellp. Installed the latest GPU and Chipset driver and things have gotten weird. Even after a DDU on the GPU driver and Revo on the Chipset, and rolling back to the ones I was using previously - my card is now showing a max Power W of 380w~ and will not boost above 3024mhz. But doesn't break 52c in warzone LOL 🤦‍♂️
I'll mess with it more in the morning lol

 

[kfxsti]

Anything less than 1080mV will cause driver timeouts for me. I have just been increasing the core clock before each session to see what is stable. Warzone and Warframe are the only two games I play right now. And warzone being the more taxing of the two. I will pull the voltage back up to 1100mv to see what happens in a few matches of Warzone. As for the PSU its 3 individual 8 pins from the PSU to the GPU . I dunno, maybe i just had an issue to begin with-with the power spikes being so high. And maybe the fresh driver installs fixed it? Maybe i dont need to boost beyond 3100 in warzone to maintain my 240fps in warzone since clean chipset/gpu driver installs. I dunno, I will tinker around a bit more with it later today.
Thanks Alastair and Phantom for the help and guidance . I will report back once i have done some more testing

 

[kfxsti]

Thats the thing, im not crashing or anything at 1080m @ 3070mhz . Im just no longer boosting above 3100 mhz. The more i sit here and think about it still, maybe it was a an issue to begin with. Hell , even my temperatures on my GPU are lower in the game now. I just checked and have a System restore point at the time before installing the new drivers . I wonder what will happen if i roll back to that time lol
Update: Ok that didn't work either with the system restore. Lol I think I'm just going to leave it be

 

[kfxsti]

You are not going to believe this ... I have been uninstalling ,reinstalling, playing with the voltage, core clock.. nothing doing anything . And then I find this . 🤦‍♂️

Evidently I must have hit the switch sometime yesterday when I was doing my weekly dusting lol

 

[kfxsti]

Good luck mate!

I think now that I have everything back to normal. I may wait until the next WHQL Driver is released for the GPU. I hate how they have tried to bundle the download of the chipset driver in with it all now too with that extra software .