The Radeon RX 7900 XT(X) Owners Thread.

thefogo · 30 Dec 2022 at 16:31

Trypta said:
What stable underclocks are people getting? I seem to be getting driver crashes if I go below 1100mV

Mine crashes eventually at anything under 1135mv so I’ve just left my voltage at default.

Gibbo · 30 Dec 2022 at 17:17

Can a Nitro owner test their card both horizontal and vertical please to see if the vapor x chamber works as well in any position.

LtMatt · 30 Dec 2022 at 17:22

Gibbo said:
Can a Nitro owner test their card both horizontal and vertical please to see if the vapor x chamber works as well in any position.

Yes, I'll test that now and put a YouTube video up of both horizontal and vertical positions. Will post it here and tag you back once done.

Trypta · 30 Dec 2022 at 17:24

Gibbo said:
Can a Nitro owner test their card both horizontal and vertical please to see if the vapor x chamber works as well in any position.

Sure, I'll run a stress test now while I make dinner.

Edit:
Although, my sapphire MBA will hit 110c in any orientation within 10 mins even without high power limits, delta is around 50c worse case scenarios almost 60c. Any powerlimit above +5% just worsens this gap and the rate it hits those temperatures.
Noticeably, junction temp will jump from 110c to down to 55c almost instantly when the test ends. It doesn't gradually lower which is odd.

So far the highest I've seen the nitro+ hit is 91c and tends to stay there even with with +15% on the power limit.

So far with my case on its side I'm looking at temps of 71c and 91c on the nitro+, with stock fan settings. Will keep it going to see if that changes though. In comparison the largest delta I've seen is 20c on the nitro compared to 59c and above on the mba.

Vimes · 30 Dec 2022 at 17:38

@LtMatt - you have noted the wattage that your new Nitro can pull, not sure if that was at max...? But not the 483w mentioned somewhere else in this forum that might be possible with a Nitro. Compared to the TUF you had how much of a difference is that making in terms of performance and is that max performance compromised at all due to to bigger and better (?) cooler on the TUF meaning it runs hooter or louder with the Nitro....?

Thanks.

Trypta · 30 Dec 2022 at 17:39

Precaution be careful when touching a glass side panel when testing horizontally, the glass reached a temperature I wasn't expecting lol

delta0 · 30 Dec 2022 at 17:51

Vimes said:
@LtMatt - you have noted the wattage that your new Nitro can pull, not sure if that was at max...? But not the 483w mentioned somewhere else in this forum that might be possible with a Nitro. Compared to the TUF you had how much of a difference is that making in terms of performance and is that max performance compromised at all due to to bigger and better (?) cooler on the TUF meaning it runs hooter or louder with the Nitro....?

Thanks.

There’s also a YouTube video of the nitro hitting 500w. https://youtu.be/XaEWPgPDF7U

Trypta · 30 Dec 2022 at 17:54

@Gibbo nitro+ 72c/91c in both orientations after 20 loops of speedway.

LtMatt · 30 Dec 2022 at 18:09

Vimes said:
@LtMatt - you have noted the wattage that your new Nitro can pull, not sure if that was at max...? But not the 483w mentioned somewhere else in this forum that might be possible with a Nitro. Compared to the TUF you had how much of a difference is that making in terms of performance and is that max performance compromised at all due to to bigger and better (?) cooler on the TUF meaning it runs hooter or louder with the Nitro....?

Thanks.

delta0 said:
There’s also a YouTube video of the nitro hitting 500w. https://youtu.be/XaEWPgPDF7U

The Nitro pulls 465W total board power according to HWINFO64 with +15% set. . If you measure actual power draw like Kitguru did, I suspect it will be higher and closer to the 483W-500W value that was quoted. The TUF pulls 431W total board power in comparison with +15%. The performance difference can be seen in my Port Royal Ray Tracing tests for the Nitro here and the TUF here. The performance difference is about 3%, so not huge. The TUF has the better cooler in terms of temperatures, its huge. Both are good though, and both quiet. TUF louder at 100% fan speed, think the fans move faster and move more air.

LtMatt · 30 Dec 2022 at 18:10

Trypta said:
@Gibbo nitro+ 72c/91c in both orientations after 20 loops of speedway.

What clocks, power limits and fan speeds were you using for this? I will copy and run Speedway too for a comparison but it will take me some time as I'll capture it on video.

Gibbo · 30 Dec 2022 at 18:10

Trypta said:
@Gibbo nitro+ 72c/91c in both orientations after 20 loops of speedway.

That’s good so Sapphire does not have the issue MBA cards have then, good to know.

Mr Evil · 30 Dec 2022 at 18:12

Gibbo said:
Can a Nitro owner test their card both horizontal and vertical please to see if the vapor x chamber works as well in any position.

I'd be interested to see this tested for all the custom cards, in all six possible orientations. Especially the one orientation that the MBA can't handle (monitor outputs at the top).

Trypta · 30 Dec 2022 at 18:25

LtMatt said:
What clocks, power limits and fan speeds were you using for this? I will copy and run Speedway too for a comparison but it will take me some time as I'll capture it on video.

Core:
Min Freq: 500 Mhz
Max Freq: 3400 Mhz
1100mV

Vram:
Fast timings
2648 Mhz

Power limit 15%
Stock fan speeds

Vimes · 30 Dec 2022 at 18:26

@LtMatt thanks for that, you linked the Time Spy results for the Nitro, not the Port Royal

LtMatt · 30 Dec 2022 at 18:28

Trypta said:
Core:
Min Freq: 500 Mhz
Max Freq: 3400 Mhz
1100mV

Vram:
Fast timings
2648 Mhz

Power limit 15%
Stock fan speeds

Thanks.

Vimes said:
@LtMatt thanks for that, you linked the Time Spy results for the Nitro, not the Port Royal

Fixed.

Trypta · 30 Dec 2022 at 18:56

LtMatt said:
I've managed to surpass my stock 4090 graphics scores in Timespy with the Nitro. @Gibbo

I scored 32 410 in Time Spy

AMD Ryzen 9 7950X, AMD Radeon RX 7900 XTX x 1, 32768 MB, 64-bit Windows 11}

www.3dmark.com

How on earth did you manage that score and those temps by the way?

LtMatt · 30 Dec 2022 at 19:00

Trypta said:
How on earth did you manage that score and those temps by the way?

23c ambient when I did that too Lol. Was running an undervolt at 1.010mv, 100% fan speed. I'd put liquid metal on the GPU core too, so that may have helped reduce temps a little bit from stock.

The Nitro is still power limited. Give me more power and I will score more. So even the highest available XTX at this time is still held back by power limit.

Trypta · 30 Dec 2022 at 19:21

LtMatt said:
23c ambient when I did that too Lol. Was running an undervolt at 1.010mv, 100% fan speed. I'd put liquid metal on the GPU core too, so that may have helped reduce temps a little bit from stock.

The Nitro is still power limited. Give me more power and I will score more. So even the highest available XTX at this time is still held back by power limit.

1010mv?
Is there a list you need to be on to get good samples like that?

LtMatt · 30 Dec 2022 at 19:28

Trypta said:
1010mv?
Is there a list you need to be on to get good samples like that?

That is not stable in games, need way higher. But for synthetics it works. Firestrike Ultra needed 1.020v. Games need at least 1.060mv, and I think I might even need higher than that to be truly stable most likely 1070mv-1.080mv.

Been busy at work this week, but will get some proper time on the weekend to play some games and put some YouTube videos up.

Noxia · 30 Dec 2022 at 19:32

LtMatt said:
That is not stable in games, need way higher. But for synthetics it works. Firestrike Ultra needed 1.020v. Games need at least 1.060mv, and I think I might even need higher than that to be truly stable most likely 1070mv-1.080mv.

Been busy at work this week, but will get some proper time on the weekend to play some games and put some YouTube videos up.

How many PCs do you have? I'm sure you had a 4090