RX 9000 series GPU Owners Club

[Kwadratowicz]

Thanks!
I'd still like to see those in maxed out 4K for a complete GPU comparison.

Unfortunately my display is max at 3440x1440, and If You could add the settings tab?

 

[AVATARAT]

Sorry for the stitched shots. Win10 applied some compatibility settings, resulting in messed up UI. It would seem that the 7900xtx does better in legacy APIs. Of course, this finding is mostly of academic value with the number of fps either GPU churns out.

May I still ask for those feature tests in 06? Pretty please?

Okay, here they are
And yes, the 7900 XTX has more hardware inside, so where it can work, it works.

I scored 0 in 3DMark06 Custom

AMD Ryzen 7 9700X, AMD Radeon RX 9070 XT x 1, 65536 MB, 64-bit Windows 11}

www.3dmark.com

Lol, that's interesting, did you see it in 3DMark 2001?
Didn't expect it, but it looks like the High Details are faster on the 9070 XT

06, why it looks a lot better than newest benchmarks? new benchmarks are like soulless and surgically crafted, not a speck of dust there, not like in old ones .

Yeah, did you see the mirrors and the trees?
No RT, but it looks very nice... and the trees, look far better with all those details.

 

[Kwadratowicz]

Yeah, did you see the mirrors and the trees?
No RT, but it looks very nice... and the trees, look far better with all those details.

I was looking in 1280x1024, and it was still better than 4k new benchmarks
Today I will see them in 2.4 k
And the game was fun ^^

 

[NSR]

Anyone play with their GPU set to the lowest power limit?

Recently have mine at -30% which is 170W. 9060 XT is suspected to be 180W for reference.

I lose about 15% performance going from Stock TDP (240) to -30% (170). Still plenty left to work with.

But doesn't seem worthwhile since upscaling reduces power draw/temps also.

 

[Kwadratowicz]

Anyone play with their GPU set to the lowest power limit?

The lowest I could get was 55W, but only in linux it's possible .
(everything lower than 55W is using 55W anyway).

 

[NSR]

The lowest I could get was 55W, but only in linux it's possible .
(everything lower than 55W is using 55W anyway).

Oh cool, what was the performance at that point? I'd imagine it'd be around a RX 7600 unless it's just way too limited.

 

[Kwadratowicz]

Oh cool, what was the performance at that point? I'd imagine it'd be around a RX 7600 unless it's just way too limited.

Dunno, but I'm waiting to change some pc parts, and after that I will dive in with benchmarks, and other stuff, so I will then add strange benchmarks , with black myth wukong at different wattage
But black Myth wukong was around 8 fps, with all on full, and no upscale i think without rt too.

Haha it's working

40 fps is from menu, so 8 is from benchmark.

 

[NSR]

Dunno, but I'm waiting to change some pc parts, and after that I will dive in with benchmarks, and other stuff, so I will then add strange benchmarks , with black myth wukong at different wattage
But black Myth wukong was around 8 fps, with all on full, and no upscale i think without rt too.

40 fps is from menu, so 8 is from benchmark.

This makes me miss the RDNA 2/3 GPU tuning.

We could limit GPU frequency to 500MHz which in turn limited max power usage. It was fun to play with.

 

[QuietBob]

If You could add the settings tab?

Sure!

I scored 0 in 3DMark06 Custom

AMD Ryzen 7 5800X3D, AMD Radeon RX 7900 XTX x 1, 32768 MB, 64-bit Windows 10}

www.3dmark.com

Lol, that's interesting, did you see it in 3DMark 2001? Didn't expect it, but it looks like the High Details are faster on the 9070 XT

Yeah, I analyzed every single result

In 3DMark01SE game tests 1 to 3 utilize the fixed function T&L rendering pipeline of dx7 for the most part, so they benefit from a faster CPU. In Car Chase with high details the 9070xt is 15% faster, but the 7900xtx is 71% faster in Dragothic (low detail) and 24% faster in Lobby (low detail). They're matched in other iterations of these tests, but I'm pretty sure the 7900xtx would come out on top with something like an oc'd 14900KS.

This makes me miss the RDNA 2/3 GPU tuning. We could limit GPU frequency to 500MHz which in turn limited max power usage. It was fun to play with.

Those are my daily settings. With all sliders at minimum the 7900xtx is effectively capped to 140 W. It performs similarly to the rx6600, which is enough for old and indie games. The clock will still boost to over 1200 MHz in less demanding titles/scenes, but the consumption won't increase.

 

[elano68]

Hi guys.
In my country;

Gigabyte Gaming oc model is 865 Dollars in 3 installments without interest.

The Taichi model is 1020 Dollars in 6 installments without interest on Amazon.

The Nitro model is 1000 Dollars in 6 installments without interest on Amazon

I am a gamer who does uv/oc and uses the card to its fullest extent. I have no aesthetic expectations such as rgb or appearance. My only expectation is performance. When viewed from this perspective, does the gigabyte gaming oc model provide the same performance or oc performance as the others?
Or should I prefer nitro or taichi?

By the way, the nitro models here come with Samsung VRAM. Taichis come with Hynix.

Samsung VRAMs work cooler as far as I can see. However, it is said that Hynix is better in terms of OC performance. I would be very happy if you could comment on this situation.

Thanks in advance for the answers.

 

[NSR]

Sure!

I scored 0 in 3DMark06 Custom

AMD Ryzen 7 5800X3D, AMD Radeon RX 7900 XTX x 1, 32768 MB, 64-bit Windows 10}

www.3dmark.com

Yeah, I analyzed every single result

In 3DMark01SE game tests 1 to 3 utilize the fixed function T&L rendering pipeline of dx7 for the most part, so they benefit from a faster CPU. In Car Chase with high details the 9070xt is 15% faster, but the 7900xtx is 71% faster in Dragothic (low detail) and 24% faster in Lobby (low detail). They're matched in other iterations of these tests, but I'm pretty sure the 7900xtx would come out on top with something like an oc'd 14900KS.


View attachment 400889
Those are my daily settings. With all sliders at minimum the 7900xtx is effectively capped to 140 W. It performs similarly to the rx6600, which is enough for old and indie games. The clock will still boost to over 1200 MHz in less demanding titles/scenes, but the consumption won't increase.

Nice, I did the same as at the time I was only playing low demand games. Was using less than 100W on my 6800. It was a great feature, no reason to use power I don't need.

Did they explain exactly why it was changed?

 

[AVATARAT]

Did they explain exactly why it was changed?

AMD probably did this to make better market segmentation of the cards, since locking down all hardware is good for their revenue and bad for us consumers.

 

[Kwadratowicz]

Hi guys.
In my country;

Gigabyte Gaming oc model is 865 Dollars in 3 installments without interest.

The Taichi model is 1020 Dollars in 6 installments without interest on Amazon.

The Nitro model is 1000 Dollars in 6 installments without interest on Amazon

I am a gamer who does uv/oc and uses the card to its fullest extent. I have no aesthetic expectations such as rgb or appearance. My only expectation is performance. When viewed from this perspective, does the gigabyte gaming oc model provide the same performance or oc performance as the others?
Or should I prefer nitro or taichi?

By the way, the nitro models here come with Samsung VRAM. Taichis come with Hynix.

Samsung VRAMs work cooler as far as I can see. However, it is said that Hynix is better in terms of OC performance. I would be very happy if you could comment on this situation.

Thanks in advance for the answers.

Cause of the unclear details about the gigabyte and thermal gel problem I would for now did not buy gigabyte card (especially that gigabyte was trying to scrape it under the rug).
I've got one, and it is great card, but I don't know if it's a ticking bomb, or good component.

About Taichi and Nitro

ASRock Radeon RX 9070 XT Taichi OC Review - Excellent Cooling

The ASRock Radeon RX 9070 XT Taichi features the most powerful cooler of all the models that we've tested so far, resulting in a whisper-quiet gaming experience. It also comes with the highest power limit and biggest factory overclock, which makes it the fastest RX 9070 XT.

www.techpowerup.com

Sapphire Radeon RX 9070 XT Nitro+ Review - Beating NVIDIA

AMD's new RDNA 4 GPUs are launching today. The Sapphire Radeon RX 9070 XT Nitro+ is a fantastic custom design with powerful cooling and whisper-quiet fans. In terms of FPS, the RX 9070 XT offers competitive performance with similarly priced NVIDIA options in both raster and ray tracing.

www.techpowerup.com

TPU has great reviews of this two cards (unfortunately it won't be apples to apples with Your place, cause both of them has sk hynix in review).

 

[elano68]

Cause of the unclear details about the gigabyte and thermal gel problem I would for now did not buy gigabyte card (especially that gigabyte was trying to scrape it under the rug).
I've got one, and it is great card, but I don't know if it's a ticking bomb, or good component.

About Taichi and Nitro

ASRock Radeon RX 9070 XT Taichi OC Review - Excellent Cooling

The ASRock Radeon RX 9070 XT Taichi features the most powerful cooler of all the models that we've tested so far, resulting in a whisper-quiet gaming experience. It also comes with the highest power limit and biggest factory overclock, which makes it the fastest RX 9070 XT.

www.techpowerup.com

Sapphire Radeon RX 9070 XT Nitro+ Review - Beating NVIDIA

AMD's new RDNA 4 GPUs are launching today. The Sapphire Radeon RX 9070 XT Nitro+ is a fantastic custom design with powerful cooling and whisper-quiet fans. In terms of FPS, the RX 9070 XT offers competitive performance with similarly priced NVIDIA options in both raster and ray tracing.

www.techpowerup.com

TPU has great reviews of this two cards (unfortunately it won't be apples to apples with Your place, cause both of them has sk hynix in review).

Thanks for your answer. You have a 9070xt gaming oc model, did I understand correctly?

Frankly, I didn't worry much about the gel status since I will be using the card horizontally.

What is your experience with using Siin? Have you done a steelnomad test with stock and ubv/oclu?

Besides, which model do you think is better, taichi or nitro? It has been said that taichi is better as a PCB, but practical performance is important to me. I have also watched most of the videos and reviews.

 

[Kwadratowicz]

What is your experience with using Siin? Have you done a steelnomad test with stock and ubv/oclu?

Could You elaborate Siin and ubv/oclu? I'm a little green to overclocking .
About the steel nomad, I was doing test with stock, then -80uV but only speedway with 2800 Mhz, I will test later it again, and post the results here.
Around 8 PM of CEST.
And yes it's Gigabyte gaming oc 9070 xt.

 

[elano68]

Could You elaborate Siin and ubv/oclu? I'm a little green to overclocking .
About the steel nomad, I was doing test with stock, then -80uV but only speedway with 2800 Mhz, I will test later it again, and post the results here.
Around 8 PM of CEST.
And yes it's Gigabyte gaming oc 9070 xt.

Sorry I wrote it wrong. What was your steel nomad score with 9070 xt gaming oc? I would appreciate it if you could post the UV/OC result as well. Thank you.

 

[Kwadratowicz]

Sorry I wrote it wrong. What was your steel nomad score with 9070 xt gaming oc? I would appreciate it if you could post the UV/OC result as well. Thank you.

Steel nomad:

Edit: OVER 8000
Fan set to 75% and finally I was able to break through 8000 mark .

I scored 8 002 in Steel Nomad

AMD Ryzen 7 5700X, AMD Radeon RX 9070 XT x 1, 16384 MB, 64-bit Windows 10}

www.3dmark.com

Is it enough

 

[elano68]

Steel nomad:

Edit: OVER 8000
Fan set to 75% and finally I was able to break through 8000 mark .

I scored 8 002 in Steel Nomad

AMD Ryzen 7 5700X, AMD Radeon RX 9070 XT x 1, 16384 MB, 64-bit Windows 10}

www.3dmark.com

View attachment 400981

View attachment 400982
Is it enough

Thank you very much my friend. What you added was quite sufficient. You got incredible points. Congratulations

You came across a very good silicon. By the way, I also bought the 9070xt nitro with a momentary campaign. It is almost the same price as the Gaming OC. I hope I will be satisfied.

 

[Kwadratowicz]

I've taken card from system for inspection, and for now it looks like it's ok, maybe radeon is not hot enough to melt the thermal gel.


It's a little different angle, but the thermal gel does not touch vapor chamber in both cases .
exactly month between pictures.
(I hoped that it was not the case, and I could mod my gpu , cause 2x120 fits

)

Thank you very much my friend. What you added was quite sufficient. You got incredible points. Congratulations

Now I need just check which is stable for gaming , but that is for another day

 

[iSpeakVeryWell]

Steel nomad:

Edit: OVER 8000
Fan set to 75% and finally I was able to break through 8000 mark .

I scored 8 002 in Steel Nomad

AMD Ryzen 7 5700X, AMD Radeon RX 9070 XT x 1, 16384 MB, 64-bit Windows 10}

www.3dmark.com

View attachment 400981

View attachment 400982
Is it enough

That's crazy! The fact that you did 8000 with only -85mV undervolt blow my mind!

I did 7950 with -100mV! Either something's wrong with my windows installation, or you used a secret sauce!

 

[QuietBob]

Did they explain exactly why it was changed?

I don't think there's ever been an official statement. Perhaps they concluded that no one in their right mind would want to restrict their GPU to 40% TBP
On a more serious note, removing voltage adjustment from the control panel makes it more foolproof and safer for the average user. Fewer RMAs due to inept overclocking and less bad publicity for the manufacturer.

 

[Kwadratowicz]

That's crazy! The fact that you did 8000 with only -85mV undervolt blow my mind!

I did 7950 with -100mV! Either something's wrong with my windows installation, or you used a secret sauce!

I've waited a liitle to cool down the gpu, and then blasted it with best settings from benchmarks
Just that, the gpu was cool in the beginning of the run, has made difference (7990 was with hot gpu, and 8002 was with cold gpu at start , and vram was overclocked to 2830, and pl was 110 in both cases).
Hot gpu had 2805 MHz max in vram, but cold gpu had 2815 MHz max in vram.
run 8 vs run 25, look at the gpuz
(the runs are numbered, and pictures are correlated )
Edit, -90 is unstable even in steel nomad ^^

 

[Braegnok]

I've waited a liitle to cool down the gpu, and then blasted it with best settings from benchmarks
Just that, the gpu was cool in the beginning of the run, has made difference (7990 was with hot gpu, and 8002 was with cold gpu at start , and vram was overclocked to 2830, and pl was 110 in both cases).
Hot gpu had 2805 MHz max in vram, but cold gpu had 2815 MHz max in vram.
run 8 vs run 25, look at the gpuz
(the runs are numbered, and pictures are correlated )
Edit, -90 is unstable even in steel nomad ^^

8002 awesome score

5635 seems to be my max for VRAM, or 2817 running Steel Nomad,.. my points will start to drop running above 5635,..

 

[Solaris17]

What a day, indeed. In my neighborhood, we spent 11 hours without GSM or mobile data, and 12 hours and 30 minutes without power. Luckily, nothing serious happened on my end, hope everything's good on your side too.
---
@Camper7, @Solaris17 and @Hogar you've been added to the list! Feel free to share any additional details or corrections whenever possible, and let me know if you'd like to be removed.
@cadaveca I updated your card, but still there are some missing details, same comments as above apply

Keep them coming!!

sorry for getting back so late. Its samsung memory. I was doing other things with this board before I could get the card on it. Now I can start playing in earnest.

 

[iSpeakVeryWell]

I'm I the only one who is Steel Nomad score goes up after a driver crash? I normally reboot immediately after a driver crash, but yesterday I wanted to test my settings was correct before rebooting and the run I did bested my previous best score. I was really confused and the next run I did was even better.

Based on the screenshot below, driver crashed during the run following the one that gave me a score of 7966. The next 3 runs 8008, 8044 and 8041 was after driver crashed. I'm only able to pass 8000 if driver crashed first! That's so weird!

This was on Win11 Pro. I thought something was wrong with my Windows installation so I reinstalled using Windows 11 LTSC and I was immediately able to reproduce. It seems it doesn't happen every times tho. I'd say it happened more often than not.

What is your thought on this?