Linux 6.13 To Allow Controlling Zero RPM Feature For Radeon RX 7000 Series GPUs

[Anux]

Actually I hate zero rpm and don't see any point of it. You always want at least some flow to get the hotter air out of the PC chassis as soon as possible, even if it is very low rpm. On my 7900 XTX GPU, even without the aftermarket fans I have on them now, sub 800ish rpm is totally inaudible

That depends on the fans you use and how much power your system needs in idle mode. I also prefer to run my fans at 900 RPM on idle just because they are inaudible and it gives a little heat buffer so they don't spin up with short usage spikes.
But if you have low quality fans they might even make noises in their lowest RPM setting. Or if you have a low power system that can run fan less most of the time with fans just kicking in for long full load scenarios or on hot days.

[Ferrum Master]

That depends on the fans you use and how much power your system needs in idle mode. I also prefer to run my fans at 900 RPM on idle just because they are inaudible and it gives a little heat buffer so they don't spin up with short usage spikes.
But if you have low quality fans they might even make noises in their lowest RPM setting. Or if you have a low power system that can run fan less most of the time with fans just kicking in for long full load scenarios or on hot days.

Nope. The thermal performance for air coolers are very similar, look at the benchmarks, the deviation are few dBs, you may argue about main dominant harmonics, but not today. The thing that matters is how many monitors you have hooked to the card and at what refresh rate those panels are, I have 3 and it eats 50W while in idle... and actually windows has better numbers in that regard ~15-20W. But I suppose it will never be fixed.

But just as another bloke here, I haven't ran GPU on air for decades, it is wasted performance, massively more noise and better longevity due to better thermals across the board surrounding components.

[Anux]

The thermal performance for air coolers are very similar

What no? There couldn't be more difference, every fan has different characteristics and of course different heat pipe configs, materials, fin distances do heavily influence the cooling performance.

look at the benchmarks, the deviation are few dBs

Which benchmarks?

you may argue about main dominant harmonics, but not today.

I'm talking about bearing and motor noises. Typical fan (wind) noise is hardly a problem on low RPMs.

The thing that matters is how many monitors you have hooked to the card and at what refresh rate those panels are

This just influences your heat output and might prevent you from shutting down fans but you are using a high end video card that sucks > 300W on load, which is a choice you made. And since you're using a water pump you are somewhat immune to low background noise.
Other people have much higher standards what they consider as being silent and it also matters how loud your environment is.

[Ferrum Master]

What no? There couldn't be more difference, every fan has different characteristics and of course different heat pipe configs, materials, fin distances do heavily influence the.

It is the same thermal envelope, same size and most importantly same price or you think the OEMs are doing charity there? It is real life. They perform very similarly, we could argue how many care for memory and VRM cooling, but the core is same as otherwise it would look bad in gaming benchmarks and it would drop out and if you look on sites like Techpowerup for same card reviews on the GPU release day... they perform the same per clock, no one drops out in performance, the industry knows what they do. It is all bullshit, and RGB bling that matters for some in the looks, otherwise they perform the same, some variation is introduced with your ASIC quality also as you need more voltage to remain stable, but that's matter of luck, but 7900XTX specific if you have it, don't know I have one, the most important issue is high consumption doing multiple displays and that is it, cooling idle 5W or 50W like I do is the difference. Don't be one of those marketing victims, just my five cents.

[Anux]

if you look on sites like Techpowerup for same card reviews on the GPU release day... they perform the same per clock

That is because all cooling solutions are sufficient, it doesn't mean they all perform the same. Since you don't provide a link to a benchmark that would prove your point I just took a random one: https://www.techpowerup.com/review/g...ing-oc/39.html
There you can see temp differences of around 10°C, noise differences of 10 db (which is huge) and different RPMs. How would you explain those if all coolers are the same? Look even further to the OC-section and you will see that one card can dissipate 100 W more than another. And of course the coolers differ in size with bigger ones performing better most of the time.

cooling idle 5W or 50W like I do is the difference.

Even 50W should be easy to cool passive with such a big metal chunk. And also remember there are more cards that just your 7900.

Don't be one of those marketing victims, just my five cents.

I'm using an 8 year old GPU that I bought used for less than 200 € 7 years ago and you use an overpriced 300W monster that seams to be buggy. Who is the marketing victim here?

[Ferrum Master]

That is because all cooling solutions are sufficient

You understand you just talk rubbish and contradict yourself? Spare me and others from it.

Products are designed to perform the same, you pick a product X and expect it to work like any other, and it actually is so, otherwise you RMA the thing. We don't care for for any DIY design made in your shed and personal doings here, you are free to do it. We have a reference design and everyone aligns to it, that is how industry works.

I cannot even understand the rant about using 8 year old gpu, we need to give you a medal for it?

[Anux]

You understand you just talk rubbish and contradict yourself?

I posted a link that proves what I said you just blabbered about magical benchmarks that you can't link here because they don't exist.

Products are designed to perform the same

What is wrong with you, even a child can see that this is just plain wrong.

you pick a product X and expect it to work like any other

No.

otherwise you RMA the thing.

"Dear Sapphire Team, the cooler on your card is better then reference, please send me a card with a worse cooler."
Yeah sure, that would totally happen.

I cannot even understand the rant about using 8 year old gpu, we need to give you a medal for it?

I couldn't care less, you accused me of being a victim of marketing, I just showed you that the fool is you.
And of course you don't understand it, you can't even look at benchmarks and make logical conclusions from the data.