• Welcome to TechPowerUp Forums, Guest! Please check out our forum guidelines for info related to our community.

AMD Ryzen 9 7950X Cooling Requirements & Thermal Throttling

Are you seriously suggesting that Tjmax doesn't apply when part of the heat is coming from an outside source while the component is under full load?

I’m just saying that the clickbait video has nothing to do with concerns about silicon degradation over normal usage , unless you live in an room where ambient temperature is around 70/80° C.
 
I’m just saying that the clickbait video has nothing to do with concerns about silicon degradation over normal usage , unless you live in an room where ambient temperature is around 70/80° C.
Not even then, because the chip decreases its frequency and voltage until it's safe within the temperature limit, so you'll have less current going through it (current was your concern, wasn't it?). If it can't anymore because it's basically overheating while idling, then it'll shut the PC off. TLDR: The more heat it gets from outside sources, the less heat it will generate internally to compensate.

You clearly didn't bother watching the video, yet you keep arguing, which is sad. Very sad.
 
Do you really think you’re funny ?

I'm sorry but I couldn't resist. The temptation was just too great.

Snag_5e130fa.png
 
Look to me, to get the temperatures down from the 95 degrees C, you need to implement a refrigerated CPU cooler! I cannot afford the cost!
 
Look to me, to get the temperatures down from the 95 degrees C, you need to implement a refrigerated CPU cooler! I cannot afford the cost!
If you even read just the conclusion never mind the whole OP, you would know that's not necessarily the case but you do you.
 
Look to me, to get the temperatures down from the 95 degrees C, you need to implement a refrigerated CPU cooler! I cannot afford the cost!
No, you just need to adjust the temp limit in the UEFI wherever you want.
 
Yes they do. This could risk law suit from EU, under planned obsolescence, which is illegal.
Apple and Samsung have got hit by fines regarding down-tuning their phones (for alleged battery performance), AMD could risk the same by having "planned degradation"

and no, saying boost up till, and failing that is also a 9law suit risk under false marketing/advertisment
They have to gurantee it'll be stable and reliable


It's not like the CPU will magically decide "Oops i need to drop 200MHz today, this core isn't doing so great"
If that tech existed, we'd never have overclocking instability


They're saying if it loads to 95C, it'll be running and functional for the duration of the 3 year warranty and hopefully beyond at stock settings (including stock boost), and if it not - it wasnt the heat that killed it.

No, you just need to adjust the temp limit in the UEFI wherever you want.
Which no one talks about, despite being an easily accesible BIOS option

Changing the limit to 80C may reduce MT performance, but not any more than a weaker cooler throttling would (And same as on both intel and AMD, tweaking any power settings can fix this)
 
No, you just need to adjust the temp limit in the UEFI wherever you want.
Or if you're too lazy to do that (like I am), then just enable Eco mode.
 
They have to gurantee it'll be stable and reliable


It's not like the CPU will magically decide "Oops i need to drop 200MHz today, this core isn't doing so great"
If that tech existed, we'd never have overclocking instability


They're saying if it loads to 95C, it'll be running and functional for the duration of the 3 year warranty and hopefully beyond at stock settings (including stock boost), and if it not - it wasnt the heat that killed it.
what, no? In the US maybe.

if it has the alleged "planned degradation", AMD will put themselves at risk of being sued in EU court. It could potentially go under planned obsolescence, and/or false marketing.
I'm not saying it is guaranteed, but they do put themselves at risk for it (under the presumption that normal usage will degrade, as a "planned degradation")
These are not dependent on the warranty. planned obsolescence and false marketing is illegal in EU, as mentioned both Samsung and Apple (iPhone) have lost for this, after warranty period.
Warranty Period doesn't mean shit for that.


iPhone 6 (2014 phone), class action 2021, lotsa year after it was released.
 
what, no? In the US maybe.

if it has the alleged "planned degradation", AMD will put themselves at risk of being sued in EU court. It could potentially go under planned obsolescence, and/or false marketing.
I'm not saying it is guaranteed, but they do put themselves at risk for it (under the presumption that normal usage will degrade, as a "planned degradation")
These are not dependent on the warranty. planned obsolescence and false marketing is illegal in EU, as mentioned both Samsung and Apple (iPhone) have lost for this, after warranty period.
Warranty Period doesn't mean shit for that.


iPhone 6 (2014 phone), class action 2021, lotsa year after it was released.
That's true, but you can't prove that your CPU has degraded. Maybe it maxes out at 5.1 GHz instead of 5.2, but the guaranteed clock speed is 4.5 GHz, so it still runs above spec, technically.
 
That's true, but you can't prove that your CPU has degraded. Maybe it maxes out at 5.1 GHz instead of 5.2, but the guaranteed clock speed is 4.5 GHz, so it still runs above spec, technically.
eh, they advertise with boost up till 5.7 GHz (based on applications yaddayadda), so they kinda need to prove it can do that, if it falls bellow, you have a possible case (surprised 3000-series didn't get lawsuited tbh)
sure, losing 0.1Ghz is nothing, and I'd argue that's not degradation (at least within "planned degradation"), it's if it's within many hundreds MHz, at that point you'd probably have a case.
 
eh, they advertise with boost up till 5.7 GHz (based on applications yaddayadda), so they kinda need to prove it can do that, if it falls bellow, you have a possible case (surprised 3000-series didn't get lawsuited tbh)
sure, losing 0.1Ghz is nothing, and I'd argue that's not degradation (at least within "planned degradation"), it's if it's within many hundreds MHz, at that point you'd probably have a case.
"Up to" means nothing. "Under certain conditions" can mean conditions that are impossible to meet, like LN cooling, or being ejected into outer space or something. You can't prove that your CPU has degraded unless it can't even maintain base clock anymore with default settings and adequate cooling. Otherwise, AMD's lawyers can easily blame it on your use case or the silicon lottery.
 
And I think 5600 doing 70°C in games and hovers around 85°C while rendering in conjunction with PBO is alarming, I knew it's pretty safe because it's still under TJMax. If I upgrade to Zen 4, I have to forget that as this was a new normal, there is no need to worry about temperature.
 
if it has the alleged "planned degradation", AMD will put themselves at risk of being sued in EU court. It could potentially go under planned obsolescence, and/or false marketing.

Did you know any CPU, and even at stock operation, will eventually degrade?

Degradation.png


Heat, power, voltage, all plays a role in accelerating this proces. At 95 degrees they tested it to be safe. And i dont think any of us here is taxing such a CPU for 24/7 rendering or something simular, while constantly running at 95 degrees. It has a max of 115 degrees but AMD is warning that degradation for long term periods on that temperature can or will cause degradation.

I think after a few years these CPU's will still function perfectly as intended. Even after warranty. However Bullzoid already blew up a 7950x for running it at 1.41V or so.
 
However Bullzoid already blew up a 7950x for running it at 1.41V or so

What's the max voltage AMD claims to be safe for full load tasks?
 
eh, they advertise with boost up till 5.7 GHz (based on applications yaddayadda), so they kinda need to prove it can do that, if it falls bellow, you have a possible case (surprised 3000-series didn't get lawsuited tbh)
sure, losing 0.1Ghz is nothing, and I'd argue that's not degradation (at least within "planned degradation"), it's if it's within many hundreds MHz, at that point you'd probably have a case.
You either didn't read what i wrote, or you mis interpreted all of it.


1. "up to" with boost
2. There is NOTHING in the hardware that allows them to monitor "Degradation" so this concept of lowering clocks over time is utterly impossible.
 
You either didn't read what i wrote, or you mis interpreted all of it.


1. "up to" with boost
2. There is NOTHING in the hardware that allows them to monitor "Degradation" so this concept of lowering clocks over time is utterly impossible.

There are safe limits setup in the CPU; it's called FIS and it has bin in the first generation of Ryzen already. It basicly monitors the CPU's workload, temperature, current consumption and stuff and sticks to a (maximum) predefined boost setting or algorithm. Since AMD proberly pretested the CPU's theirself they can pretty much determine which is the upmost maximum the silicon in all conditions could handle for at least the warranty period.

I'm running on a 2700X and it's boosting happy over 4.5Ghz single core and 4.2Ghz all core. No issues till this day really. CPU operates as designed. And there's no need for manual all core OC'ing really.
 
There are safe limits setup in the CPU; it's called FIS and it has bin in the first generation of Ryzen already. It basicly monitors the CPU's workload, temperature, current consumption and stuff and sticks to a (maximum) predefined boost setting or algorithm. Since AMD proberly pretested the CPU's theirself they can pretty much determine which is the upmost maximum the silicon in all conditions could handle for at least the warranty period.

I'm running on a 2700X and it's boosting happy over 4.5Ghz single core and 4.2Ghz all core. No issues till this day really. CPU operates as designed. And there's no need for manual all core OC'ing really.
That is not the same thing
All that information is hardcoded, and relies on sensor readings

Degradation cannot be measured, there is no sensors for it - therefore it cannot ever happen
 
That is not the same thing
All that information is hardcoded, and relies on sensor readings

Degradation cannot be measured, there is no sensors for it - therefore it cannot ever happen

Nobody ever said that, right? the sensors and hardcoded values are in place to protect the silicon's life. There's a guy who set the AMD FX world record, who wrote a complete essay on the FIS part of the chip. It is there in place to provide the most optimal values and right out of the box "free" performance.

Intel does the same. The boosting is a safe feature and it extracts the best performance out of the box with no interaction on your behalf. It can do that up to 50 times a second. I've seen people do OC's on their 2700x, ramp up the core voltage from stock to 1.4 ~ 1.45V and have their chip degraded in weeks to months.

The degradation was it could not hold clocks and in worst case not even hold boost anymore. These nodes are'nt like the FX era, where stock 1.3V chips could eat 1.65V all day and night.
 
Did you know any CPU, and even at stock operation, will eventually degrade?

View attachment 265187

Heat, power, voltage, all plays a role in accelerating this proces. At 95 degrees they tested it to be safe. And i dont think any of us here is taxing such a CPU for 24/7 rendering or something simular, while constantly running at 95 degrees. It has a max of 115 degrees but AMD is warning that degradation for long term periods on that temperature can or will cause degradation.

I think after a few years these CPU's will still function perfectly as intended. Even after warranty. However Bullzoid already blew up a 7950x for running it at 1.41V or so.
yes?
You're kinda ignoring the "PLANNED DEGRADATION" part, the one close to, or is planned obsolescence.
Of course degradation of tools over time (including PC parts) is natural, no one expects them to last forever.
BUT planned obsolescence isn't the same as normal wear and tear.

The user that this was a response to alleged AMD is doing a planned degradation/obsolescence, by having the CPU running up to 95c "all the time".
It's not guaranteed, but if that is the true, and it's a major degradation after the warranty period (it doesn't matter necessarily matter here, basing on the mentioned lawsuits), AMD could possibly get sued, and pay fines.

You either didn't read what i wrote, or you mis interpreted all of it.


1. "up to" with boost
2. There is NOTHING in the hardware that allows them to monitor "Degradation" so this concept of lowering clocks over time is utterly impossible.
And you seems to have done the same.

1.
During the early 3000-series, was many who claimed their CPU's couldn't reach the advertised boost. Referencing among other, this: https://wccftech.com/der8auers-ryzen-3000-series-boost-survey-reveals-worse-than-expected-boosting/

Advertisements which mislead or which may mislead the people who receive them are forbidden.
It is desirable that misleading commercial practices cover those practices, including misleading advertising, which by deceiving the consumer prevent him from making an informed and thus efficient choice.

You have the right to truthful advertising

the new CPU not reaching boost could have been a court case in EU.

2.

Where the f does minor degradation comes from?
There is a reason I included "Planned obsolescence" in every post here. I had hoped it would imply MAJOR degradation.
A user earlier in the thread alleged that the 95c limit by AMD was bad, and that it would degrade the CPU after warranty period.
IF it is a planned degradation (outside of normal wear and tear), in the form of planned obsolescence, then that is illegal (in EU), and they put themselves at risk of getting sued (see, the keyword "at risk"), that risk is still there after warranty period.

Which among others, this is:
(e) the need for a service, part, replacement or repair;
Under the Directive you have a minimum 2-year legal guarantee against faulty products, or products that do not look or work as advertised. These are known as your statutory rights. The national law in some Member States may allow for longer periods. In Ireland the time limit (also known as the limitation period) is 6 years.

Manufacturer is not protected from widespread failure/obsolescence AFTER warranty, as shown in the iPhone 7 link. Where Apple was sued after warranty period.
 
BUT planned obsolescence isn't the same as normal wear and tear.
But since you can't predict how these products will be used in each & very system out there, alongside cooling, & of course if the user OC it himself ~ so planning to make it obsolete by AMD themselves outside warranty would be akin to putting your two feet & both arms on a chainsaw! Why would they even do that, they got burned with FX "8 core" lawsuit do you really they'd try an even more brain dead move?
 
They settled it likely because they didn't want to risk it going to trial in a court, they had a strong case IMO but the litigation would be needless bad PR at the time!
 
Back
Top