Ryzen 5800 owners complain about very high MT load temps

[birdie]

Source 1.
Source 2.
Source 3.

Many have resorted to enabling Eco Mode in Ryzen Master which basically turns the 5800X into a yet to be announced 5700X. From the Ryzen Master manual:

Applying Eco-Mode lowers the processor’s power consumption from default stock to AMD’s lower, standard AM4 infrastructure power level (TDP).
‒ 105W and 95W TDP models shift to 65W
‒ 65W TDP models shift to 45W
• The processor continues to manage core voltage and frequency automatically to the reduced power levels - expect frequencies may be lower
• PBO, AOC and Manual overclocking are not supported when in Eco-Mode
• Expect to see lower peak frequencies from the cores under heavy multi-threaded workloads, though at lower power consumption.
• Eco-Mode may be useful when trading off Ryzen processor power for more graphics card power at the system level.

A number of people disagreed with me when I said that AMD basically OC'ed Ryzen 5000 CPUs as much as possible right out of the gate to gain that coveted single threaded performance lead and it looks like it's exactly what people are dealing with now. Luckily enabling Eco Mode translates into a ~8% performance loss in MT Cinebench R20 scores which shouldn't affect games but it's still not pleasant for those who have shelled out $450. But here's a problem: the stock Ryzen 3700X scores around 4750 points while TDP limited 5800X scores around 5350 which is a 12% performance increase while costing 36% more.

 

[Zach_01]

That guy on reddit said that he was hitting 90C on R20MT and he re-seated his cooler and got -5C. But he didn't mention his cooler.
Sounds like aggressive boosting just like 3000 with not(maybe) good cooling. And the spikes from idle 30C to 50/60s or even 70s is same as 3000 depending cooling again and windows power plan. Its the hotspot sensor (Tctl/Tdie) reading that only ZEN2/3 reveal.

Is there any other examples?

 

[birdie]

That guy on reddit said that he was hitting 90C on R20MT and he re-seated his cooler and got -5C. But he didn't mention his cooler.
Sounds like aggressive boosting just like 3000 with not(maybe) good cooling. And the spikes from idle 30C to 50/60s or even 70s is same as 3000 depending cooling again and windows power plan. Its the hotspot sensor (Tctl/Tdie) reading that only ZEN2/3 reveal.

Is there any other examples?

Anandtech review? The same 90C:

Also of note are the last two processors – both processors are reporting 4450 MHz all-core turbo frequency, however the 5800X is doing it with 14.55 W per core, but the 5600X can do it with only 10.20 W per core. In this instance, this seems that the voltage of the 5800X is a lot higher than the other processors, and this is forcing higher thermals – we were measuring 90ºC at full load after 30 seconds (compared to 73ºC on the 5600X or 64ºC on the 5950X), which might be stunting the frequency here. The motherboard might be over-egging the voltage a little here, going way above what is actually required for the core.

 

[LFaWolf]

That guy on reddit said that he was hitting 90C on R20MT and he re-seated his cooler and got -5C. But he didn't mention his cooler.
Sounds like aggressive boosting just like 3000 with not(maybe) good cooling. And the spikes from idle 30C to 50/60s or even 70s is same as 3000 depending cooling again and windows power plan. Its the hotspot sensor (Tctl/Tdie) reading that only ZEN2/3 reveal.

Is there any other examples?

He did - be Quiet dark rock on b550 MSI mortar.

 

[Zach_01]

And I forgot to mention the Power Reporting Deviation matter, where a board is "lying" to the CPU about its consumption to push it more. Although this may happening on all CPUs , but not so sure...
At this point we need the kind of board in use for those cases.

Or may be its early AGESA in general.

 

[Vayra86]

142W peak will do that to any cooling solution but the very biggest.

These temp spikes aren't new. I guess you're just going to have to settle for 'within spec' and pray it won't burn a hole in your socket at some point. The engineers say this is fine, so yeah.

Not too happy about it either though.

 

[Frick]

Yeah I mean you want to use the capabilites of the CPU right? That means more heat, and spiking while doing anything is fine. Wasn't there an outrage for some Ryzen generation because measuring software read the ms-duration spikes and reported those values as max? Also I assume those are die temps and not package temps.

 

[birdie]

Yeah I mean you want to use the capabilities of the CPU right? That means more heat, and spiking while doing anything is fine. Wasn't there an outrage for some Ryzen generation because measuring software read the ms-duration spikes and reported those values as max? Also I assume those are die temps and not package temps.

I personally mean nothing while people report Ryzen Master temps. No one uses HWiNFO64 from mid 2019.

 

[tabascosauz]

I guess this is why the early 5800X adopters should have waited for reviews and opted for a 5600X or 5900X. So far there isn't a single Vermeer chip that's being pushed as hard as the 5800X.

• The 6-core has lower TDP, is less dense, and sustains the same all-core frequency with much less Vcore. Counterintuitively, it almost looks like it has better chiplet quality.
• The 12-core has the second lowest thermal density, and runs lower Vcore and all core.
• While the 16-core is the only other SKU with full chiplets, it runs more conservatively than the 12-core and runs even lower all-core.

That just leaves the 5800X, running a blistering 4.5GHz all core at 1.35V+ on one chiplet, while all of its other brethren have dialed all core Vcore back to sub-1.3V this time around.

The 3800X ran barely higher all-core stock than a 3700X, and already didn't run much cooler than a 3900X. Not entirely surprising that the 5800X is like this.

AMD firmware restricted the 3700X in AGESA somewhere after 1004, to have it run cooler and closer to 65W and stop cannibalizing the 3800X. AMD could make firmware changes to dial back the Vcore in P95 and Cinebench here, but how likely idk.

 

[EarthDog]

No one uses HWiNFO64 from mid 2019.

That's because it's been updated many times since then.

In fact, isn't that the application which tells you about the power deviation that @Zach_01 mentioned earlier?

 

[Zach_01]

That's because it's been updated many times since then.

In fact, isn't that the application which tells you about the power deviation that @Zach_01 mentioned earlier?

And keep evolving....

HWiNFO v6.35-4305 Beta released

HWiNFO v6.35-4305 Beta available. Changes: Enhanced sensor monitoring on ASUS H570, B560, H510 and Q570 series. Added reporting of Precision Boost Clock Limit and Automatic OC Offset on AMD Vermeer. Fixed monitoring of CPU power and HTC status on AMD Zen3.

www.hwinfo.com

 

[Deleted member 24505]

So no jokes about about hot and sweaty Intel now eh

Someone on reddit said this-
The 5800x and 5950x have the exact same 105W TDP and 142W max power limit. So there is a larger power and thermal budget available per core on the 5800x vs the 5950x.

On the 5800x, all that power goes to one die, and is divided amongst only 8 cores. While on the 5950x, you have the exact same 142W divided amongst 16 cores and two separate dies which will be easier to cool. So the 5800x running hotter and being able to draw more power per core makes total sense.

And the 5600x is 65W(88W power budget), so it has a 38% lower max power draw than the 105W chips, but only two fewer cores. So again this ratio makes total sense.

sense or not?

 

[thesmokingman]

So no jokes about about hot and sweaty Intel now eh

Someone on reddit said this-
The 5800x and 5950x have the exact same 105W TDP and 142W max power limit. So there is a larger power and thermal budget available per core on the 5800x vs the 5950x.

On the 5800x, all that power goes to one die, and is divided amongst only 8 cores. While on the 5950x, you have the exact same 142W divided amongst 16 cores and two separate dies which will be easier to cool. So the 5800x running hotter and being able to draw more power per core makes total sense.

And the 5600x is 65W(88W power budget), so it has a 38% lower max power draw than the 105W chips, but only two fewer cores. So again this ratio makes total sense.

sense or not?

Total sense. The problem is that the complainers did not do their due diligence. The 5800x TDP is well advertised. It only take minimal logic to see how much more power is available to that chip given its TDP vs the number of cores. PPL seem to think the 5800x is the same as a 5600x... smh.

 

[TumbleGeorge]

Better chiplets vs worse chiplets by powerdrow? Better for flagship?

 

[thesmokingman]

Better chiplets vs worse chiplets by powerdrow? Better for flagship?

Read the label on a 5800x. It has a 105w TDP to feed its 8 cores instead of 12 or 16 cores.

 

[Zach_01]

So no jokes about about hot and sweaty Intel now eh

Someone on reddit said this-
The 5800x and 5950x have the exact same 105W TDP and 142W max power limit. So there is a larger power and thermal budget available per core on the 5800x vs the 5950x.

On the 5800x, all that power goes to one die, and is divided amongst only 8 cores. While on the 5950x, you have the exact same 142W divided amongst 16 cores and two separate dies which will be easier to cool. So the 5800x running hotter and being able to draw more power per core makes total sense.

And the 5600x is 65W(88W power budget), so it has a 38% lower max power draw than the 105W chips, but only two fewer cores. So again this ratio makes total sense.

sense or not?

Yes it does...

Total sense. The problem is that the complainers did not do their due diligence. The 5800x TDP is well advertised. It only take minimal logic to see how much more power is available to that chip given its TDP vs the number of cores. PPL seem to think the 5800x is the same as a 5600x... smh.

...or a 3700X, 65W TDP (88W PPT) also

 

[TheoneandonlyMrK]

Source 1.
Source 2.

Many have resorted to enabling Eco Mode in Ryzen Master which basically turns the 5800X into a yet to be announced 5700X. From the Ryzen Master manual:



A number of people disagreed with me when I said that AMD basically OC'ed Ryzen 5000 CPUs as much as possible right out of the gate to gain that coveted single threaded performance lead and it looks like it's exactly what people are dealing with now. Luckily enabling Eco Mode translates into a ~8% performance loss in MT Cinebench R20 scores which shouldn't affect games but it's still not pleasant for those who have shelled out $450. But here's a problem: the stock Ryzen 3700X scores around 4750 points while TDP limited 5800X scores around 5350 which is a 12% performance increase while costing 36% more.

Well surprise, AMD and Intel adopt boost algorithms to maximize performance beyond stock and shit gets hot.
Are you a Noob , since when was this new my Ryzen third gen on water gets hot as did my Q6600 Fx8350 and many others.

 

[Vayra86]

Well surprise, AMD and Intel adopt boost algorithms to maximize performance beyond stock and shit gets hot.
Are you a Noob , since when was this new my Ryzen third gen on water gets hot as did my Q6600 Fx8350 and many others.

The difference now is that you're very limited in controlling those spikes. The algorithm thinks for you and that takes some getting used to, at least. People like control.

 

[TumbleGeorge]

Read the label on a 5800x. It has a 105w TDP to feed its 8 cores instead of 12 or 16 cores.

Same TDP for one chiplet vs two and also boost low. That speak of much worse efficiency. This depend more electrical loses= more heat to achieve same or close frequency.

 

[Dux]

Interesting. My max temp on R9 5900X was 65C with OCCT and small data set. Fortron Windale 6 cooler with 2 X 120mm fans.

 

[8BitZ80]

I guess this is why the early 5800X adopters should have waited for reviews and opted for a 5600X or 5900X. So far there isn't a single Vermeer chip that's being pushed as hard as the 5800X.

• The 6-core has lower TDP, is less dense, and sustains the same all-core frequency with much less Vcore. Counterintuitively, it almost looks like it has better chiplet quality.
• The 12-core has the second lowest thermal density, and runs lower Vcore and all core.
• While the 16-core is the only other SKU with full chiplets, it runs more conservatively than the 12-core and runs even lower all-core.

That just leaves the 5800X, running a blistering 4.5GHz all core at 1.35V+ on one chiplet, while all of its other brethren have dialed all core Vcore back to sub-1.3V this time around.

The 3800X ran barely higher all-core stock than a 3700X, and already didn't run much cooler than a 3900X. Not entirely surprising that the 5800X is like this.

AMD firmware restricted the 3700X in AGESA somewhere after 1004, to have it run cooler and closer to 65W and stop cannibalizing the 3800X. AMD could make firmware changes to dial back the Vcore in P95 and Cinebench here, but how likely idk.

The 5900X actually has an 8+4 core configuration this time around, so it does get one fully-enabled CCD like the 5800X.
So even though it's running higher frequencies and the temperatures should be worse on the 5900X by nature, it seems like AMD has something going on to actually make it run cooler.

 

[Frick]

The difference now is that you're very limited in controlling those spikes. The algorithm thinks for you and that takes some getting used to, at least. People like control.

Yeah well the CPU knows its own silicone better than people so...

 

[R0H1T]

The difference now is that you're very limited in controlling those spikes. The algorithm thinks for you and that takes some getting used to, at least. People like control.

You know that scheduler on an OS, that also takes control away from the user ~ I bet most people don't like that either.

 

[Athlonite]

...or a 3700X, 65W TDP (88W PPT) also

my 3700X pulls way more than 88W try double it's TDP at 135.8W during CBr20 multi mind you it's also pulling an all core of 4450MHz and reaches 77c

 

[ratirt]

Funny because I looked over the TPUs reviews and it would seem the 5800X doesn't reach temps the 5900X does. The power draw on the other hand is similar with the 5800X drawing a bit less power. Also the voltage is quite high on the 5800X 1.355v compared to 1.238v on the 5900X.

The 5900X actually has an 8+4 core configuration this time around

I think you are wrong. The 5900X uses 6+6 and that is why the price is so competitive comparing to the 5800X because this one has to use 8 core chiplet. So the 5800X and 5950X are the only chips using 8core chiplets. Also, the 5950X uses the top quality silicon while the 5800X uses any 8core chiplet which passed testing process.
That is why there is a discrepancy in the voltage that the 5800X uses (all 5000 series chips use around 1.2v while the 5800x uses 1.35v and that is a noticeable difference). The voltage is higher than 3800X and that is the only chip in the 5000 series family which has temps and voltage higher than the previous gen counterpart.