ASUS ROG Crosshair VIII Hero (WiFi)

[Jism]

Lol.

We're talking 24/7 manual voltages. Not boost states. The boost states is dependend on current. If the current goes up too high then the boost voltage will lower. Lmao man. Your a joke.

360mm AIO are more towards exotic and not budget cooling. That is if you have a decent set with a push/pull configuration going on. Because the chips are made on 7nm, it is more difficult then to cool a 32nm vishera for that matter. You can push alot more current through a 32nm chip and have it well cooled under 60 degrees with a 360mm rad, but with 7nm it's a different ballpark.

These cpu's start to behave like small nuclear plants. Tremendous amount of heat coming out of one tiny area. Intel suffers from the same issues, esp their K series and all that. They are just difficult to cool at a certain point and that hardens the proces of a succesfull overclock. Apart from that, 7nm is very slim, small, tiny, high current could actually brake it way faster then high currents flowing through a Bulldozer FX at 5GHz absorbing 220W like it's breakfast.

Your just plain wrong. The 1.5V in a boost state is a temporary and not a permanent voltage supplied to the CPU. People who put manual overclocks with a pre-set voltage are the ones who suffer. That is exactly where all these reviews are about. https://www.techpowerup.com/review/asus-rog-crosshair-viii-hero-wifi/14.html

TPU: 1.4V. All these review websites send out a wrong signal. 1.4V is'nt for 24/7 usage. Oh you can run a succesfull benchmark tho for a short period of time, but this is'nt 32nm anymore. These cpu show signs of degrade already in weeks once overvolted too much.

 

[Oberon]

Yes, the boost state depends on the current; very good! It's convenient that you still see voltages in that range for higher-current loads (i.e. loading up a single core with a heavy workload.) Boost states are further limited by the package current limit, but you'll be overriding that when manual overclocking anyhow. Note that I also mentioned you're not going to be running up to 1.5 V 24/7 for a manual overclock because it won't net you anything. Those voltages will be significantly lower as you'll be limited by the speeds attainable by the lowest quality core.

*EDIT* If you want proof of my statement regarding the voltages seen under high-current workloads, see here:

 

[EarthDog]

There you go. Specifies up to 1.5 V for boost behavior. That does NOT state 1.5 V for 24/7 usage, but A) you aren't running 1.5 V for a manual overclock because it doesn't gain you anything, and B) if one core can handle it they all can. The cooling bit is a red herring. If you can't cool it, get better cooling. 360mm AIOs are still in the realm of budget cooling if you're actually chasing max overclocks.

As for the Gigabyte thing, you might want to read it again:

So AMD mentions its ok in short bursts.... gigabyte says 1.45V. Clear as mud!

360mm AIOs are sure as hell NOT in the 'budget realm'... no way. And I would like to see these chips running all cores and threads over 1.35-1.4V.... unless its a cherry chip, it isn't happening for most. But yeah.. dude we are not talking stock behavior and boost, but all c/t overclocking (how most do it) and a 24/7 voltage... come on now... that point was lost???????

 

[Jism]

Yes, the boost state depends on the current; very good! It's convenient that you still see voltages in that range for higher-current loads (i.e. loading up a single core with a heavy workload.) Boost states are further limited by the package current limit, but you'll be overriding that when manual overclocking anyhow. Note that I also mentioned you're not going to be running up to 1.5 V 24/7 for a manual overclock because it won't net you anything. Those voltages will be significantly lower as you'll be limited by the speeds attainable by the lowest quality core.

*EDIT* If you want proof of my statement regarding the voltages seen under high-current workloads, see here:

Sigh. A manual dialled in voltage that is going to hurt the Ryzen CPU's. Your coming up with boosts and all that but people who overclock manually (i.e all core 4.2Ghz or so) do not need boost. They exchange the single threaded performance up to 4.35 ~ 4.5Ghz for a all core 4.2 ~ 4.4Ghz overclock.

You have no idea what your talking about really. Perhaps you should reread the subject first before coming in with boost arguments and voltages. We know what the boost state does.

 

[Oberon]

So AMD mentions its ok in short bursts.... gigabyte says 1.45V. Clear as mud!

360mm AIOs are sure as hell NOT in the 'budget realm'... no way. And I would like to see these chips running all cores and threads over 1.35-1.4V.... unless its a cherry chip, it isn't happening for most. But yeah.. dude we are not talking stock behavior and boost, but all c/t overclocking (how most do it) and a 24/7 voltage... come on now... that point was lost???????

AMD doesn't say anything about short bursts, they just state that 1.45 - 1.5 V is within normal specification for the chip. You can make a 3700X run 1.44 V all day long by loading up a single core as seen in buildzoid's video above. The only reason you don't see the same behavior under multi-core loads is because the other cores aren't capable of achieving the same speeds at reasonable voltages, so all voltages drop in order to 1) achieve the boost clocks capable by each core and 2) respect the package current limit (which, again, you're going to override in order to do all-core overclocking anyhow.) I'm just repeating myself at this point, but the fact that a single core is capable of operating at those voltages under completely stock conditions means that the voltage obviously isn't a concern for the silicon as it pertains to longevity (unless you're one of these people that claims every Ryzen 3000 is going to die within months because AMD has a bunch of idiots as engineers who don't know what they're doing.)

Sigh. A manual dialled in voltage that is going to hurt the Ryzen CPU's. Your coming up with boosts and all that but people who overclock manually (i.e all core 4.2Ghz or so) do not need boost. They exchange the single threaded performance up to 4.35 ~ 4.5Ghz for a all core 4.2 ~ 4.4Ghz overclock.

You have no idea what your talking about really. Perhaps you should reread the subject first before coming in with boost arguments and voltages. We know what the boost state does.

I'm not talking about boost, I'm using that video as an example that the chip is perfectly capable of running those voltages for as long as it needs to even under heavy load. The fact that you can't extrapolate from there (even with help) isn't a failing of mine.

 

[EarthDog]

The fact that you can't extrapolate from there (even with help) isn't a failing of mine.

Weird, I was thinking the same thing from your take on the voltage from that /r link... LOL

But, it even says as much straight up anyway.

It is perfectly okay if your CPU is periodically using 1.4-1.5V to achieve boost frequencies, and you should see dips into sub-1.0V as the CPU goes into idle. These dips may be brief, and that's okay. Load voltages of around 1.2-1.3V are perfectly okay also. This is the processor working as expected.

If 1.35V or 1.4V was "perfectly okay" one would expect it to be mentioned, no???????????

but the fact that a single core is capable of operating at those voltages under completely stock conditions means that the voltage obviously isn't a concern for the silicon as it pertains to longevity

It isn't???!!!??? I think you are taking entirely too many liberties on that assumption.

Again, the scope here is overclocking... and overclocking using all cores and threads like most people do. Not stock, not boost, overclocking all cores and threads. Single core and its temps/abilities have little to do with things here.

 

[Jism]

No man.

The only reason you don't see the same behavior under multi-core loads is because the other cores aren't capable of achieving the same speeds at reasonable voltages, so all voltages drop in order to 1) achieve the boost clocks capable by each core and 2) respect the package current limit (which, again, you're going to override in order to do all-core overclocking anyhow.)

So explain to us internets how people's ryzen are showing serious signs of degradation after running extensively at 1.38 ~ 1.45v for that matter, and we're not talking years but utterly weeks (!) CPU's not able to hold a overclock let alone run on boost at all at some point.

No it's not your package current limit, as that is a software value preset for things like PBO or OEM's who want a absolute 65W CPU in their little NUC or so. The chips suffer from serious degradation once a too high voltage is manually set, and things like CB15 or Linpack is being bashed onto the CPU for extensive amount of time.

ryzen degradation voltage - Google Search

Your like denying that degradation does not exist. Lol. Any CPU on this planet will degrade over time. It's just that within spec it will last at least 10 years. Once overclocking is applied your shortening that lifespan. How much nobody really knows, but your doing it. A 360mm rad not high-end? Lol. I bet you proberly never overclocked serious in your life did you?

Really, push up to 1.4V in your Ryzen, and go by your own knowledge that your ryzen is fine. I'll see you in approx 3 to 6 months complaining that the CPU is no longer holding boost without crashing. That alone is a sign up degradation. Users are reporting it. Nobody knows the safe limit, except for a mild 1.36v for the 2x00 series. For the 3x00 series it's something in between 1.37V and 1.38V. But nothing more if your planning to run 24/7 with that!

 

[Oberon]

*sigh*

You guys are free to believe whatever you want. I've spent enough of my time today trying to dispel the misinformation. Unfortunately, I don't have the time to draw everyone a picture.

 

[EarthDog]

*sigh*

You guys are free to believe whatever you want. I've spent enough of my time today trying to dispel the misinformation.

You've assumed a lot, and proved little.

That said, I appreciate you actually trying and sharing links... it's more than most of the members here can manage and still try to force their misinformed opinions down your throat.



But yeah. All I see is up to 1.3V under load is normal and OK (assuming from all c/t).... from your own link. It says boost up to 1.5V is PERIODICALLY ok with a single thread in boost. Further extrapolation of that statement isn't from the horse's mouth.

 

[Oberon]

You've assumed a lot, and proved little.

That said, I appreciate you actually trying and sharing links... that's more than most of the members here can manage and still try to force their misinformed opinions down your throat.

Hard to prove it when I lack the time to draw the connections that are being missed, but I appreciate the civility (unlike from some others here!)

 

[EarthDog]

There isn't anything being missed. I don't agree with your extrapolation of the quoted material. (see edit above)

Cheers bud.

 

[Jism]

Ryzen are just very sensitive CPU's related to "All core overclocking". A voltage, and we're not talking short burst in boost conditions, set too high can cause fast degradation of the CPU. The 1x00 series suffered from that, the 2x00 series suffered from that, and sure is the 3x00 series. I woud'nt take an advice from an employee working at Gigabyte alone but i prefer statements from AMD themself, or, people like the stilth, who've done tremendous amount of research on the Ryzen arch in general. Pretty much his word is like a golden reference here on the internet.

These CPU's degrade over time; the ryzen 2x00 series is no different from the 3x00 series. A all core overclock at a 1.4V ~ 1.45V is asking for problems. Materials get so small and way more fragile then it was back in the 32mm days, where a stock 1.3V cpu could eat 1.7V to 1.8V on water for breakfast. Its just aint going to happen anymore and really with the small headroom the 2700x has to offer related to max 24/7 voltages, i'll simply slap on a big cooler, slight undervolt and have it boost as long as possible with AMD's specs.

But really, go ahead if you think 1.4 ~ 1.45V into your 3x00 series CPU is safe. We dont make this up. People's CPU's are falling after hard pushed overclocks.

 

[clark75]

ASUS ROG Crosshair VIII WI-FI
Where is the memOK! button?
I cannot find this button...