Buildzoid's 3700x Static Overclock Degrades Processor

[milewski1015]

You are reffering to 1xxx and 2xxx chips ? I ask because my 2700X defaults to 1.45V in BIOS.

When PBO/XFR2 is enabled? I recall a while back Buildzoid complaining about the PBO algorithm. Something about how it's extremely stupid that the first thing PBO does is crank the voltage up.

You may want to read this post from r/overclocking...

https://www.reddit.com/r/overclocking/comments/amtnt4

"Conclusion, based on my testing and several others that have seen degradation above 1.38v on this subreddit, I can solidly say, since AMD has released zero safe voltages for this chip, if you value the longevity of your Ryzen gen 2 chip for more than around 4 months, stick to 1.375v as your max voltage. "

and this from the Anandtech forums:

Page 72 - Ryzen: Strictly technical

Page 72 - Seeking answers? Join the AnandTech community: where nearly half-a-million members share solutions and discuss the latest tech.

forums.anandtech.com

"The seen behavior suggests that the full silicon reliability can be maintained up to around 1.330V in all-core workloads (i.e. high current) and up to 1.425V in single core workloads (i.e. low current). Use of higher voltages is definitely possible (as FIT will allow up to 1.380V / 1.480V when scalar is increased by 10x), but it more than likely results in reduced silicon lifetime / reliability."

 

[droopyRO]

Yup. PBO enabled, with it off i can go as low as -1.250V -1.000V offset for 4.1Ghz manual setting. Testing it now to see if it is stable in Prime95. Good thing i ran in to this thread.

 

[Waldorf]

At least for 3xxx cpus, Pbo is more often counter productive to voltages/temps.
And no consistency that performance improves.
Turning it off (which leaves PB running) is usually the better choice for the masses/most ppl..

 

[Zach_01]

You are reffering to 1xxx and 2xxx chips ? I ask because my 2700X defaults to 1.45V in BIOS.

I'm reffering to 3xxx. And some keep posting here info for 2xxx. I'm always reffering to 3xxx from the beggining of this thread.
1.375V for 3xxx is definately not safe in most full load sircumstances. Buildzoid degrade his 3700X with 1.375V.
Even 1.3~1.325V could be lethal potentially. All and all, 3xxx is not for static speeds and voltages. Its not the same with auto boosting and voltage. Its night and day.

Pbo is more often counter productive to voltages/temps.
And no consistency that performance improves.
Turning it off (which leaves PB running) is usually the better choice for the masses/most ppl..

I'm in the middle of DRAM OC and PBO testing. Need a few more days... Wait and see.
All I can say right now is that you can benefit from some PBO settings in a specific way but nothing groundbreaking.

 

[biffzinker]

Turning it off (which leaves PB running)

You mean turning off PBO but leaving the factory boost on - XFR2?

 

[Schmuckley]

You are reffering to 1xxx and 2xxx chips ? I ask because my 2700X defaults to 1.45V in BIOS.

3000 series does too. At least in my BIOS.

AMD Ryzen 5 3600X @ 4423.97 MHz - CPU-Z VALIDATOR

[6a39qf] Validated Dump by Schmuckley (2020-02-28 21:22:48) - MB: Asus ROG CROSSHAIR VII HERO - RAM: 32768 MB

valid.x86.fr

Zoooohh-Nooooeeessss! Teh degradation!

IN b4 I'm eating my snarkiness in 6 mos. So I'm actually running this chip with .035 less voltage than stock auto settings apply.
This is no stellar sample, it wants 1.46v.
It seems to me this is what the chip wants to be stable.
No crashes in this configuration. With PBO there were crashes every other boot and running hot.
'bout to get some crashes bcuz tinkering with RAM.

 

[Bones]

Now I'm wondering if my RAM's gonna degrade too along with the 3600x.......

Bones`s Memory Frequency score: 2232.8 MHz with a DDR4 SDRAM

The DDR4 SDRAM @ 2232.8MHzscores getScoreFormatted in the Memory Frequency benchmark. Bonesranks #295 worldwide and #293 in the hardware class. Find out more at HWBOT.

hwbot.org

Yep - Defaults spec'ed by the board maker tends to be higher than it needs but thats because not every chip will work with less voltage, they have to spec a "Blanket" voltage to be sure any non-defective chip dropped in would work.

 

[moproblems99]

This is no stellar sample, it wants 1.46v.

I am pretty sure it doesn't want 1.46v all core.

 

[Zach_01]

3000 series does too. At least in my BIOS.

AMD Ryzen 5 3600X @ 4423.97 MHz - CPU-Z VALIDATOR

[6a39qf] Validated Dump by Schmuckley (2020-02-28 21:22:48) - MB: Asus ROG CROSSHAIR VII HERO - RAM: 32768 MB

valid.x86.fr

Zoooohh-Nooooeeessss! Teh degradation!

IN b4 I'm eating my snarkiness in 6 mos. So I'm actually running this chip with .035 less voltage than stock auto settings apply.
This is no stellar sample, it wants 1.46v.
It seems to me this is what the chip wants to be stable.
No crashes in this configuration. With PBO there were crashes every other boot and running hot.
'bout to get some crashes bcuz tinkering with RAM.

Now I'm wondering if my RAM's gonna degrade too along with the 3600x.......

Bones`s Memory Frequency score: 2232.8 MHz with a DDR4 SDRAM

The DDR4 SDRAM @ 2232.8MHzscores getScoreFormatted in the Memory Frequency benchmark. Bonesranks #295 worldwide and #293 in the hardware class. Find out more at HWBOT.

hwbot.org

Yep - Defaults spec'ed by the board maker tends to be higher than it needs but thats because not every chip will work with less voltage, they have to spec a "Blanket" voltage to be sure any non-defective chip dropped in would work.

Am I not visible to all people in here? Seem that some are on their own world. Are they trolling or something?

1.42V all core voltage and load with high current will degrade the CPU eventually.
Ryzen 3000 doesn’t need 1.46 for all core speeds and loads. Don’t post such things that could mislead other users. Please read the posts in here and maybe try to understand...

 

[looniam]

not that i want to quell a spirited discussion but ya know this is ONE SAMPLE.

carry on.

 

[Zach_01]

Please do so...

 

[R-T-B]

He is a kid and gullible.

He is neither. He's literally made a media business out of overclocking, something many of us could never hope to do.

 

[mtcn77]

He is neither. He's literally made a media business out of overclocking, something many of us could never hope to do.

I don't say so needlessly. It is that he is using msi and that entails its own conditions on his supposed 1.375v which is 1.475v in fact.

 

[Zach_01]

I don't say so needlessly. It is that he is using msi and that entails its own conditions on his supposed 1.375v which is 1.475v in fact.

Come on man... did you monitor his test bed?
If that was true be sure about it that he would know. He doesn’t relay on software monitoring.

 

[Bones]

Am I not visible to all people in here? Seem that some are on their own world. Are they trolling or something?

1.42V all core voltage and load with high current will degrade the CPU eventually.
Ryzen 3000 doesn’t need 1.46 for all core speeds and loads. Don’t post such things that could mislead other users. Please read the posts in here and maybe try to understand...

I'm not trolling on my part - At least I'm not trying to.

Anyone can look at what I did with my 3600x, see the speed I had it running at AND voltage used to get there.

In turn most (Hopefully) would realize if it can go that high with such low voltage there is no need to really volt it up to get that from a typical setup running at or not too far from stock speed.

Was using that to show what speeds could be obtained with what amount of voltage in use and it does make the case you probrably don't need 1.40v's or higher to get some nice clocks from a Ryzen 3rd Gen.

 

[moproblems99]

I still don't see the point of all core clocking Ryzen. Why lose out on single/low thread performance?

 

[mtcn77]

Come on man... did you monitor his test bed?
If that was true be sure about it that he would know. He doesn’t relay on software monitoring.

I am doing this guy a favor. He is young and should be given the chance to make a mistake in a clear headed way. If you start manipulating, you start speculating... suddenly, you find yourself despirited. That is how you get emotional burnout.
Don't idolize him, let it come forth. It is not his mistake to make anyway, it is MSI's. It is just that he is underreporting and in all his pure heartedness, that is unlike him, he would have left no stone unturned such as this.
Anyway, I am just highlighting that bold 'MSI' disclaimer in his quandary. It confounds with any logical statement said previously and is in direct conflict with any conclusion thereafter, imo. It is a big no-no.

I still don't see the point of all core clocking Ryzen. Why lose out on single/low thread performance?

I agree fully, however since the cpu adopts fivr regime, it overrides any undervolting in practice unless you make it pack up shop.

Notice Ryzen is right on target where Intel hit home with Ivy Bridge/Haswell. Intel had gained genuineness points for adopting fivr; however cpu-vrm was not mentioned as an overall improvement to vrm mosfets so it be good AMD did it the way it is.

 

[Schmuckley]

I am pretty sure it doesn't want 1.46v all core.

Am I not visible to all people in here? Seem that some are on their own world. Are they trolling or something?

1.42V all core voltage and load with high current will degrade the CPU eventually.
Ryzen 3000 doesn’t need 1.46 for all core speeds and loads. Don’t post such things that could mislead other users. Please read the posts in here and maybe try to understand...

I do understand Chicken Littleism. And yes, it sucks 1.46v on auto.

 

[moproblems99]

I am doing this guy a favor. He is young and should be given the chance to make a mistake in a clear headed way. If you start manipulating, you start speculating... suddenly, you find yourself despirited. That is how you get emotional burnout.
Don't idolize him, let it come forth. It is not his mistake to make anyway, it is MSI's. It is just that he is underreporting and in all his pure heartedness, that is unlike him, he would have left no stone unturned such as this.
Anyway, I am just highlighting that bold 'MSI' disclaimer in his quandary. It confounds with any logical statement said previously and is in direct conflict with any conclusion thereafter, imo. It is a big no-no.

Did I miss something? I didn't watch the video because YouTube sucks and I avoid it at all costs but how do you know he didn't verify with a meter? How do you know he didn't set the voltage in the uefi to reflect the msi overvolt? How do you know MSI under reports? It doesn't make any sense for them to not report the correct voltages.

And yes, it sucks 1.46v on auto.

At load, on all cores? In action? Mine draws 1.46V but that is only when 1 or 2 threads is active.

 

[Schmuckley]

I still don't see the point of all core clocking Ryzen. Why lose out on single/low thread performance?

What single core performance?

Did I miss something? I didn't watch the video because YouTube sucks and I avoid it at all costs but how do you know he didn't verify with a meter? How do you know he didn't set the voltage in the uefi to reflect the msi overvolt? How do you know MSI under reports? It doesn't make any sense for them to not report the correct voltages.



At load, on all cores? In action? Mine draws 1.46V but that is only when 1 or 2 threads is active.

During Cinebench R20.

 

[moproblems99]

What single core performance?

I don't understand what you are asking.

During Cinebench R20.

Are you doing an all-core fixed multiplier overclock or did you leave everything on auto so the processor is functioning how it was intended from the factory?

 

[Zach_01]

What single core performance?


During Cinebench R20.

There is no chance that Ryzen 3000 will take by its own and on full stock more than 1.35-1.36V on full, all thread load. Higher chips take even less (1.15~1.25V)
And again after a thousand times I said it it’s not the same with static voltage. Night and day...

Look the Core voltage SVI2 TFN. That is the actual Vcore. This under R20 all core load full stock settings.

I do understand Chicken Littleism. And yes, it sucks 1.46v on auto.

Only when idle and very low load and current. It’s not for all core loads

Yet another one with better temp(more cooling)

 

[Schmuckley]

There is no chance that Ryzen 3000 will take by its own and on full stock more than 1.35-1.36V on full, all thread load. Higher chips take even less (1.15~1.25V)
And again after a thousand times I said it it’s not the same with static voltage. Night and day...

Look the Core voltage SVI2 TFN. That is the actual Vcore. This under R20 all core load full stock settings.
View attachment 146223


Only when idle and very low load and current. It’s not for all core loads

Yet another one with better temp(more cooling)
View attachment 146226

That's not not going any higher than 4.05Ghz, though.
Manually you could (should be able to) get that down to 1.3v or less more often than not..
Or maybe clock a little higher.

I don't understand what you are asking.



Are you doing an all-core fixed multiplier overclock or did you leave everything on auto so the processor is functioning how it was intended from the factory?

Neither. it's a manual overclock with offset voltage and C n Q enabled. All cores do ramp up...yes.

PBO is disabled as well as boosting. The single core performance is just dandy!

543 with CPU-Z

 

[Zach_01]

That's not not going any higher than 4.05Ghz, though.
Manually you could (should be able to) get that down to 1.3v or less more often than not..

Are we talking about stock settings or not? You said that it takes 1.46V at stock... where is that exactly?
Changing the subject?

Neither. it's a manual overclock with offset voltage and C n Q enabled. All cores do ramp up...yes.

PBO is disabled as well as boosting. The single core performance is just dandy!

543 with CPU-Z

How is that neither? Manual OC is fixed multi...

 

[Schmuckley]

Are we talking about stock settings or not? You said that it takes 1.46V at stock... where is that exactly?
Changing the subject?


How is that neither? Manual OC is fixed multi...

A Ryzen 5 3600 allegedly should turbo up to 4.2 Ghz, which the one in those charts is not doing.
I guess I am doing fixed multi then.
It takes 1.46v at all-auto settings. Gets very hot under load, too.
I know this newer stuff can take more heat, but I'm from the days of keep it below 55c if it's AMD.