2700X boosting beyond 4.3GHz on its own?

[Redwoodz]

Nice example of why I like W7. You know my figures are true.

W10 is a liar.

Yeah I know but unfortunately that's a dead end.

 

[iuliug]

Idle boost is irrelevant as it not sustained but it si just a peek single moment. Use Cinebench and see what it does under load single or multi threaded.

 

[Deleted member 185158]

Idle boost is irrelevant as it not sustained but it si just a peek single moment. Use Cinebench and see what it does under load single or multi threaded.

The lying cpu-z validation requires no load to validate the frequency.

Turbo clocks should vary from system to system as well, but pretty certain all the boost states happen in 100mhz frequencies depending on PBO and CPB comfigurations, boost sustained clocks for 2700x is as follows.

4.0ghz sustained
4.1ghz sustained
4.2ghz sustained
4.3ghz sustained.

Frequencies between above figures are from reference (bus clocks of you will) fluctuating.
I do believe of you have the function of cpu spread spectrum, you can control the reference clock and hold it stead.
I unfortunately do not have the option to disable cpu spread spectrum.

 

[sam_86314]

Aha, I caught HWMonitor being drunk on my system...

I wish it was boosting that high...

 

[Zach_01]

At this point I trust only HWiNFO for multi-reporting tool.
Never really liked HWMonitor as always (years ago) some readings were totally off...

 

[Redwoodz]

The lying cpu-z validation requires no load to validate the frequency.

Turbo clocks should vary from system to system as well, but pretty certain all the boost states happen in 100mhz frequencies depending on PBO and CPB comfigurations, boost sustained clocks for 2700x is as follows.

4.0ghz sustained
4.1ghz sustained
4.2ghz sustained
4.3ghz sustained.

Frequencies between above figures are from reference (bus clocks of you will) fluctuating.
I do believe of you have the function of cpu spread spectrum, you can control the reference clock and hold it stead.
I unfortunately do not have the option to disable cpu spread spectrum.

In my case I believe I had cpuz and hwinfo running, along with an insider preview build of W10. Also on a new bios so I dont believe I had the HPET enabled and spread spectrum was probably on. That cpuz sub did load the cpu, but somehow captured the speed as it was returning to idle. I have not been able to reproduce it.

They are doing somethings with Ryzen and W10 as far as boosting and fine grain boost control(like 25mhz inc), not sure all what's going on.

 

[EarthDog]

They are doing somethings with Ryzen and W10 as far as boosting and fine grain boost control(like 25mhz inc), not sure all what's going on.

Ryzen 3000 works that way. I dont think 2000 series does..

 

[Deleted member 185158]

Ryzen 3000 works that way. I dont think 2000 series does..

I dont think Ive paid enough attention to notice a quarter multi, but 2700x does max boost with a half multi a like 4350mhz.
Not sure the difference of influence between PBO and CPB both seem to have same effects when set in manual oc mode.

 

[EarthDog]

If that gen has it, you can see it in the CPU multiplier in the BIOS. I vaguely recall it as .5 multi's only... but, I didn't start the deep dives in Ryzen until Zen 2/X570.

 

[Darmok N Jalad]

I see 4.35GHz with regularity across all cores for the 2700x. AMD has that as the rated max boost.

 

[Deleted member 185158]

I can throttle the cpu via refernce clock and a certain program.
Also with manual OC the system will throttle refernce clock if vcore is too low.

But seemingly and generally speaking the Zen+ chips throttle or boost 100mhz increments.

Might be usefull at quarter multi but the performance gains would be so minimal youd never see it even with most benchmarks.
Neat concept never the less....

 

[Melvis]

My 2700X does 4.35GHz on 1 core and around the 4GHz on all 8 cores, if anyone says theres can do 4.35GHz on all 8 cores is lieing and is reading the software incorrectly. I have never tried PBO honestly.

 

[Darmok N Jalad]

My 2700X does 4.35GHz on 1 core and around the 4GHz on all 8 cores, if anyone says theres can do 4.35GHz on all 8 cores is lieing and is reading the software incorrectly. I have never tried PBO honestly.

I think you misunderstand what I'm saying. Every core reports to hit 4.35GHz eventually, just not all at the same time. 4.0 all-core is also what I get.

 

[JAYGUNNER]

Is this normal behavior for the 2700X? I'm not overclocking at all, but HWMonitor is showing it boosting to 4.5GHz? CPU should only be hitting 4.35GHz, right? Not complaining, but I didn't expect this behavior on the previous Zen generation. Or am I just dealing with a bug in the software?
View attachment 138404

vb

Nice example of why I like W7. You know my figures are true.

W10 is a liar.

So which app has the most accurate reading then?

I think you misunderstand what I'm saying. Every core reports to hit 4.35GHz eventually, just not all at the same time. 4.0 all-core is also what I get.

So which app has the most accurate reading then?

 

[Darmok N Jalad]

HWiNFO had the most accuracy.

 

[INSTG8R]

Not seen mine do it yet but I was in my BIOS half an hour ago and found a whole new overclocking section. I only built in the last week so I have no deep secrets but there was another PBO tab in the new section I found but I had to nuke my install and it’s too hot to put it back together again to test the setting

HWiNFO had the most accuracy.

I would argue AIDA apparently HWinfo puts an artificial load on while using it . I ’ve been using AIDA since it was Everest so I can only go by what others have said