Question regarding PBO on B450 Pro4f

[liquidplasma]

So my new R5 5600 arrived a while ago and I activated PBO on my motherboard and here are the settings it automatically set up:

I want to known if these settings are safe or not.

Here are the relevant power and clocks while doing a cinebench run.
My question here isn't almost 105w quite high for a 65w? I dunno if it's safe leaving at these settings.
The motherboard in question.

 

[Deleted member 185158]

65w is a description for base clock frequencies.

Not to be confused with all core boost wattage.

Everything looks OK

 

[izy]

So my new R5 5600 arrived a while ago and I activated PBO on my motherboard and here are the settings it automatically set up:
View attachment 271657
I want to known if these settings are safe or not.
View attachment 271658
View attachment 271660
Here are the relevant power and clocks while doing a cinebench run.
My question here isn't almost 105w quite high for a 65w? I dunno if it's safe leaving at these settings.
The motherboard in question.

You should use the Curve optimizer combined with PBO and AutoOC , you will have lower voltages + less heat and better boost clocks.

 

[liquidplasma]

You should use the Curve optimizer combined with PBO and AutoOC , you will have lower voltages + less heat and better boost clocks.

I have no idea how curve optimizer works

 

[izy]

I have no idea how curve optimizer works

You can undervolt your CPU with it based on a curve , you just have to add a number (max -30) then check see if its stable or not , you can do it all core or per core. You can use Ryzen Master first but its better if you do it in Bios.

The menu i think it looks like this on Asrock https://albertherdcom.files.wordpress.com/2021/01/img_20210110_233656.jpg?w=809

 

[liquidplasma]

You can undervolt your CPU with it based on a curve , you just have to add a number (max -30) then check see if its stable or not , you can do it all core or per core. You can use Ryzen Master first but its better if you do it in Bios.

The menu i think it looks like this on Asrock https://albertherdcom.files.wordpress.com/2021/01/img_20210110_233656.jpg?w=809

So in case of that picture the is it 11% or 11mV on vcore?

 

[izy]

So in case of that picture the is it 11% or 11mV on vcore?

It based on a curve , i think the cpu voltage curve goes up or down based on your input i dont think its a flat voltage. You need to do some stress tests after like OCCT or/and CoreCycler.

EDIT:
I tuned mine using -30 and AutoOC + 200Mhz , tested with CoreCycler (it shows what core is failing) and after it was stable in that i ran OCCT (SSE stress test at start , but ran all in the end) seem that i had to go a bit higher on 2 core to make it 100% stable.

Here its mine for example:

 

[Chrispy_]

Automatic PBO set to motherboard defaults aren't great because they vastly exceed the capabilities of most coolers.

You're better off working out what sort of max PPT your cooler can handle at a noise level you're comfortable with, and setting your peak PPT at that level, or slightly higher, before working on the voltage curve for some efficiency gains.

 

[izy]

Automatic PBO set to motherboard defaults aren't great because they vastly exceed the capabilities of most coolers.

You're better off working out what sort of max PPT your cooler can handle at a noise level you're comfortable with, and setting your peak PPT at that level, or slightly higher, before working on the voltage curve for some efficiency gains.

I dont think he will go too high after tweaking the curve with that CPU , my 5700x goes ~78C in stress test on air with PPT 140W , TDC 88 , EDC 125.
I think he should tweak the curve first then lower the PBO values and check the gains.

 

[liquidplasma]

I dont think he will go too high after tweaking the curve with that CPU , my 5700x goes ~78C in stress test on air with PPT 140W , TDC 88 , EDC 125.
I think he should tweak the curve first then lower the PBO values and check the gains.

OK here's what I got so far

 

[izy]

OK here's what I got so far
View attachment 271805

It looks good , is it all stable? Core3 and Core4 cant go more than that? Did you test with CoreCycler ? AutoOC +200?

 

[liquidplasma]

It looks good , is it all stable? Core3 and Core4 cant go more than that? Did you test with CoreCycler ? AutoOC +200?

Yeah, running at +200mhz. Core3 and 4 haven't really tested much yet.

 

[izy]

Yeah, running at +200mhz. Core3 and 4 haven't really tested much yet.
View attachment 271806

After you done with CoreCycler you should run OCCT (SSE) and check for stability, how are the temps/boosts?
Also check for WHEA , you can see them in hwinfo.

 

[liquidplasma]

After you done with CoreCycler you should run OCCT (SSE) and check for stability, how are the temps/boosts?
Also check for WHEA , you can see them in hwinfo.

Temps seems to stay around the 72c range and clocks go all the way to 4650mhz. No whea errors so far either.

 

[izy]

Temps seems to stay around the 72c range and clocks go all the way to 4650mhz. No whea errors so far either.

Thats great , you can OC your RAM too and you are ready to go

 

[Dr. Dro]

If you're running AGESA v2 1.2.0.7, PBO is broken due to EDC bug. On 5600/5600X, do not exceed 90 A and your performance will significantly improve, of course, this means you effectively can't overclock either. If running 105W spec CPU, keep EDC below 140 A. These are stock values.

Welcome to AMD's wondrous world of AGESA bugs.

 

[liquidplasma]

If you're running AGESA v2 1.2.0.7, PBO is broken due to EDC bug. On 5600/5600X, do not exceed 90 A and your performance will significantly improve, of course, this means you effectively can't overclock either. If running 105W spec CPU, keep EDC below 140 A. These are stock values.

Welcome to AMD's wondrous world of AGESA bugs.

I'm on agesa 1.2.0.2

 

[DR4G00N]

If you're running AGESA v2 1.2.0.7, PBO is broken due to EDC bug. On 5600/5600X, do not exceed 90 A and your performance will significantly improve, of course, this means you effectively can't overclock either. If running 105W spec CPU, keep EDC below 140 A. These are stock values.

Welcome to AMD's wondrous world of AGESA bugs.

Maybe it's just an Asus thing? Running 1.2.0.7 on a X370 Taichi & 5600 and it seems to work just fine, 4650 w/ CO -30 all core PPT 130, TDC 80, EDC 160.

 

[Dr. Dro]

Maybe it's just an Asus thing? Running 1.2.0.7 on a X370 Taichi & 5600 and it seems to work just fine, 4650 w/ CO -30 all core PPT 130, TDC 80, EDC 160.

AGESA-level, any brand. The issue manifested is that once EDC exceeds processor fused value, SMU will apply VID_LIMIT around 1.4V, this behavior is technically only intended for the 5800X3D. Benchmark and you'll see it, I personally lose around 3k points on Cinebench R23 w/ my 5950X CPU.

The AGESA 1.2.0.7 PBO bug combined with a 65W TDP processor is hilarious: My 5600X is effectively stuck at 90A EDC : Amd (reddit.com)