[bmagnien]

Guys what if we set PBO to advanced and +0 but use Multi 55 and CO -30 will this work?

Or whats the best way to get 5500 stable and not use eclk? (my b650e cant eclk)

No - setting a static OC (core clock or voltage) will disable any PBO (including CO) you have active. Only way to get higher static core clock OC stable (that I’m aware of) is to try to dial in the right accompanying static voltage, and/or to reduce temps. Even then, you may not be able to get past a certain clock threshold potentially constrained by silicon quality. I spent several hours straight last night trying to find the highest core I could run cb23, and the highest core I could boot into windows. Maxed out at 5665 and 5875 respectively, at 1.25 and 1.3v. But for each I tried every voltage between 1 and 1.5 going by .05 intervals. Going from 1.3 to 1.5 didn’t get me anywhere, even though temps were kept at or below 70 up to 200w, so maybe those clocks are just my chip’s limits without going sub 0 or increasing voltage beyond 1.5 which I’m not going to do on just direct die water.

 

[bmagnien]

That's the magic of eclk / pbo. The voltage it gets is about 1.341v. If I set this static even 5550 wouldn't be anywhere near stable. I don't know how it works but it works..

If we compare PBO + eCLK vs PBO only so 5425 vs 5650 the difference in Dawntrail was about 3k points (52.7k vs 55.8k).

Yah that combo of PBO and ECLK is very strong. So you can boot into windows and run dawntrail at +200 offset, 0 CO, 104.2 ECLK. Can you run cb23 there? CPUz bench? 3dmark cpu profile?

 

[Xdrqgol]

Why is the mobo training ram every restart? Is there an option I have missed or this is “normal?”

i understand the training memory when you change the settings but why during restart?

thanks

 

[digitalfrost]

Why is the mobo training ram every restart? Is there an option I have missed or this is “normal?”

i understand the training memory when you change the settings but why during restart?

thanks

Did you enable memory context restore?

 

[domdtxdissar]

Thank you for this amazing work.

Quick question did you adjust the Scaler setting 10x in bios?

Yeah for the GB6 tests i ran with 10x scaler

 

[bmagnien]

So apparently manual cpu oc does not override CO settings. Even though in the tool tip it says it overrides PBO settings. I thought CO was a part of PBO? It lives within the PBO context menu…🤨. So what parts of PBO get deactivated when setting a manual OC? Just core offset? Scalar? PPT/TDC/EDC limits? I’m going to have to redo all my testing from last night because I did it with an all core -30 CO set, assuming that it didn’t matter because manual oc was deactivating all PBO settings…😡

 

[HippoLOL]

Yeah for the GB6 tests i ran with 10x scaler

Thank you. Also, 1 more question did you disable Spread spectrum?

 

[Xdrqgol]

Did you enable memory context restore?

It was disabled apparently. Hehe maybe I did it by mistake.
I enabled it and the issue is gone! ✌

 

[dk_mic]

Going from 1.3 to 1.5

very brave XD

So apparently manual cpu oc does not override CO settings. Even though in the tool tip it says it overrides PBO settings. I thought CO was a part of PBO? It lives within the PBO context menu…🤨. So what parts of PBO get deactivated when setting a manual OC? Just core offset? Scalar? PPT/TDC/EDC limits? I’m going to have to redo all my testing from last night because I did it with an all core -30 CO set, assuming that it didn’t matter because manual oc was deactivating all PBO settings…😡

what exactly did you notice?

 

[bmagnien]

very brave XD

what exactly did you notice?

Well core vids at idle shot up in hwinfo after booting with the static oc and then changing -30 to 0 in the SMU tool. Which tells me the -CO was active with a static OC. But whether that means anything remains to be seen…I’m rechecking all my static core cocks:static voltage combos and so far they’re largely in line with my limits from last night. So maybe it’s irrelevant. If anything temperature seems to be the driving factor now. My water is ~5c warmer this afternoon and it’s causing instability, which is a bit annoying as I’m trying to dial in a daily that’ll work independent of a few degrees room fluctuation.

 

[dk_mic]

as long as the actual vcore value from the motherboard sensor reports the same, there should not be any difference

 

[bmagnien]

as long as the actual vcore value from the motherboard sensor reports the same, there should not be any difference

Yah I believe that remained unchanged, overridden to the set static voltage. I was just surprised that releasing the CO in SMU tool showed any change to voltage readings, but probably overreacted to seeing the core vids jump.

 

[Nico67]

PBO power limits, CO and curve shaper should still work, but boost and fmax won't have any effect if you pass that with base freq. Scalar being related to boost probably won't do anything either. I have just been using higher clock multiplier with -vcore offset for the time being. p-states still enabled so it idles still.

 

[dk_mic]

PBO power limits, CO and curve shaper should still work, but boost and fmax won't have any effect if you pass that with base freq. Scalar being related to boost probably won't do anything either. I have just been using higher clock multiplier with -vcore offset for the time being. p-states still enabled so it idles still.

there is no curve when using a static OC, vid requests from the CPU are ignored, same goes for powerlimits, it consumes what it consumes

 

[Arizor]

If anyone is watercooling I'm selling my Optimus water block for cheap.

 

[donnieyeen]

5665mhz, 1.25v, 163w, 57c
View attachment 2682508

voided my warranty just to beat @LtMatt by 200 points. worth it. #framegen4lyfe
5850mhz, 1.3v
View attachment 2682509

Hi! Time to void that warranty some more.

Shout out to @LtMatt

 

[LtMatt]

Hi! Time to void that warranty some more.

View attachment 2682662

Shout out to @LtMatt

Damn Bro, destroyed a direct die 9800X3D on an AIO, respect.

Is your warranty intact?

Spoiler

 

[Ph0enix42]

perhaps im unlucky but i dont think most 9800x3d are able to do -40 just fine.
my unit (sp rating = 114) doesnt do -25 on at least 1 core.
i´ve been using -20 all core, no pbo and have been stable so far.
but I havent testet -40 on any single core yet.

I have similar results. -20 seems fine. -30 all cores report error in y cruncher. Testing -25 now but 4 cores already reported errors

 

[Benni231990]

I Have SP111 and can go -40 And Cinebench no Crash but only Warzone cant run -40 on long gaming session so i must go -35 and now no Problems any more

 

[NateST]

SP seems to be a crap shoot based on what I've seen from others. Eckl + pbo looks to be more potent than pbo + CO.