cancel
Showing results for 
Search instead for 
Did you mean: 

Having trouble switching to Adaptive voltage mode on Z390 Maximus XI Formula + 9900k

Zammin
Level 9
Hi everyone

I've just recently finished dialing in my stable 5Ghz overclock in manual mode with the following settings:

- AI Overclock Tuner XMP I (clicked Yes at the prompt)
- Sync all cores
- AVX Offset 0
- Manual Voltage Override 1.33V
- LLC7
- IA/DC Loadlines 0.01
- Loadline Synch diabled
- CPU Current Capability 170%
- Long and Short Duration Package Power Limit 4095
- Package Power Time Window 127
- CPU Core/Cache Current Limite Max 255.75
- Cache Ratios Auto
- VCCIO 1.15V
- System Agent Voltage Auto (1.2V)
- BCLCK Aware Adaptive Voltage disabled
- MCE disabled
- SVID disabled
- Intel SpeedShift/Step disabled
- C-States disabled

Now that I'm stable in manual mode I'm currently trying to switch to adaptive voltage mode. I've done this before on Z370 and my 8700k with success, however with the 9900k and my ASUS Z390 motherboard it is exhibiting some strange behavior.

In manual mode it sits at 1.314V idle and drops to 1.27V under RealBench (or AIDA64) types of loads. When I re-enable C SpeedShift/Step, C-States and SVID, then switch to Adaptive voltage and set 1.33V as the additional turbo voltage it idles between 1.33-1.35V and sits around 1.33V under load, causing it to run much hotter than before. The only way I can get it to come down to 1.33V idle and sag to 1.27V under load again is to set LLC to level 6. This is all very strange to me as I would expect the LLC behavior as well as the idle voltage and load voltages to be the same as before until I set the windows power plan to balanced to allow the CPU to clock down when at idle.

This is not in line with my experience overclocking my 8700k on my ASUS Z370 Maximus X Code. In this case on the Z390 motherboard LLC6 in adaptive mode behaves closer to the way LLC7 behaves in manual mode, although the voltage level is not as stable under load, it fluctuates up and down a lot instead of sitting at 1.27V for the duration of the test. This appears to cause more heat and definitely isn't stable as I get BSOD's in RealBench with these settings in adaptive mode.

I'm not really sure what to do, I want to use adaptive mode to allow my CPU to clock down (and bring the voltage down) at idle but I just can't get it to work. I had no trouble doing it on Z370 following Raja's Z270 overclocking guide but it just doesn't want to work on Z390 for some reason.

If anyone has any advice for me I would greatly appreciate it. I've listed my settings above but if anyone requires BIOS screenshots I can get them for you as well.
12,334 Views
39 REPLIES 39

Zammin wrote:
Hi, thanks for your reply

So if I understand you correctly, I'm seeing this strange behavior in adaptive mode because I am not exceeding the maximum single core boost speed, even though all cores are sync'd? I suppose that would explain why I did not see this behavior on my 8700k.



Yes I am aware of this already, I've seen elmors post about it on OCN along with the graph he posted. Thanks for the link to that video though.



Sorry but I don't fully understand this part. I know that LLC7 is the minimum stable LLC for manual mode, I've tried LLC6 and increased target voltage to compensate but it just won't be stable. I spent a few hours trying today. Someone on OCN said the same thing about lower LLC having better transient response and requiring a lower VMIN but I can't seem to replicate that. Even at the same voltage under load with LLC6 it's not stable, where as LLC7 is.

So knowing that LLC7 appears to be the most stable for my CPU in manual mode, what should I do for adaptive mode? When I run LLC7 in adaptive mode I get overshoot at idle and at most, only a very small drop in voltage under load (from target 1.33V to 1.32V). When I run LLC6 it idles at target voltage and drops about the right amount under load but is unstable. The only thing I haven't tried is maintaining LLC7 in adaptive mode and lowering the voltage target. Is that worth trying?

Thanks for your advice.


That would be me "I'm a "someone?" Ouch that hurt =.= that said "Lower LLC= better transient response=lower VMIN required"
But it was Shamino himself that said it before anyone else (except maybe Raja).
I got that information directly from here.

https://rog.asus.com/forum/showthread.php?106375-MCE-explanations-and-others&highlight=explanations

But I am only using manual voltage, NOT auto.
This issue with lower VMIN able to be used with lower LLC + higher VID (or bios voltage) to compensate becomes EXTREMELY obvious when testing small FFT prime95 FMA3 enabled (especially the 15K fixed in-place FFT custom test--15K draws the most amps).

Falkentyne wrote:
That would be me "I'm a "someone?" Ouch that hurt =.= that said "Lower LLC= better transient response=lower VMIN required"
But it was Shamino himself that said it before anyone else (except maybe Raja).
I got that information directly from here.

https://rog.asus.com/forum/showthread.php?106375-MCE-explanations-and-others&highlight=explanations

But I am only using manual voltage, NOT auto.
This issue with lower VMIN able to be used with lower LLC + higher VID (or bios voltage) to compensate becomes EXTREMELY obvious when testing small FFT prime95 FMA3 enabled (especially the 15K fixed in-place FFT custom test--15K draws the most amps).


Hey man! Yes that was you haha, my apologies I just didn't know you were on this forum as well. You were the first person I heard it from, but I haven't really been looking much into it until recently.

Problem is I just can't replicate it. If I use lower LLC (e.g. LLC6) with higher VID to compensate, it's not stable at the same load voltage. It will either detect instability in RealBench or BSOD during the 30 min test. Not saying that the whole thing isn't true, I just can't seem to replicate it on my end (in manual mode). I might be doing something wrong though since I've only just started tweaking this CPU over the last week. I'm going to try Shamino's suggestion of tweaking the IA/DC LoadLines today and see how I go.

Zammin wrote:
Hey man! Yes that was you haha, my apologies I just didn't know you were on this forum as well. You were the first person I heard it from, but I haven't really been looking much into it until recently.

Problem is I just can't replicate it. If I use lower LLC (e.g. LLC6) with higher VID to compensate, it's not stable at the same load voltage. It will either detect instability in RealBench or BSOD during the 30 min test. Not saying that the whole thing isn't true, I just can't seem to replicate it on my end (in manual mode). I might be doing something wrong though since I've only just started tweaking this CPU over the last week. I'm going to try Shamino's suggestion of tweaking the IA/DC LoadLines today and see how I go.


I just replied on OCN to your post just now.
But this thing about the IA AC DC Loadlines...
When you are using manual voltages (override), the IA AC and IA DC loadlines are completely ignored, or should be completely ignored.
They only affect the CPU VID.

Do you have "Synch AC DC loadline to VRM loadline" enabled or disabled?

Because Loadline Calibration" is "VRM loadline", and assuming it is not set to "Auto" (idk how Asus handles this), the VRM "loadline voltage droop" should completely, totally ignore the DC loadline "VID droop" (DC loadline controls power measurements and vdroop along the CPU VID).

CPU Package Power however is based on the CPU VID, not the CPU Vcore, so your CPU package power is VID * Amps, so if the VID is extremely lower (e.g. if set to 0.01 / 0.01 for IA AC and DC) and you are using higher manual voltages, your package power will be much lower than expected. Check the VRM readout to see if it has "Power (Output) or POUT, and IOUT (Amps).

In your test you did, you said that you had IA AC DC loadline at auto and were unstable, when using *manual* voltages,
But then you set it to 0.01 / 0.01 and were stable at the same manual voltages. Did you check the cpu temp readout and the load voltage to verify the vcore and temps were identical in both the "unstable" IA AC loadlines and the 0.01 stable ones?

Or am I being stupid here and your "manual" overclock means that you were using adaptive voltage + manually setting the CPU core and cache ratios (so you were not using manual voltage?).

IA AC loadilne affects the CPU Power supply and boosts the VID up by a certain amount (based on something I don't know about; I presume Shamino and Raja know), even though it's based on mOhms, it does not seem to function like DC loadline does.

DC loadline causes the VID to drop based on current, in mOhms: example: 100 amps of current * DC value of 1.6 is 100 amps * 1.6 mOhms=160mv of the VID dropping at full load, compared to full idle.

Theoretically, if you set IA AC and IA DC loadline manually to 1.6 in your bios, AND set Loadline Calibration to Auto (or normal) AND used no offsets or adaptive voltage setting at all, just "Auto" (or possibly, adaptive without additional turbo voltage, sorry I don't know how the Asus UEFI works), then your CPU VID should 100% match the CPU vcore, within 5mv.
(Reference AC / DC loadline for 9th gen 8 core S series CPU is 1.6 mOhms; for 4 and 6 core it's 2.1 mOhms).

Falkentyne wrote:
I just replied on OCN to your post just now.
But this thing about the IA AC DC Loadlines...
When you are using manual voltages (override), the IA AC and IA DC loadlines are completely ignored, or should be completely ignored.
They only affect the CPU VID.

Do you have "Synch AC DC loadline to VRM loadline" enabled or disabled?

Because Loadline Calibration" is "VRM loadline", and assuming it is not set to "Auto" (idk how Asus handles this), the VRM "loadline voltage droop" should completely, totally ignore the DC loadline "VID droop" (DC loadline controls power measurements and vdroop along the CPU VID).

CPU Package Power however is based on the CPU VID, not the CPU Vcore, so your CPU package power is VID * Amps, so if the VID is extremely lower (e.g. if set to 0.01 / 0.01 for IA AC and DC) and you are using higher manual voltages, your package power will be much lower than expected. Check the VRM readout to see if it has "Power (Output) or POUT, and IOUT (Amps).

In your test you did, you said that you had IA AC DC loadline at auto and were unstable, when using *manual* voltages,
But then you set it to 0.01 / 0.01 and were stable at the same manual voltages. Did you check the cpu temp readout and the load voltage to verify the vcore and temps were identical in both the "unstable" IA AC loadlines and the 0.01 stable ones?

Or am I being stupid here and your "manual" overclock means that you were using adaptive voltage + manually setting the CPU core and cache ratios (so you were not using manual voltage?).

IA AC loadilne affects the CPU Power supply and boosts the VID up by a certain amount (based on something I don't know about; I presume Shamino and Raja know), even though it's based on mOhms, it does not seem to function like DC loadline does.

DC loadline causes the VID to drop based on current, in mOhms: example: 100 amps of current * DC value of 1.6 is 100 amps * 1.6 mOhms=160mv of the VID dropping at full load, compared to full idle.

Theoretically, if you set IA AC and IA DC loadline manually to 1.6 in your bios, AND set Loadline Calibration to Auto (or normal) AND used no offsets or adaptive voltage setting at all, just "Auto" (or possibly, adaptive without additional turbo voltage, sorry I don't know how the Asus UEFI works), then your CPU VID should 100% match the CPU vcore, within 5mv.
(Reference AC / DC loadline for 9th gen 8 core S series CPU is 1.6 mOhms; for 4 and 6 core it's 2.1 mOhms).


Hi mate, sorry this is a lot to reply to and my knowledge of this stuff is not at the same level, compared to you guys I am still fairly new to this stuff so i will try answer what I can to the best of my ability.

Loadline Synch is disabled in my Manual OC and also when I am trying Adaptive mode. I originally tried enabling it when I was first dialing in my Manual OC but it required more voltage than disabling it and setting the loadlines to 0.01.

LLC is not set to auto, although I have tried that once with adaptive mode and it finds it's way up to LLC7 on it's own during RealBench.

I think the load voltages were the same when I tested with loadlines set to auto and 0.01, but for the sake of being sure I could run another 1 hour of realbench tonight with them on auto to double check if necessary. I don't remember what the exact temps were as it was over a week ago and I have run many tests and made a number of tweaks since then, but in all my RealBench runs the temps generally do hit momentary peaks of around 92-95C depending on ambient.

When I say Manual, I do mean Manual Voltage Override. Not Adaptive.

I hope that addresses your questions and comments, let me know if I missed anything,

Zammin wrote:
Hi mate, sorry this is a lot to reply to and my knowledge of this stuff is not at the same level, compared to you guys I am still fairly new to this stuff so i will try answer what I can to the best of my ability.

Loadline Synch is disabled in my Manual OC and also when I am trying Adaptive mode. I originally tried enabling it when I was first dialing in my Manual OC but it required more voltage than disabling it and setting the loadlines to 0.01.

LLC is not set to auto, although I have tried that once with adaptive mode and it finds it's way up to LLC7 on it's own during RealBench.

I think the load voltages were the same when I tested with loadlines set to auto and 0.01, but for the sake of being sure I could run another 1 hour of realbench tonight with them on auto to double check if necessary. I don't remember what the exact temps were as it was over a week ago and I have run many tests and made a number of tweaks since then, but in all my RealBench runs the temps generally do hit momentary peaks of around 92-95C depending on ambient.

When I say Manual, I do mean Manual Voltage Override. Not Adaptive.

I hope that addresses your questions and comments, let me know if I missed anything,


Yes that test would be helpful if possible.
See if you wind up being stable in both tests (only changing AC/DC loadliine from 0.01 to auto and nothing else).
of course make sure you use manual voltage and a loadline calibration you know to be stable (e.g. LLC6 or whatever you choose).
Keep that sync ac/dc to vrm loadline disabled. I'm not sure if enabling it changes the AC DC loadline to match the VRM (LLC) in mOhms, or if it changes the vRM (LLC) to match the AC DC value or not, so lets try to get that out of the picture.

I've done repeated testing on a GB board and with a fixed LLC value (e.g. high), and manual voltage, there is no change at all in temps, amps draw or stability, when changing IA AC/IA DC to 0.01 mOhms or auto or the default (1.6 mOhms) for 9900K.

What does affect "VMIN" stability is loadline calibration. At a 'target' VR VOUT (CPU on-die sense voltage, which the Maximus XI boards all use now as vcore), you will require a higher VMIN for bypassing borderline stability or instability at higher LLC, than with a lower (more vdroop or steeper) LLC+higher bios voltage, when stressing heavy loads (especially AVX/FMA3!)

Example:
Bios voltage: 1.295v, max load VR VOUT voltage (FMA3): 1.237v, LLC6=unstable
Bios voltage: 1.345v, max load VR VOUT voltage (FMA3): 1.237v, LLC5=stable (better transient response, smaller guardband /less voltage dips below 1.237v in microseconds).

I have been using adaptive voltage since day 1 on my Hero XI. I simply like the power and heat savings.

Am I missing out on something by using Adaptive voltage the way that I do?

I leave everything on Auto except:

Set XMP
Sync all cores (at X51)
AVX Offset 1
Set Adaptive additional voltage to 1.3v
VCCIO 1.2V
SA 1.2V
LLC 6

Works like a charm. I don't touch all the other settings. Am I missing something? Is something throttling down that I don't see occurring? I have no heat issues on my 9900k.

mdzcpa wrote:
I have been using adaptive voltage since day 1 on my Hero XI. I simply like the power and heat savings.

Am I missing out on something by using Adaptive voltage the way that I do?

I leave everything on Auto except:

Set XMP
Sync all cores (at X51)
AVX Offset 1
Set Adaptive additional voltage to 1.3v
VCCIO 1.2V
SA 1.2V
LLC 6

Works like a charm. I don't touch all the other settings. Am I missing something? Is something throttling down that I don't see occurring? I have no heat issues on my 9900k.



No, you’re not missing anything if stable during AVX workloads.
13900KS / 8000 CAS36 / ROG APEX Z790 / ROG TUF RTX 4090

Falkentyne wrote:
Yes that test would be helpful if possible.
See if you wind up being stable in both tests (only changing AC/DC loadliine from 0.01 to auto and nothing else).
of course make sure you use manual voltage and a loadline calibration you know to be stable (e.g. LLC6 or whatever you choose).
Keep that sync ac/dc to vrm loadline disabled. I'm not sure if enabling it changes the AC DC loadline to match the VRM (LLC) in mOhms, or if it changes the vRM (LLC) to match the AC DC value or not, so lets try to get that out of the picture.

I've done repeated testing on a GB board and with a fixed LLC value (e.g. high), and manual voltage, there is no change at all in temps, amps draw or stability, when changing IA AC/IA DC to 0.01 mOhms or auto or the default (1.6 mOhms) for 9900K.

What does affect "VMIN" stability is loadline calibration. At a 'target' VR VOUT (CPU on-die sense voltage, which the Maximus XI boards all use now as vcore), you will require a higher VMIN for bypassing borderline stability or instability at higher LLC, than with a lower (more vdroop or steeper) LLC+higher bios voltage, when stressing heavy loads (especially AVX/FMA3!)

Example:
Bios voltage: 1.295v, max load VR VOUT voltage (FMA3): 1.237v, LLC6=unstable
Bios voltage: 1.345v, max load VR VOUT voltage (FMA3): 1.237v, LLC5=stable (better transient response, smaller guardband /less voltage dips below 1.237v in microseconds).


Alright I'll run the stress test tonight. It's been pretty hot today (Australia) and our evap AC has been struggling so I'm just waiting for it to cool down before I run the 1 hour stress test.

mdzcpa wrote:
I have been using adaptive voltage since day 1 on my Hero XI. I simply like the power and heat savings.

Am I missing out on something by using Adaptive voltage the way that I do?

I leave everything on Auto except:

Set XMP
Sync all cores (at X51)
AVX Offset 1
Set Adaptive additional voltage to 1.3v
VCCIO 1.2V
SA 1.2V
LLC 6

Works like a charm. I don't touch all the other settings. Am I missing something? Is something throttling down that I don't see occurring? I have no heat issues on my 9900k.


Silent Scone@ASUS wrote:
No, you’re not missing anything if stable during AVX workloads.


If your CPU can do 5.1 with those settings under heavy AVX loads you have a waaaaay better CPU than me. Mine cannot go above 5Ghz as it's pretty much hit a temperature wall. I've tried 5.1 with an AVX offset of -1 and it can't even run cinebench lol.

Zammin wrote:
Alright I'll run the stress test tonight. It's been pretty hot today (Australia) and our evap AC has been struggling so I'm just waiting for it to cool down before I run the 1 hour stress test.

If your CPU can do 5.1 with those settings under heavy AVX loads you have a waaaaay better CPU than me. Mine cannot go above 5Ghz as it's pretty much hit a temperature wall. I've tried 5.1 with an AVX offset of -1 and it can't even run cinebench lol.



His sample might, my own personal sample isn't so great. I'm in the 1.35v for 5GHZ club 😉 (No AVX offset)
13900KS / 8000 CAS36 / ROG APEX Z790 / ROG TUF RTX 4090

Silent Scone@ASUS wrote:
His sample might, my own personal sample isn't so great. I'm in the 1.35v for 5GHZ club 😉 (No AVX offset)


I think I might be right there with you lol, mine might even be worse in light of the results of my test just now.

This is gonna make me look kinda stupid, but I just ran RealBench with the loadlines (IA AC/DC) set to auto, as well as with my usual settings of 0.01. Both tests hit a BSOD within 10 minutes. This blows my freakin' mind because just a few days ago, when I was dialing in the settings, I ran multiple 1 hour passes of RealBench and OCCT and it passed every time with these settings. Now it fails. Either I haven't done something right or my CPU is rapidly deteriorating lol

If I had to guess, I've probably not done this right and there must be some better settings I can use. I can't really increase the voltage as I'm pretty well up against a temperature wall with these settings even with liquid metal between the die and IHS, and a dual rad custom loop. The only way I can increase the voltage is if I can get the temps down somehow. I can give Shamino's suggestion for adaptive mode a try but if I can't get it stable that way I'm stumped on what to do. Might have to aim lower and settle for 4.9Ghz..