X570 Aqua SOC Voltage Wrong

[Whiskeyomega]

Ive found that the soc voltage tends to be read wrongly in Hwinfo at the moment. I have adjusted my soc voltage to around 1.05-1.1 yet its still reading 1.2.
Ryzen Master reads this correctly. I should note I am running a experimental bios that isnt publicly available 2.03

 

[Martin]

Which exact value do you follow as the SoC voltage? Is it the "SoC Voltage (SVI2 TFN)" value in sensors?

 

[Whiskeyomega]

Which exact value do you follow as the SoC voltage? Is it the "SoC Voltage (SVI2 TFN)" value in sensors?

yes SoC Voltage (SVI2 TFN) . I know you just updated the aqua but thought you'd like to know about that one.

 

[Martin]

Is this problem only with the experimental BIOS?
It's quite strange as the SVI2 TFN value comes straight from the VRM telemetry and it has always been the most reliable one.
Please post the HWiNFO Debug File with sensor data for analysis.

 

[Whiskeyomega]

I think this was an issue on Ryzen master with 1.60 and above. I can no longer change bios' for the time being since getting my memory and fclk stable took so much work I do not want to spend 2 hours trying to get my pc to boot in order to even see. But I remember at some point others having issues with it reading 1.2 on Ryzen master as a whole. So I think something changed. Anyway https://drive.google.com/file/d/1qfzeoY6uM3MIzdwZy0MunXbVHF_4q0SI/view?usp=sharing Is the Debug file too big to attach here.

 

[Martin]

Thanks, so the above dump was made @ VSOC=1.2V ?
That file doesn't contain sensor data, so I can't check those details. But if it was @ 1.2V then it won't help anyway.

 

[Whiskeyomega]

Thanks, so the above dump was made @ VSOC=1.2V ?
That file doesn't contain sensor data, so I can't check those details. But if it was @ 1.2V then it won't help anyway.

Oh not sure how to include sensor data if its an option. My VSOC is currently 1.1 but hwinfo is still showing as 1.2

 

[Martin]

With Debug Mode enabled, just make sure to open the sensors window before closing HWiNFO and grabbing the Debug File.

 

[Whiskeyomega]

HWiNFO64.DBG

drive.google.com

Ok done.

 

[Martin]

Thanks. I analyzed the data and I'm not sure whether you want to trust what you see in BIOS/RM.
The system reports:
- VDDCR_SOC voltage SET = 1.10 V
- VDDCR_SOC voltage MEASURED (telemetry) = 1.20 V
So I'm not keen to change the value currently reported as measured to the setpoint instead.

 

[Whiskeyomega]

Yep I went and did some digging around in the bios. Turns out that if SOC/Uncore OC is on and "SOC/Uncore OC Voltage" is set it should be reading from that setting instead of "VDDCR_SOC Voltage" however it doesnt and mearly sends it has SET as you reported. I have now changed the VDDCR_SOC Voltage setting to 1.05 and its showing correctly in hwinfo. thanks again

 

[Zach]

Yep I went and did some digging around in the bios. Turns out that if SOC/Uncore OC is on and "SOC/Uncore OC Voltage" is set it should be reading from that setting instead of "VDDCR_SOC Voltage" however it doesnt and mearly sends it has SET as you reported. I have now changed the VDDCR_SOC Voltage setting to 1.05 and its showing correctly in hwinfo. thanks again

Know also that CLDO VDDP(memory controller) and CLDO VDDG(InfinityFabric) is CPU SOC derived voltages. So if your actual SOC voltage is 1.05V, the CLDO VDDP/G cannot exceed 1.0V. Always must be set at least 50mV lower than SOC voltage.
And what about the EDC limit set to 20A? Are you trying to limit power consumption (I doubt that cause the PPT is set to 300W) or to make the CPU boost more? If the second try setting it to 1A. Works for my little R5 3600 but with with marginal results (+50~100MHz) due to worst possible silicon quality.

 

[Whiskeyomega]

Know also that CLDO VDDP(memory controller) and CLDO VDDG(InfinityFabric) is CPU SOC derived voltages. So if your actual SOC voltage is 1.05V, the CLDO VDDP/G cannot exceed 1.0V. Always must be set at least 50mV lower than SOC voltage.
And what about the EDC limit set to 20A? Are you trying to limit power consumption (I doubt that cause the PPT is set to 300W) or to make the CPU boost more? If the second try setting it to 1A. Works for my little R5 3600 but with with marginal results (+50~100MHz) due to worst possible silicon quality.

Yes Im very aware of that at this point due to my bios and motherboard not exactly being smooth booting due to my memory being incompatible and also the voltages are way too high on Auto on my board. I run i run CCD & IOD and 0.02 less than soc. It doesnt need to be 50mv just doesnt exceed it and no my EDC is broken with the PBO i am running due to AGESA 1004. Its actually around 180A from what ive seen.
Trust me ive spent enough time on this now to know whats going on.

 

[NIoSaT]

Hey,
I think I have similar Issues here. X570 Taichi 2.70 BIOS

I set VDDCR_SOC to 1.1V via RyzenMaster and RM shows 1.1V but HWInfo shows 1.2V

So if I understand you correctly 1.1V is set somewhere in the BIOS but my mainboard just ignores this and still uses 1.2V?

 

[Zach]

Hey,
I think I have similar Issues here. X570 Taichi 2.70 BIOS

I set VDDCR_SOC to 1.1V via RyzenMaster and RM shows 1.1V but HWInfo shows 1.2V

So if I understand you correctly 1.1V is set somewhere in the BIOS but my mainboard just ignores this and still uses 1.2V?

Most likely in BIOS is set to AUTO so the board it self sets it to 1.2V. If you want to lower it you should set it manual in BIOS to 1.05~1.08V (VDDCR_SOC voltage)

 

[Martin]

I think the problem is that the BIOS doesn't properly apply some manual VDDCR_SOC values.

 

[Zach]

Didnt @Whiskeyomega said (post #11) that he change it to manual 1.05V and then he saw it correctly?
Well, boards BIOSs all have flaws and bugs and all shorts of stuff, but what I read from users... AsRock's are the worst ones

 

[Martin]

Yes, 1.05V worked, but 1.1V didn't seem to.

 

[NIoSaT]

Sorry for the late answer.
I managed to get it working, it seems like one setting that sound like it would change the VDDCR_SOC does... nothing but another setting changes it correctly.

I ended up using 1.020V in one setting (it did not let me use 1.025V) and 1.25V in the other setting and it seems the 1.020V are dominant.
HWInfo reads 1.019V
RM reads 1.025V

 

[Cosminnn]

Now I'm confused. Where should I set manually SOC Voltage and really read it correctly on HWiNFO64? I am on Taichi X570 and strangely, I do not have an option to set SoC Voltage in RM ... only VDDG and VDDP.