Question IA voltage offset is stuck on 0.098v in HWiNFO for some reason - - - shows as working in Throttlestop ?

[placeboing]

EDIT: I have found that HWMonitor properly displays the offset! This seems to be a bug with HWiNFO on some Legion laptops! See here:

_______________________
Original post:

This is a problem that I am seeing repeatedly asked, but no true solutions have been given:

Question - IA Voltage Offset is different, while P and E cores reflects the undervolted voltage ?

Laptop is a Lenovo Legion 5i Specs i7-14650HX 4050 6gb 32gb ram 1TB ssd 80wh Li-Po IA Voltage Offset is different while P-cores and E-cores reflect the undervolted voltage. Is there anything wrong? Settings I changed: For High Performance (Plugged In) Speed Shift : On Speed Stop : Off CPU...

forums.tomshardware.com

https://www.reddit.com/r/LenovoLegion/comments/1c6lh34/need_help_with_undervolting/

https://www.reddit.com/r/LenovoLegion/comments/1eq5mxd/legion_7i_overvolted_by_default/

I have my Legion 7i, i9 14900HX, I have changed the necessary BIOS settings, I have uninstalled Lenovo Vantage.

If I change the CPU Core Offset Voltage in Throttlestop, it shows up as properly working in the Throttlestop window.

However, in HWiNFO (of course, I restart it), it ALWAYS shows 0.098 for IA voltage offset. (however, undervolting the cache displays as working fine in HWiNFO)
What is going wrong here?

I have tried rebooting, making sure Throttlestop runs first, deleted my Throttlestop settings file.

All necessary settings such as Undervolt Protection are properly set in BIOS.
Uninstalled Vantage, and do not have Intel XTU.
mvboost is NOT enabled in throttlestop.
Thank you for any help.

 

[uWebb429]

@placeboing

HWiNFO does not report the offset voltages in real time. If you use ThrottleStop to make a voltage change, you need to completely exit HWiNFO and restart it. After doing that, HWiNFO should report the same offset voltage value that ThrottleStop reports in its FIVR monitoring table.

HWMonitor updates the offset voltages it reports in real time but it has a small 2% error in the values it reports. The ThrottleStop FIVR window updates in real time and is the most accurate source of offset voltage information for Intel CPUs.

Edit - If you used the ThrottleStop mV Boost feature, you need to reset that. Check mV Boost, set it to 0 and press the Apply button. If you uncheck the mV Boost box without applying a 0 value, the CPU will be left at the previous mV Boost value that you were using. I am guessing that is why HWiNFO still shows a positive IA Voltage Offset value of 0.098V.

A fast Windows restart will not reset all of the voltage control registers on some computers. Try holding down the Shift key on the keyboard when selecting Reboot to do a full reboot and not a fast reboot. You might need to completely shut down and power off instead of rebooting to reset everything.

FYI - The 14900HX likes mV Boost values in the 150 to 200 range. Over 40K in Cinebench R23 is possible with good cooling and the right ThrottleStop settings.

https://imgur.com/Wom9OTN

View: https://i.imgur.com/Wom9OTN.jpg

 

[placeboing]

Thank you for your help.
The mvboost change did nothing for the HWiNFO value. I did full reboot and shutdown too. And I am always sure to restart HWiNFO.

However, the author of HWiNFO said that this is because of an issue, and he will update the program to display this differently:

HWiNFO showing 0.098v overvolt on some Legion laptops, despite undervolting - a bug?

I am completely new to this sort of stuff, so I'm confused if I've actually discovered a bug on my first-ever attempt at working with undervolting... First, I should probably mention that I have finally now discovered that HWMonitor properly displays the undervolt offset, as seen here: Here...

www.hwinfo.com

 

[uWebb429]

the author of HWiNFO said that this is because of an issue, and he will update the program to display this differently:

That is good news. The mV Boost and offset voltages should be reported separately. Hopefully he also starts reporting any changes to the voltage values as they happen without having to restart HWiNFO.

 

[placeboing]

That is good news. The mV Boost and offset voltages should be reported separately. Hopefully he also starts reporting any changes to the voltage values as they happen without having to restart HWiNFO.

Isn't it so weird though, that some Lenovos are overvolted at 58x, all of them by 98mv? This is just on the regular BIOS.

Of course thanks to Throttlestop, I have this completely under control. But don't I have to worry about these overvolt boosts occurring while the PC is still booting up? (btw, can it hit the 58x while booting, before Windows has loaded?)
Maybe now I will find how to limit the ratios in the BIOS so it just never hits 58x, since that's the only place the boost is applied, I guess

 

[uWebb429]

some Lenovos are overvolted at 58x, all of them by 98mv?

How are you determining this? Wait until HWiNFO is updated. Then you can use use the HWiNFO data to find out what is really going on.

But don't I have to worry about these overvolt boosts occurring while the PC is still booting up?

I think you are worrying too much. The 58 multiplier is likely rarely if ever used while booting up or during normal use.

Intel has said that the HX processors do not have the same degradation problems that the 13th and 14th Gen desktop CPUs have. This is likely because laptops do not have the necessary cooling so you cannot run an HX CPU continuously at high power levels like you can run a desktop CPU. Most laptops will need to thermal throttle if you try to run them at over 150 Watts for any length of time. Desktop computers with a decent AIO cooler can run at twice that power level indefinitely without overheating. That seems to be the main source of the degradation issue.

 

[placeboing]

How are you determining this? Wait until HWiNFO is updated. Then you can use use the HWiNFO data to find out what is really going on.


I think you are worrying too much. The 58 multiplier is likely rarely if ever used while booting up or during normal use.

Intel has said that the HX processors do not have the same degradation problems that the 13th and 14th Gen desktop CPUs have. This is likely because laptops do not have the necessary cooling so you cannot run an HX CPU continuously at high power levels like you can run a desktop CPU. Most laptops will need to thermal throttle if you try to run them at over 150 Watts for any length of time. Desktop computers with a decent AIO cooler can run at twice that power level indefinitely without overheating. That seems to be the main source of the degradation issue.

Yeah, I am totally new to all of this, so I don't know anything.
But, when I got the laptop, it was spiking up to 1.67v occasionally according to HWiNFO Core VIDS, while just browsing the web. This scared me.
So I learned about undervolting, and now it never hits high numbers. (also, big change from changing 58 to 56 ratio)
In my HWiNFO report, it says:

Voltage/Frequency Curve:

98 mV @ 58x

But if you still feel that I shouldn't worry, then I won't.

 

[uWebb429]

I am totally new to all of this

I did not fully understand your problem before but I think I do now. Here is some more info.

Intel's unlocked HX and K series CPUs allow software to fine tune the voltage at 8 different VF (Voltage / Frequency) points from 800 MHz to the max MHz. Does your HWiNFO report show values for all 8 frequency points? If a VF point is set to zero and is not using an offset, HWiNFO might not report anything for that VF point.

The ThrottleStop mV Boost feature allows users to make a voltage adjustment but only at the lowest 800 MHz VF point. Try setting mV Boost to 50, restart HWiNFO, run a report and see if HWiNFO reports this information correctly. Does it also report the +98 mV value at the 58 multiplier? You might have to reboot and set the BIOS maximum multiplier back to 58X before doing this test.

It is either Intel, the Lenovo BIOS or possibly some Lenovo CPU management software running on your computer that is guilty. The default VF point for the 58 multiplier should never be automatically set to +98 mV. A CPU should be 100% stable at default voltage. It is never a good idea to randomly add more voltage to cover up any stability problems.

1.67v

I thought your computer was using the default Intel voltage frequency curve with all of the VF offset points set to zero like they are supposed to be set to. The default Intel voltage curve is guaranteed to be OK for long term use. A computer that has jacked the 58X VF point up +98 mV might not be OK for long term use. That is running the CPU at the high end of the Intel spec or perhaps a little beyond. Reducing the maximum multiplier to 56 was a good idea. This can help reduce the maximum voltage.

Good work digging into this issue some more. You are obviously not the only one that has this problem. Perhaps a new feature can be added to ThrottleStop to allow Lenovo HX owners to zero the highest VF offset point.

 

[placeboing]

Try setting mV Boost to 50, restart HWiNFO, run a report and see if HWiNFO reports this information correctly. Does it also report the +98 mV value at the 58 multiplier?

Voltage/Frequency Curve:

50 mV @ 8x, 98 mV @ 58x

This is all I see, so I guess the rest are at 0

Perhaps a new feature can be added to ThrottleStop to allow Lenovo HX owners to zero the highest VF offset point.

Cool!