undervolting i7 14700HX

[Grumix]

Hi everyone! I finally give up and seek for your help after a bit over a month of researching on this forum and others on how to undervolt this cpu properly after I bought a lenovo legion pro 5 16irx9.
My main objective would be to lower temperatures without having a big impact on performance or disabling turbo.
First thing I thought to lower the numbers on Long Power PL1 and Short Power PL2 (lower the watts-lower the temperatures).
I reach for those numbers without BSOD but I get some errors in HWiNFO
Indeed, lowering the watts lowered the temperatures but I get those errors in HWiNFO and I worry on the long term if it snaps something.
I even get some higher numbers on CPU Core -150,4 mV, CPU P Cache -145,5 mV and CPU E Cache -145,5 mV but those errors in HWiNFO somewhat increased a bit. (still no BSOD after running some games)
The thing is I get random BSOD when I'm downloading games over steam or when I start the laptop, it restarts one more time when the desktop appears...and that's it.
My last settings are these:
If it matters, I bought a high-end cooler pad but I use it exclusively at home. Every other day I take the laptop in different locations without cooler pad to game or to browse the internet.
Any help is appreciated, thank you!

 

[unclewebb]

WHEA errors are bad. If HWiNFO reports WHEA errors, your undervolt settings are not stable. Reduce your undervolt until there are no errors and no blue screens.

Start by setting the E Cache back to default, +0.0000. Most people do not see any improvement by adjusting this voltage. Undervolting this might cause instability without improving anything.

Are you using the V/F Point feature? This is a must. At 800 MHz, set V/F Point 1 for both the core and the cache to 150.

Lowering the turbo power limits like you have done can cause instability. A hot CPU is better than a CPU that is not stable. Increase the power limits back towards where they originally were. When MMIO Lock is checked, do not check the Sync MMIO box.

Reduce the core and cache undervolt to -100.6 mV. Test for stability and watch for WHEA errors at this level. Spend some time at this level. If there are any errors then try increasing the voltage. Do not try to come up with the perfect voltage settings in one day. You have to use your computer as you normally would, play a few games, and see if there are any WHEA errors or blue screens. You should never be seeing either.

Check the ThrottleStop Log File box on the main screen and run a log file while playing for at least 15 minutes. When finished testing, exit the game and then exit ThrottleStop so it can finalize the log file. Attach a log file to your next post. Also post updated FIVR and TPL screenshots so I can see what changes you have made.

Consider setting all of the turbo groups to 51. This will sacrifice some light load performance. You will probably be able to come up with a much more stable undervolt with less heat, fewer voltage spikes and fewer errors.

There a link to the Valour549 Ultimate ThrottleStop Guide that is included in the ThrottleStop folder that you downloaded. It shows the basics of setting ThrottleStop up on a 14900HX. It is well worth reading this info.

 

[Grumix]

Thank you for answering, unclewebb! Indeed, with these settings I had no more errors in HWiNFO, only "problems" I had was the temperatures even with the cooling pad. I attached below the pictures and the log file you requested. I will keep testing in the upcoming days and if no more blue screens and errors appear I may raise core and cache undervolt gradually.
When using TS Bench only messages I got thru Limits tab were BD Prochot and THERMAL switching between red and yellow.
In V/F Point feature I managed to set 150 in both the core and the cache
Temperatures doesn't really bother me that much, I still have a Legion 5 from 2021 with a ryzen 7 4800H. That cpu managed to run with 100°C in all games and still survives to this day.
Update: I unchecked C1E, after that I've seen that all cores stays at 5,1Ghz
Edit: On CPU Package maximum temperature I even reached 106°C while gaming, is it normal that only 3 cores to reach such temperatures? I feel like my keyboard is boiling at one point XD
I will test this laptop tonight without cooling pad to see if smoke is coming out.lol

 

[unclewebb]

The log file shows that your Lenovo laptop has a BD PROCHOT throttling issue. It looks like it is being triggered by the CPU temperature. When the CPU core temperature gets up to 100°C, there will be a quick, less than a second, shot of BD PROCHOT throttling to try and get control of the CPU temperature.

   DATE       TIME    MULTI   C0%   CKMOD  BAT_mW  TEMP   NVIDIA GPU     VID   POWER
2025-02-10  10:35:20  49.79   29.7  100.0       0   93    2490    73   1.3550  102.1
2025-02-10  10:35:21  49.87   31.7  100.0       0  100    2490    73   1.2375  105.6   TEMP
2025-02-10  10:35:22  48.14   33.7  100.0       0  100    2460    73   1.3540   97.1   TEMP
2025-02-10  10:35:23  47.57   33.5  100.0       0   98    2460    73   1.3307   99.8   TEMP
2025-02-10  10:35:24  49.26   30.2  100.0       0   98    2430    73   1.3477  103.2
2025-02-10  10:35:25   4.94   30.3  100.0       0   74    2430    73   0.7451   20.2   BDPRO
2025-02-10  10:35:26  48.35   26.7  100.0       0   98    2490    72   1.3246   91.4
2025-02-10  10:35:27  49.56   30.0  100.0       0   97    2490    72   1.3342  102.7

BD PROCHOT throttling is the worst kind of throttling. The CPU goes from happily running at almost 5000 MHz down to 400 MHz for almost a second. The log file shows a big drop in power from 100W down to 20W. I can only imagine that a drastic drop in performance like this is going to cause a big FPS drop while gaming.

I recently bought a MSI Vector laptop that has an advanced BIOS where you can unlock and disable BD PROCHOT to avoid any ugly and unnecessary throttling. Most recent laptops, including your Lenovo laptop, have enabled BD PROCHOT in the BIOS and have locked out adjustment of this setting. That means ThrottleStop cannot be used to disable BD PROCHOT. Until someone comes up with a BIOS mod to disable BD PROCHOT, my advice would be to reduce the power consumption and maximum temperature of your computer. Avoid letting the CPU get up to 100°C and perhaps the drastic BD PROCHOT throttling issue will go away.

You could try setting PL1 and PL2 to somewhere in the 90W to 100W range. This will reduce maximum performance but if it helps keep the CPU from reaching 100°C, it might result in smoother overall performance.

Another possibility would be to play around with the PROCHOT Offset setting. Right now your TPL screenshot shows that this is locked to 0. If it is possible to unlock this, (you will need to reboot), try setting PROCHOT Offset to 5 or 10. This would tell the CPU to start thermal throttling at 95°C or 90°C. This might help avoid BD PROCHOT throttling. That is an important goal.

However you decide to try and fix this issue, run another log file and watch the far right column for any more BD PROCHOT throttling. Ideally there should not be any BD PROCHOT throttling.

Have a look in the BIOS for a BD PROCHOT setting. I have a hunch that this important setting has been deliberately hidden by Lenovo.

106°C

Did you see that big number in the Enhanced section of HWiNFO? I am not 100% sure if that information is correct. I am also not sure if the DTS information is 100% correct. The DTS info is capped at 100°C so even if a CPU core goes beyond 100°C, the DTS info will still report a maximum temperature of 100°C. This means the HWiNFO Enhanced data may in fact be correct. Usually thermal throttling kicks in right at 100°C and it would be very difficult for the CPU to overshoot the 100°C maximum by more than a degree or two.

Is this important? Not really. All I know is that you have to reduce the CPU temperature to ensure smooth performance with hopefully less BD PROCHOT throttling. It is possible that the factory did a crap job applying the thermal paste. Buy some Honeywell PTM 7950. Perhaps you can do a better job applying thermal paste than a bored assembly line worker can do.

 

[Grumix]

Thank you once again for this valuable information! At first, I thought that was something wrong with my internet or the server I was playing, not even managed to think about this BD PROCHOT. As you said, my performance dropped significantly for about one second in the game, multiple times.
I didn't see any BD PROCHOT in the bios, as you said, it is most likely hidden, but I managed in the PROCHOT Offset to set the value 5. Indeed, by doing this I didn't saw any BD PROCHOT in the log file and my game was much smoother. Throttling increased as expected.

On PL1 and PL2 I let the settings as they are (140 and 190) or it would be better to switch to 90W and 100W until I grab some Honeywell PTM 7950?

I have one more question which is bugging me out. I have left some liquid metal which I succesfully applied on my old Legion last week. Could I apply on this new Lenovo as well or it would be better with Honeywell PTM 7950? Would one of these two things resolve this temperature thing or I'm doomed?

Thank you once again!
The difference beetwen two log files, on the right the PROCHOT Offset is set to 5

 

[pavko222]

hi uncle. I have the same laptop only different processor and can get into the hidden Bios. I changed the PROCHOT Lock setting to Disabled and it seems to work

 

[Grumix]

hi uncle. I have the same laptop only different processor and can get into the hidden Bios. I changed the PROCHOT Lock setting to Disabled and it seems to work

Which version of bios do you have? btw it's the same cpu but more likely different version of bios

 

[unclewebb]

I changed the PROCHOT Lock

It is actually the Bi-directional PROCHOT that needs to be disabled in the BIOS. Can you try doing that and see if ThrottleStop shows BD PROCHOT as disabled?

BD PROCHOT disabled and locked will look like this in ThrottleStop. There will be no check mark in the BD PROCHOT box.

@Grumix

Great to see that my suggestion to change the PROCHOT Offset value worked to get rid of BD PROCHOT throttling.

Some people have a fear of using liquid metal in their laptops. You are obviously comfortable doing this so give it a try. I do not know what would work better. I think liquid metal is usually best but I have no hands on experience comparing liquid metal to PTM 7950.

No need to adjust the power limits. Thermal throttling at 95°C is keeping your laptop from having a melt down.

Edit - Just noticed in your first log file that even when BD PROCHOT does not show up in the log file, you can still see the huge and regular drops in performance. That makes sense when BD PROCHOT is being enabled for less than a second at a time. Dirty dogs at Lenovo trying to pull a fast one hoping that no one will notice. Busted!

Now that you are getting control over your laptop, maybe you can try bumping the undervolt to -125 mV. I prefer using the High Performance power plan. Less latency and you can usually get away with reliably using a larger undervolt because the MHz are not jumping up and down.

For all the people that think using the High Performance profile in a laptop is some sort of crime, it is hard to argue with these low temps and power consumption numbers when fast and idle.

 

[harooniz]

Hi,

I thought I'd share my setup in case anyone comes around looking (I also don't mind recommendations in case there are any improvements to be made). I have the same laptop as Grumix.

Also those who do have a Lenovo laptop, I strongly suggest that you utilize Lenovo Legion Toolkit (Instead of Lenovo Vantage. It has the same features of Lenovo Vantage with the benefits of no background services, less memory hog, limited CPU usage, and contains no telemetry or marketing). You can adjust how you want the discrete GPU to work, adjust the fans accordingly and power levels just to mention a few things.

My PC runs stable at these settings and temps usually stay around 70-85C at most. It can sometimes peak to 90 in games, but it never stays there. Only change I made recently (Today in fact) I upped the PL1 to 90 from 57 (Datasheet shows shows 55 \ 157) and unclamped it. I did notice my BD PROCHOT is greyed out too, but this has not been an issue in my case. Haven't applied any paste yet since I bought it in November 2024.

 

[pablosky]

Hi, it's a relief to find this forum and see I'm not the only one who could get grilled with this processor :´).

I have a problem with the P-cores, and it's that two of them run extremely hot when gaming, 95-7º constant and 100Wº peak. If I lower the PL1/2 enough (60-70W) I avoid it, but I lose a lot of multi-core performance (in multi, 28T, I can reach 100W without much thermal throttling). It would be really nice to be able to set power limits individually for P and E cores.

I switched to a Noctua NT-H1 without violating the warranty, and the fans with a custom curve with hardly any change. All this test without using the 4070, imagine the temps in the P-Cores playing... what a oven.

Also idk if it´s normal but some P-Cores didn´t even reach 82 deg while playing, mabye some silice issue.

Greatings!

 

[unclewebb]

If you want to use ThrottleStop to control the CPU voltage, you have to disable Windows core isolation memory integrity. Reboot and run msinfo32 to make sure Virtualization Based Security is disabled. Delete the ThrottleStop.INI configuration file and start again. Post an updated FIVR screenshot.

Follow the guides in my signature if you have any problems.

The Offset column of the FIVR monitoring table shows that none of your undervolt values are being applied. ThrottleStop cannot do anything useful until VBS is disabled.

Edit - Use ThrottleStop 9.7.3

Noctua NT-H1

Hard to recommend using Noctua NT-H1 on a mobile CPU.

 

[pablosky]

Thanks for the reply, these two Windows security options were already disabled, it turned out to be virtualization, I can now adjust the voltage.

Edit - Use ThrottleStop 9.7.3

I have already switched to this version.

Hard to recommend using Noctua NT-H1 on a mobile CPU.

I am already aware of this phenomenon, even if it has good conductivity it is likely to happen in some months, I already have my eye on Thermal Grizzly Kryonaut or Kryonaut Extreme.

Anxious to lower those temps now, happy undervolting!

 

[unclewebb]

Did you delete the ThrottleStop.INI file? Your screenshot shows an odd ball number in the Voltage column of the FIVR monitoring table. You usually only see this if the ThrottleStop.INI file was corrupted because ThrottleStop was running when VBS was enabled.

Double check with msinfo32. Delete the ThrottleStop.INI file and shutdown your computer so all of the voltage registers are reset.

My computer shows the word Default in the Voltage column.

Honeywell PTM-7950 should be on your short list for thermal paste. It works well and works long term. Many other pastes do not.

 

[pablosky]

Yup, i deleted the old .ini when i managed to modify offsets.

Honeywell PTM-7950 should be on your short list for thermal paste. It works well and works long term. Many other pastes do not.

Yes, in my opinion, it's the best option: an easy and accurate for mobile chips. The problem is finding one that's authentic and not an Aliexpress clone.
If you know any reliable seller on Ebay/Aliexpress/Amazon/etc. it would be appreciated. (I live in Europe, specifically Spain)