Need guidance undervolting i7 9750H with ThrottleStop

[AOne]

Pretty good result. Congratulations Now I'm jealous and start thinking to repaste mine again, cause it was like that in April, but lately the Cores' differences are increasing up to 11 degrees in tests and the temp is 2-3 degrees up. And it was just as yours in the beginning. Strangely, all my core temps are pretty close and only one couple is lower(!) than the others.

 

[unclewebb]

I'm seeing some crazy drop in temps.

I have not done any hands on testing but another user in this forum had great results with Noctua NT-H2 on his laptop. What he found was that it was much more consistent during long term use. No need to re-paste every other day.

Many users with mobile CPUs report the same thing when using MX4. It may work great day one which all of the review sites show. Long term on a mobile CPU, not so great.

Hopefully your fantastic results last longer than a week or two.

Edit - Now I remember. That other user that had great results with Noctua was @AOne

 

[nguyen]

I have not done any hands on testing but another user in this forum had great results with Noctua NT-H2 on his laptop. What he found was that it was much more consistent during long term use. No need to re-paste every other day.

Many users with mobile CPUs report the same thing when using MX4. It may work great day one which all of the review sites show. Long term on a mobile CPU, not so great.

Hopefully your fantastic results last longer than a week or two.

Edit - Now I remember. That other user that had great results with Noctua was @AOne

Funny that I bought TF-X and NT-H2 at the same time but didn't get to try NT-H2 yet since I tried TF-X first and the result are too good .
Has been 3 months since I applied TF-X and there is no sign of pumping out, the other user @budgetgaming that I recommended TF-X to didn't get good result with it though.

Pretty good result. Congratulations Now I'm jealous and start thinking to repaste mine again, cause it was like that in April, but lately the Cores' differences are increasing up to 11 degrees in tests and the temp is 2-3 degrees up. And it was just as yours in the beginning. Strangely, all my core temps are pretty close and only one couple is lower(!) than the others.

Maybe you can try putting your laptop under extended load, then tighten the heatsink screws when it is still warm (the laptop that is).

Yea I'm really pleased to see the core temps so close to each other. Used to really irk me to see a 20C difference. And I'm honestly surprised to see the max power draw at 81W. Used to only see a max of 71-72. Though my PL1 is set to 70 and PL2 set to 90.

Yea I've saved the log file so I can use it for later comparisons. I really am very happy with the current core temps.

And yea that makes sense. Lower temps, less pump out. I'm hoping these figures stay as they are so I don't have to repaste again.

On a side note, I am quite disappointed by the MX4 though. Many reviews I read had only good things to say about it. Feels like I got conned. Lol.

Your earlier log indicated the CPU was thermal throttling at 70W, so the end result is that you get >15C cooler while the CPU is pulling 10 extra Watts. Pretty impressive just for a repaste eh.

 

[unclewebb]

...didn't get good result with it though

Every situation is unique. A paste that works well on one laptop might work like crap on another. That is why buying a handful of tubes and doing some testing on your laptop is always the best idea. If your paste pumps out after a week and your temps are going up and up, that is probably not the best paste for your laptop.

 

[nguyen]

Try a suicide run with PL1/PL2 of 90/105W on my laptop



TF-X and my laptop are meant to be together

 

[AOne]

I actually tried retightening them two days ago as I was replacing my secondary SSD with bigger one (and blowed out all the dust), but to no avail. I'll wait a little longer and maybe repaste finally, but so far no temps over 91C in CB20 so, I'll postpone it. I can't figure it out weather the last software updates are not behind this or it's just the pump out effect for the last half a year of extended use and lots of heat/cool cycles.

Try a suicide run with PL1/PL2 of 90/105W on my laptop

View attachment 173419

TF-X and my laptop are meant to be together

Wow! My limits are 90/70, but I've never seen more than 68 used. And yes, I'm with 9750, the predecessor of yours.

 

[nguyen]

I actually tried retightening them two days ago as I was replacing my secondary SSD with bigger one (and blowed out all the dust), but to no avail. I'll wait a little longer and maybe repaste finally, but so far no temps over 91C in CB20 so, I'll postpone it. I can't figure it out weather the last software updates are not behind this or it's just the pump out effect for the last half a year of extended use and lots of heat/cool cycles.


Wow! My limits are 90/70, but I've never seen more than 68 used. And yes, I'm with 9750, the predecessor of yours.

Did you lower the Turbo Ratio Limits for 6 cores load ? your CPU probably reaches max core clocks before reaching PL2 of 90W.

 

[AOne]

Did you lower the Turbo Ratio Limits for 6 cores load ? your CPU probably reaches max core clocks before reaching PL2 of 90W.

No. All is set as default (maximum).

 

[nguyen]

No. All is set as default (maximum).

Well try prime95 small FFT then you will see your CPU pulling 90W .
Btw the 10875H is 8cores/16 threads CPU, your sucessor is the 10750H CPU, bloody Intel couldn't improve anything and just rename 9th gen CPU to 10th gen.

 

[Ic3m4n86]

Pretty good result. Congratulations Now I'm jealous and start thinking to repaste mine again, cause it was like that in April, but lately the Cores' differences are increasing up to 11 degrees in tests and the temp is 2-3 degrees up. And it was just as yours in the beginning. Strangely, all my core temps are pretty close and only one couple is lower(!) than the others.

Thanks man. I am really pleased with the results. I owe it to @unclewebb for brainwashing me into improving my repasting adventures.

I have not done any hands on testing but another user in this forum had great results with Noctua NT-H2 on his laptop. What he found was that it was much more consistent during long term use. No need to re-paste every other day.

Many users with mobile CPUs report the same thing when using MX4. It may work great day one which all of the review sites show. Long term on a mobile CPU, not so great.

Hopefully your fantastic results last longer than a week or two.

Edit - Now I remember. That other user that had great results with Noctua was @AOne

Hahaha. Coincidence eh? But as you said I do hope the repaste lasts longer than a week or two. As the temps are lower by more than 10C I certainly hope it can last at least 6 months. I don't mind repasting every 6 months rather every week. LOLOL.

Funny that I bought TF-X and NT-H2 at the same time but didn't get to try NT-H2 yet since I tried TF-X first and the result are too good .
Has been 3 months since I applied TF-X and there is no sign of pumping out, the other user @budgetgaming that I recommended TF-X to didn't get good result with it though.

I'll see how the Noctua works for the next couple of weeks. Fingers crossed I don't need to buy the TF-X instead.

Your earlier log indicated the CPU was thermal throttling at 70W, so the end result is that you get >15C cooler while the CPU is pulling 10 extra Watts. Pretty impressive just for a repaste eh.

Oh you have no idea!! I'm beyond ecstatic!! Thanks brother.

Every situation is unique. A paste that works well on one laptop might work like crap on another. That is why buying a handful of tubes and doing some testing on your laptop is always the best idea. If your paste pumps out after a week and your temps are going up and up, that is probably not the best paste for your laptop.

Thank you guru @unclewebb for the constant guidance and encouragement!!

Try a suicide run with PL1/PL2 of 90/105W on my laptop
TF-X and my laptop are meant to be together

90W is properly insane brother!! I wished I had better cooling so I could try that too.

 

[AOne]

Pretty good result. Congratulations Now I'm jealous and start thinking to repaste mine again, cause it was like that in April, but lately the Cores' differences are increasing up to 11 degrees in tests and the temp is 2-3 degrees up. And it was just as yours in the beginning. Strangely, all my core temps are pretty close and only one couple is lower(!) than the others.

Ok... so, there was a Windows update and NVidia drivers update today. Suddenly thought to myself, why not make some test and .... was quite surprised. All temps in CB20 and TSBench were 7-8 degrees down. Nothing else had changed and the room ambient temp is the same. Even noticed a bug in my Armory Crate Software (the Asus' one, responsible for CPU/GPU/Fans tweaking) gone away - it was holding my GPU over 1215MHz no matter used or not (before it was instantly falling down to 300, when idle). Strange, isn't it? Even my points are not close enough to the 3140-3150+ I used to achieve before in CB20?

 

[5ha5h]

Thanks to this thread I managed to undervolt my Laptop, and I would like to thank you all for the help. It still has warranty, that is why I didn't repaste it yet, but the result is really great.
Big thanks again,
have a really nice day