MSI GS65 CPU Core 4 is getting very hot

Hi everyone,

got my new MSI GS65 with the NVIDIA 1070MQ for a few days. Temps before repasting were very high, so I decided to use liquid metal and undervolted the CPU with -150mv. Everything went fine and the temps dropped drastically.

Now to the strange thing... All Cores are around 78-83°C degrees under load. I monitor all my 6 cores with MSI Afterburner. Today I tabbed out of Monster Hunter World and noticed that in XTU the package temperature of the CPU was 90-94 degreees. So I started HWinfo and saw that Core 4 was the only core which was above 90°C. All other cores are much cooler. Tabbed back to Monster Hunter World and MSI Afterbruner showed me that all cores had nearly the same temperature (Around 80°C).

Why are XTU and HWINFO showing that core 4 is so hot and why does MSI afterburner shows lower values?
If HWINFO and XTU show the correct value, why is core 4 uneven to the other 5 cores?

INFO: I saw that issue before and thought it was a bad pasting job, so I used a bit more LM. For the moment it seemed to work things out. But now Core 4 is getting hot again...

 

Here one screenshot of a shot Cinebench run of HWINFO values and one with MSI Afterburner values. This really confuses me...

 

Same problem as the "core 2" problem on 4 core MSI laptops (remembering the labeling is core 0 (top left), core 1 (bottom left), core 2 (top right), core 3 (bottom right).

In your cares core 4 is far top right and core 5 is far bottom right. The 6 core die is longer than the 4 core die, but its still basically the same format.

Convex heatsink combined with VRM's being above the CPU=less pressure on the top cores. And core 4 usually the hottest.
Notice that all of your top cores (0 2 and 4) are hotter than the bottom cores.

Difficult problem to fix until MSI stops using convex heatsinks.
Best thing to do is to :
1) take off the thermal pads on the CPU VRM's and CPU chokes (do not do this on the video card), and replace them with pads of HALF the thickness. Example: 1mm pads-->replace with 0.5mm pads. Or use thick thermal paste instead.

2) sand the heatsink flat. This would require a good sanding block like a Preppin Weapon sanding block and also requires that you remove the c-clips from the heatsink screws so you can remove the screws so they won't scratch the sandpaper. This is the ONLY way to fully fix this problem.

 

Thank you for your reply. Very usefull to know because I thought only Dell/Alienware has got issues with the 3 pod heatsink solution and uneven cores. So I will replace the original thermal pads with 0,5mm pads. Just ordered them and will keep this thread updated.

Another point is that all my cores a running on 3,9ghz all the time, even in windows and idle. Normaly there is a option to put the laptop in balance mode when pressing the "battery symbol" in the lower right corner. I do not have this option. All Windows Updates are installed. Did I miss something?

EDIT: Now I have the option. Just changed the power plan from high performance to balanced. Afterwards the trigger was available. Now powerplans can`t be selected anymore ^^

 

Well, the slider will become available when you switch your power plan profile to Balanced mode. And as for the multiplier that is running at the highest multiplier all the time, you can try clicking on revert to default and see if it gets fixed.

I'm using high power plan on mine, but I did some advanced tweaking to make the CPU work the way like it should

Mine has fulfilled the following items:

1. Core parking feature enabled so the multiplier works according to the multiplier you set for for the Turbo boost setting, that's the original concept of the i series CPU, the higher multiplier for less cores active only works if the cores are really turned off, and the default setting in Windows 10 is set to always on.
2. Speed Shift is working the way it should. My current setting is 63 (25%) for plugged-in mode and 127 (50%) for the Battery mode.
3. C-state can now be set as a profile in ThrottleStop. Currently I use C3 for plugged-in mode and C8 for Battery mode.
4. My range of tweaking for the multiplier in the Turbo mode can now go from x8 to x41, within this range I can set it anywhere I need, I don't need to try and limit it in the power option to go only down to x11 in the battery mode anymore.

The guide is still being written, I'll open a new thread for the guide once finished.

 

Just to be sure @Falkentyne ... You mean the thermal pad wich I have marked in the attached picture. And if there are more sticking to the mobo in this area I need to replace them aswell with 0,5mm pads?

At the moment I haven't opened the laptop because I'm waiting for the pads so I can't see how many pads there are

 

Seems to be solved now with new thermal pads (0,5mm) and a bit more LM on CPU! Will do some further tests now

 

Much better!

 

Nicely done!

 

I tried the same before, but then I found that the pads on the CPU regulators aren't contacting with the Copper plate when you put 0.5mm on them, I wonder if this will introduce any damage on long term use.

So I changed back to 1.0mm for my GS65 based on this concern.

 

No problems with 0.5mm pads on these laptops. The MSI 16L13 doesn't even come with heatsinks of any sort on their CPU VRM's (apparently the CPU heatsink helps cool the GPU VRM's instead) and no one has had magic smoke come out yet.

And I've had no problems running a 7820HK @ 4.9 ghz with 0.5mm thermal pads. The (zener diodes) trip the laptop after too much current is drawn anyway.

You absolutely won't be able to draw even close to enough current on a GS65 for any problems.

 

Thanks for the example, now I understand that the CPU VRM only gets warm and not up to the Critical level, it all makes sense now that why some don't even have heatsinks for the CPU VRM.

 

Finally did some further tests, and in PUBG the CPU on alle Cores is nearly even now (75-79°C). No peaks to 90°C or above anymore.

Maybe three things which I encountered during my new repaste:

1. The original bigger thermal pad for CPU VRM was not alligned correctly. Seemed like it was a bit to high attached and overlapped with the HS. Could have been also my original problem.

2. There are two tapes (thick ones) attached to the HS. One of them is going from above the HS down under the HS. For me I decided to place the tape only above the HS because I dind't see any reason to put the tape under it. It seemed that this could also generate a small gab between HS and mainboard.

3.There is a bigger thermal pad for the very small CPU VRMs (3 pieces). I dediced to leave it where it is.

Did some more LM on the CPU DIE and everything is fine now Very happy with the results.