Confused about "Power Limit Throttling" XPS 9560

I thought I've read it all to understand, but inevitably I am not completely clear. Right now through XTU I am undervolting -0.1v.

With power plugged in and power plan on "max performance", when I render a video in davinci resolve my CPU hovered around 85c and peaked at 96c. It didn't thermal throttle once, but jumped into "power limit throttling" and basically stayed there until the end. It *seemed* like things were still running smoothly?

https://i.imgur.com/AnYdJiD.png

My laptop was sent into a depot and the motherboard was replaced, so whoever that technician was would have done a repaste (right?) which I can only assume is better than the factory.

Working with photoshop, lightroom, video editing, I definitely want to get the performance out of my laptop and not take a step back by doing a bad paste job or whatever else.

At this point since it power throttles but doesn't thermal throttle - is there a problem/need to repaste+thermal pad the mosfets?

 

Peak 96c CPU is too high, you should repaste, eighties are okay.

Power limit throttling mostly kicks in when the CPU and the GPU are loaded at once, because the VRM area overheats, as it is not cooled actively. It is difficult to avoid it entirely, but depends on the actual workload. Firstly try the other easy things, like undervolting also the GPU with MSI Afterburner and padding the VRM mosfets. If you're not happy with the outcome, you may want to try iunlock's mod.

 

Main reason for PL1 is the chip's hit 45w and that is its limit, end of the day it's a 65w desktop chip capped at 45w for the laptop market

 

What's causing it to try delivering more power or deciding it needs to cut it? Is that to say that a repaste/vrm pads wouldn't change the situation?

 

I don't know how to explain it more simply than I did but I will try, it's a 65w chip that's capped by hardware at 45w by design (with it being in a laptop that can't cope with 65w and also it helps battery life) so when it hits 45w it triggers the PL1, nothing to worry about especially if you are encoding.
They don't make 45w chips for the laptop market, they just change the TDP setting from 65 to 45 during manufacture.

I would be more worried about what Sem said, your temps are in the danger zone and I would expect to see event viewer logs saying the speed is limited due to a thermal event.

 

As much as I understand, the power limit issue is that the power gets limited by Intel DPTF to less than 45W because the limit on an "ambient" (VRM area) sensor temperature is exceeded. Blowing some cold air into the the small central grille in the gap between the display and the keyboard tends to cause throttling to cease.

Regarding the 45W power cap, I'm curious how the higher-clocked Xeons in the 5520 live with it.

 

Hmm, repasted with the arctic silver 5 (I think/hope it all went well!)

I'm finding idle temperatures basically the same. Rendering the same video clip, the temperature seemed to rise slower, and the period of time there was throttling was shorter, but it still peaked at 93c.

Seen comments of other people using the arctic silver 5 and it working well... hmmm !

 

I think it is due to the xeon being a higher quality chip that is built to run lower TDP (wild guess)

It will idle the same but the fans will not kick in as much as it can get the heat to the heatsink more efficiently yes it will eventually creep up if the CPU is being hammered during a long operation. AS5 takes a while to bed in unlike the newer pastes such as MX2/4 so you will find over the next couple of weeks it will get a bit better. If you want better results to try undervolting.

The thing with laptops is it will try to balance cooling, unlike overclocked games rigs that are set to run the fans flat out all the time with better cooling from the start.

 

Thanks for your insights!

 

93 peak is high. I wouldn't think AS5 isn't good. More likely the heatsink doesn't sit tightly on the chips, maybe the alignment isn't perfect. Now that you know how it's done, you could repeat the exercise and see if you happened to get air bubbles or if there is a thick layer of the paste on the chip somewhere. Btw are the 4 core temperatures under Prime95 roughly the same or not?