P955ER throttling

Making a run of the same now.

 

CCC v 2.0? Or an earlier version? It seems you've been to this thread:

http://forum.notebookreview.com/thr...ter-2-0-is-the-one-causing-throttling.819744/

 

Control center v 2.x. The current one offered by Obsidian's updater, et al.

 

No, the funny part is, the Control center did apparently lower the throttling. Now you're back to powerlimit throttling.

Your temps are back to normal, however now you have much less FPS in your games. In Bf1 you lost roughly 30FPS.

basicially installing Control center + 120W powerbrick does the same as you have right now. Look at your clockspeeds + Wattage and compare before and after Control center. It seems the controlcenter does overwrite the powerlimit throttle set by the EC when inserting a bigger PSU, however this makes the notebook overheat because it's cooling cannnot keep up.

So all you did was get a bigger PSU, to uninstall CCC while then getting exacly the same as you got before you purchased your PSU. Your CPU went as low as 2.7ghz on BF4.

 

Do you have the new version of CCC 2.0 though, or the original version? As @jclausius stated, that thread covers a lot of topics and it's been proven and discussed here, as @Danishblunt stated, that the limit is in the EC and not the PSU rating.

CCC will change the behaviour depending on the load type and programs on occasion, the chassis has to abide by the MAX-Q limitations for noise and so on. This is why when you remove CCC, or change the performance setting in CCC, you will get different performance results in Firestrike for example. Prime and Furmark will also react differently, Witcher is also behaving differently. The notebookcheck review update of our laptop showed that the new version of CCC 2.0 solved some of this behavior, but it still isn't behaving exactly as we would like it to so there is a way still to go.

Were the previous AIDE throttle flags caused by power limits, temperature or both. The Obsidian fan control will only affect the temperatures but on full load they should result in the same temps as the CCC. But CCC could affect the power throttling.

 

I don't wish to put you on the spot, but can you unofficially say if there will be updates coming in the next few months? If you can't say, then understood.

 

Sure - there is already a second version of CCC 2.0 ( v 1.0.34 ) is out and it's the version I believe we have on our download page, which addresses some "issues" and that's what notebookcheck tested. I mentioned in the CCC thread that 1.0.34 still isn't ideal, but we will keep working on the topic and hope that further updates come along with additional changes that we are discussing with Clevo. At the end of the day, MAX-Q chassis have to meet the Nvidia requirements in order to use MAX-Q "hardware", the manufacturer has to use what Nvidia provides in order to meet the MAX-Q levels in terms of fan noise and the rest. It's a juggling act for everyone, but more so for some manufacturers than others for reasons widely discussed on nbr!

 

Also, what do the '(!)' mean in Core Temp on the Core temperature entry? See the screenshot below:

 

That doesn't seem to be quite right, especially for a laptop that has been repasted with liquid metal. The i7-7700HQ in my P950HP6 doesn't get that hot despite it having the same TDP as your i7 8750H. Bear in mind that my laptop has been repasted with EK-TIM Ectotherm which performs similarly to Arctic MX-2.

 

Check out the wattage if you really use around 45 Watts. (scroll a little down when using handbrake on HWinfo64)

 

It seems to be using 40 W. That 5 W discrepancy still doesn't explain why I'm getting temperatures 10 C lower than his, especially considering the fact that he's using liquid metal and I'm using Ectotherm.

 

5W is more than 10% more heat


Liquid metal does not improve temps above a certain point anymore because the bottleneck is not the conducivity, but instead the fan/heatsink performance. He might have a bad fit and might get better temps with another paste such as ICDiamond.

Also I googled your thermal paste, it's high quality. Conductivity of 8.5W/MK is on par wich thermal pastes such as GC-Gelid.

Here is the link to the paste

 

I think he probably just has a bad heat sink fit. Oh, and the EK-TIM R&D chief said that the thermal paste performs similarly to MX-2: https://www.overclock.net/forum/61-...erm-enclosed-waterblocks-any-experiences.html

 

Maybe they made it better or maybe he mispoke and meant Mx4, idk.

It's not a bad paste by all means tho. Seems solid enough on a notebook.

 

You're correct.

I went back to MX-4 (the thick stuff) and my temperatures literally dropped!

What devilry is this?

PSA: This P955ER isn't meant for LM use. That issue with the tripod heatsink design may be what kicks in with LM: Too loose of a contact surface for efficient thermal dissipation, resulting in rapid LM degradation.

Cleaning LM is quite easy though. The residue (black-ish material) was minimal.

 

Switching back to MX-4 literally gave me better thermal performance.

Welp, that weak contact on the Q6 is its' own downfall. Let's see how a copper shim mod (welded on) will perform.

 

Just stick to GC-Extreme and call it a day. It's not a good idea to weld stuff to your heatsink; if the shim makes poor contact with the heatsink surface you'll need to replace the entire heat sink.

 

As I advise people, liquid metal needs perfect contact to help.

 

Where's this data come from, especially the stock paste result on Max Fan mode?

 

Tried a reverse image search and it's uploaded. No hits.

 

Right, I nkow this isn't why you posted the image but it's what I picked up on; the "stock thermal paste" figures were taken as the system was out of the box so to speak, then the CPU was delidded and then repasted to get those paste job results. Hence the massive difference - reviewer should have really mentioned that in the graph as it's pretty missleading.

 

Any reason to choose those over Kryonaut?

 

This has been stated dozens of times, but sure one more time....

Due to mediocre heatsinks with ill fits with laptop heatsinks, and often not having proper amount of downward pressure from the mountings, a thick paste that doesn't suffer from degradation of thermal transfer capacity when applied too thick, and is also non-conductive for electricity, is an ideal TIM for laptop applications. Gelid Extreme and ICD7 fit this category.

 

I've not seen kryonaught have issues with pumping.

 

Thanks, yeah I posted my quesiton too early and saw those answers afterwards. However I tried both MX4 (which I believe should be the same as GC-Extreme) and Kryonaut and got much lower temps with the Kryonaut, so hence I was unsure.

 

GC Extreme seems quite a bit thinner than ICD. It more closely resembles the stock compound in terms of viscosity, IMO, even if it is generally better at heat transfer.

 

Pumping is also impacted by what the compound does with temperature variations in terms of viscosity and volume.