ThrottleStop Settings i7-10750H

[aUserName12]

Hi all,
I have an Acer Predator Helios 300 PH315-53-71CM (i7-10750H, rtx 3070 8gb).
It was running really hot recently, so I cleaned out the fans, reapplied thermal paste and pads, etc. It's now running cooler, but I wanted to optimise it a little more.

Please see attached for my ThrottleStop settings - do these look okay to you guys? No crashes over the last few days, and no errors when running TS Bench at 960M, CBR 20.06, or OCCT. If I push the undervoltage much more, then I do start to get errors and crashes.
When running CBR, I get EDP Other under ring, and PL1 throttling under core after a period of time (as expected I think due to the TPL settings). When I run OCCT stress tests I get throttling pretty much as soon as it starts (VR Current under core, EDP Other under ring, and then eventually thermal & PL1 under core).
I have attached a log file: 09:27 time frame shows around Idle. 09:47 - 09:50 period is during CBR test, and 09:56 - 09:57 shows OCCT CPU stability test.

Any help/advice is greatly appreciated.

 

[unclewebb]

Your log file shows PL1 power limit throttling. Sometimes at 45 Watts or 50W or 60W or 70W. Acer Predator laptops typically program an embedded controller EC to dynamically adjust the turbo power limits. These turbo power limits are separate from the MSR and MMIO turbo power limits. ThrottleStop has no control over the EC power limits.

In the TPL window I would clear the Disable Controls box and I would check the MMIO Lock box which is near the top right corner of the TPL window. If you still see PL1 power limit throttling at a wide range of power values in the log file, that means you cannot use ThrottleStop to control this EC power limit throttling problem.

reapplied thermal paste

What did you use? A lot of thermal pastes that work well on desktop CPUs can quickly degrade when applied direct die to mobile CPUs.

If I push the undervolt much more, then I do start to get errors and crashes.

Always best to stay on the safe side. Some users spend infinite time trying to come up with the perfect voltage. It is better to add a few mV and not have to worry about a random crash a day or two from now. The best undervolt is one that never crashes in any stability test or game.

I like setting Power Limit 4 to the max and I also set IccMax to the max for the CPU Core, CPU Cache, Intel GPU and iGPU Unslice. Your present settings do not seem to be causing any problems so if you want, you can probably leave all of these as is.

Do whatever you can to avoid VR TDC throttling. Acer has used voltage regulators that are marginal. As they heat up during a stress test, they will likely reach a point where they cannot supply enough current to the CPU. Some throttling is inevitable when this point is reached. When playing games and not stress testing, the voltage regulators might be perfectly adequate.

 

[aUserName12]

Brilliant - thank you so much for the reply!

In the TPL window I would clear the Disable Controls box and I would check the MMIO Lock box which is near the top right corner of the TPL window. If you still see PL1 power limit throttling at a wide range of power values in the log file, that means you cannot use ThrottleStop to control this EC power limit throttling problem

I applied the suggested settings, and re-ran the benchmarks and stress tests - PL1 now only ever throttles at around 70W, so that worked brilliantly.

What did you use? A lot of thermal pastes that work well on desktop CPUs can quickly degrade when applied direct die to mobile CPUs

I used MX-6, as after some research, I found that it had quite a high viscosity and people mentioned that this would help stop the paste pumping out.

I have increased the undervolting a bit as I had an application (MatLab) crash yesterday, which is not normal in my experience.

I like setting Power Limit 4 to the max and I also set IccMax to the max for the CPU Core, CPU Cache, Intel GPU and iGPU Unslice. Your present settings do not seem to be causing any problems so if you want, you can probably leave all of these as is

I have changed all of these as suggested also.
My benchmark scores have gone up too: CBR @ 3246, OCCT Multiple threads -SSE is 363. Superposition at 8124.
Weirdly however, when testing the GPU in superposition, the GPU temperature reading in TS and superposition does not move at all from 45 C . Yesterday toward the end of the superposition benchmark, the GPU would be at around 68 C.
I checked in the Nvidia app, and the GPU temperature is being displayed correctly there. I did uninstall a couple of things today and install new graphics drivers - maybe it's something to do with that.

Again, thank you very much for your help.

 

[unclewebb]

PL1 now only ever throttles at around 70W

You might be able to bump that up further to 75W or 80W. This could cause more thermal throttling so 70W is a safe bet.

I used MX-6

Keep a close eye on temperatures. Many people have pump out troubles and really poor long term results when using MX-4 on mobile CPUs. If you notice temperatures are creeping up, order some Honeywell PTM 7950. The Honeywell product does not pump out and should have more consistent long term results.

the GPU temperature is being displayed correctly there

ThrottleStop gets its Nvidia GPU temperature data from the Nvidia driver. It is possible that this temperature data is not being updated frequently enough during a GPU stress test. I cannot remember having this problem or other users reporting this problem. It could just be, stuff happens.

Good to hear that your Predator is running like it should.