Achieving maximum parallel performance

EndeavorOS, Intel Core Ultra 5 125H

I’ve been reading about thermal management and very nice strategies to limit fan noise and battery drain, and they’ve been very useful!

Now, I’d like to achieve the opposite.

For work, I need to compile the LLVM project with some frequency. It takes a long time, and uses every core. I’ve noticed, however, that just letting everything run in every core underutilizes the processor; the base frequencies are never reached and temperatures never go over 80˚C. By using taskset to restrict the compiler calls to the P cores, I get them to reach 2.6Hz instead. Temperatures reach 84˚C.

Are there ways to reach higher temperatures / forcefully let the processor run at full base clock / increase the amount of time it boosts? I’d like to squeeze as much parallel performance out of this thing as possible.

Freezer, low humidity freezer. Seriously. Some sub -40c winter air, depending on where you are.

Or, (I did this) have an air conditioning unit blowing into it from the intake.

It’d be difficullt to fit inside of the freezer together with the laptop to get some work done :rofl:

Also set the EC to run the fanduty at 100.