@Michael_Wu thanks, I was able to pull the program from the RPM.
To do so, I did the following:
yay rpm-tools
rpmextract.sh ./nameofthermald.rpm
cp ./usr/bin/thermald-set-pref /usr/bin/thermald-set-pref
That then had it in the correct place, and I’m not entirely sure it’s fully working, but I’ll test some more and report back.
Edit: Specifying the max temp didn’t appear to be acknowledged, so I will try doing a reboot.
Edit 2: Doing a reboot seems to have gotten the settings to stick, but now it is thermal throttling when running s-tui with default settings, and slowing down to 2.5GHz and running at 65c as a result.
Using the “Better battery” power plan, I can easily hit ~90-100C and 28W+ CPU package power (HWiNFO acctually says the max is ~102C/38W) for more than a few seconds during my last Aliens: Fireteam session. This is after I did a repaste using some Noctua NT-H2.
Given the rather confusing way that the mobile CPUs handle thermal management, can anyone recommend some way to have an onscreen alert (possibly an audible one as well) if the CPU is doing thermal throttling? If it’s hitting the upper limits due to dust in the heat sink or the fan starts to go bad a few years down the road, I’d want to know about a cooling problem sooner rather than later. Having it trigger at a fixed temperature wouldn’t work here unless bad cooling would result in it rising above 100C when it otherwise wouldn’t with a clean heatsink and fully working fan.
I had an issue (batch 3) with my CPU fan that resulted in the support team sending me a new CPU cooler. When I replaced the CPU cooler I re-pasted the cooler using the same paste footprint that came from the factory but replaced the paste with Thermal Grizzly Kryonaut Extreme that I had left from my desktop build. Not sure that would make any difference vs the OEM Shin-Etsu but under a full load stress test my CPU temps don’t push past 72C anymore. Before my fan was replaced it used to hit 100. When I pulled the faulty cooler unit off, the application was not uniform. I can maintain sustained loads reliably now and all though the fan ramps up, the temps never surpass that 72C mark.
I have a batch 5 i5-1135G7. Just ran Furmark and Prime95 at full tilt for 15 minutes straight. Both GPU and CPU topped out at 67 degrees consuming a total of 50W per my calculations (GPU-Z isn’t entirely accurate on these integrated chips). That means the cooling solution is likely good for up to 65W going full blast. What’s probably happening is the fans ramp up slowly, allowing the CPU to go all the way up to 100 before they throttle back down because most workloads aren’t exactly like Furmark or Prime95 - they don’t just constantly blast every square millimeter of silicon with heavy-duty multithreaded workloads, even in 2021.
Apparently I spoke too soon. I’ve had all 4 cores (1165G7) hit 100 in the past few minutes, while CPU utilization hasn’t exceeded 40%. The laptop has good airflow (it’s in a vertical stand) and ambient temp is ~21. What’s up with this?
You were hitting 100C at boot? Are you on the most current BIOS (3.07)? That seems really high. Are your fans going ballistic when you boot? What OS are you booting?
Just to be 100% clear, you are saying that when your framework is off (powered completely down and shut off), and you turn it on that the fans go ballistic and you are reaching temps of 100C?
If this is the case I would recommend you contact Framework support because this is not normal behavior.
I was batch 5. Framework is sending me a new heat sink assembly, although I’m a little doubtful that will solve the problem. After more observation and testing, it seems to be intermittent spikes rather than general overheating. Whoever I was corresponding with says that this model CPU is built to occasionally push up to 100C, but it seems overeager if it’s doing it with minimal CPU usage. There were no spikes when running cinebench.