Yeah, I haven’t heard back in a while either. Unfortunately I’m out of both the return and warranty periods.
Is this “BD PROCHOT” ?
There are two different PROCHOT.
- PROCHOT - Core CPU hitting limit
- BD PROCHOT - A temp sensor elsewhere on the motherboard triggering the PROCHOT state.
We would need FW Engineering to tell us where the BD PROCHOT sensor actually is?
Its times like this where users having the schematics would help a lot. 
Summary, this is not a CPU getting too hot issue, its something else thinking its getting too hot, but we don’t yet know what that something else is.
What are the fan speeds when this problem is happening?
Maybe one of the fans is not working, or a VRM is getting to hot.
Is this “BD PROCHOT” ?
There are two different PROCHOT.
No idea, is there a way for me to tell based on the EC log output?
this is not a CPU getting too hot issue, its something else thinking its getting too hot, but we don’t yet know what that something else is.
This is a good theory and seems likely, since, as I’ve noted, my CPU core temperatures are normal for some time before the throttling stops.
Would all these other sensors be reported to userspace? If so, then I’m not sure if this is it, as all temperatures are shown to be low enough. (Non “Core#” sensors I can see are “PackageId0,0”, “Iwlwifi_1,0”, “Composite,0”, “Sensor1,0”, and “Sensor2,0”.
AFAICT, “PackageId0,0” is a part of the CPU, “Iwlwifi_1,0” is (obviously) the wifi card, and the final three are related to the NVMe card.
framework_tool --thermal also gives me a look at:
F75303_Local: 46 C
F75303_CPU: 50 C
F75303_DDR: 48 C
Battery: 36 C
PECI: 49 C
F57397_VCCGT: 45 C
Fan Speed: 3079 RPM
What are the fan speeds when this problem is happening?
Not sure, but I’ll try to remember to check the next time this happens. They’re always at their loudest, though, so I expect they’re running at full tilt.
Maybe one of the fans is not working
Isn’t there just one fan? So if I hear a fan, that should be it.
FW13 - one fan
FW16 - two fans.
Ah, FW13 here, so just a single fan.
This describes the design of the Intel CPUs. I have not seen the problem people are describing. This message is just be describing the design of the Intel CPU.
There is a bit of a complex pattern of events before the CPU will throttle like this.
Something too hot == H
OK temp again == O
CPU throttling == T
HTOHTOHTOHTOTOTOTOTOTOTOTOT
I.e. the CPU will behave as one would expect, too hot, throttle, OK temp again.
But, if this cycle repeats too often and close together, the CPU will just Throttle for a long period.
The challenge of course is finding out what component is getting too hot.
So, if one records temperatures over a long period, one should see the cycles of HTO until eventually the CPU stays (as per its design) in T for longer.
This is probably why FW support were asking for temperature data to be collected over time before it happens, as opposed to only collecting data after it happens.
2 Likes