Rainmeter Forums

Hello,

In recent weeks I've had two users of my skin suite report that my CPU Meter Gadget is reporting a CPU clock speed of 0 MHz. After drilling down into the problem, it appears the problem is here, isolated to a single measure:
This measure is returning zero. Both users have one thing in common: they are using AMD Ryzen CPUs, one with a Ryzen 7 3700X and another with a Ryzen 9 3950X. I have no way to test this myself, as all my PCs are Intel machines.

Is anyone else seeing this? I have no other information to pass on (like Windows version, etc.). Using CoreTemp or HWiNFO works as an alternative to display the clock speed, but PowerPlugin appears to be acting up.

Edit: This has been confirmed as a bug in Windows. This is not a Rainmeter issue.

Info from one of the users having this issue:

Rainmeter 4.3.1.3321 (64-bit)
Language: English (1033)
Build time: 2019-09-22 7:26:11
Commit Hash: 0ba4dfc4
Windows 10 Pro 1909 64-bit (build 18363) - English (1033)
Path: C:\Program Files\Rainmeter
SkinPath: C:\Users\...\Documents\Rainmeter\Skins
SettingsPath: C:\Users\...\AppData\Roaming\Rainmeter
IniFile: C:\Users\...\AppData\Roaming\Rainmeter\Rainmeter.ini

It works fine for me on an AMD Ryzen 7 2700X.

jsmorley wrote: ↑June 25th, 2020, 6:37 pm It works fine for me on an AMD Ryzen 7 2700X.

Sigh... I was afraid you'd say that. Unfortunately, for a few folks it is definitely broken.



I'm not sure what other info I can provide to debug this, so I'm kind of stuck. I'd hate to have to create some weird "fallback" measure for the few folks that are getting this.

I don't know offhand what the Power plugin is using to get this information, but if it is using PerfMon to do so, you might have them try this:

https://forum.rainmeter.net/viewtopic.php?f=118&t=23156

jsmorley wrote: ↑June 25th, 2020, 6:59 pm

Code: Select all

[MeasureCPUSpeed]
Measure=Registry
RegHKey=HKEY_LOCAL_MACHINE
RegKey=HARDWARE\DESCRIPTION\System\CentralProcessor\0
RegValue=~MHz
UpdateDivider=-1

Yes, that's my fallback exactly. I was gonna add an IfCondition to the first measure to fall back and use your measure if the PowerPlugin is returning zero. Kind of weird to have to do this, tho.

jsmorley wrote: ↑June 25th, 2020, 7:03 pm I don't know offhand what the Power plugin is using to get this information, but if it is using PerfMon to do so, you might have them try this:

https://forum.rainmeter.net/viewtopic.php?f=118&t=23156

I'll ask them to try this...

SilverAzide wrote: ↑June 25th, 2020, 7:05 pm Yes, that's my fallback exactly. I was gonna add an IfCondition to the first measure to fall back and use your measure if the PowerPlugin is returning zero. Kind of weird to have to do this, tho.


I'll ask them to try this...

For this bit of information, I don't see any advantage with the Power plugin over the built-in Registry measure.

jsmorley wrote: ↑June 25th, 2020, 7:07 pm For this bit of information, I don't see any advantage with the Power plugin over the built-in Registry measure.

Well, it IS somewhat marginal at best, it is true. The one thing I've noticed is that the Power plugin will report the clock when my laptop's CPU is being "managed" by the Power settings (i.e., the cores are being throttled to save battery). It won't report boost/turbo speeds, though, as we all know. So, yes, it is only marginally better than the static value in the registry.

jsmorley wrote: ↑June 25th, 2020, 7:03 pm I don't know offhand what the Power plugin is using to get this information, but if it is using PerfMon to do so, you might have them try this:

I tried a UsageMonitor measure to report the total processor frequency, and it seems to report the exact same thing at the exact same second as the Power plugin, so it could be that the plugins are getting the data from the same place.