Add Support for GPU Clock Frequencies (sclk and mclk) in hwmon Collector

[zxhdaze]

Hi team,

I have added a new feature to the hwmon collector that enables the monitoring of GPU clock frequencies, specifically sclk (Shader Clock) and mclk (Memory Clock).

The hwmon collector now reads the following data from the /sys/class/hwmon directory:

/sys/class/hwmon/hwmon0/freq1_input:214000000
/sys/class/hwmon/hwmon0/freq1_label:sclk
/sys/class/hwmon/hwmon0/freq2_input:300000000
/sys/class/hwmon/hwmon0/freq2_label:mclk

and generate following metrics:

# HELP node_hwmon_freq_freq_mhz Hardware monitor for GPU frequency in MHz
# TYPE node_hwmon_freq_freq_mhz gauge
node_hwmon_freq_freq_mhz{chip="0000:9d:00_0_0000:9e:00_0",sensor="mclk"} 300
node_hwmon_freq_freq_mhz{chip="0000:9d:00_0_0000:9e:00_0",sensor="sclk"} 214

Please let me know if there are any questions or if further adjustments are needed.

Thank you!

closes:#3092

[aieri]

here's the relevant kernel documentation reference: https://docs.kernel.org/gpu/amdgpu/thermal.html

[jneo8]

Hi, I know there is no unit test for hwmon_linux now. Is that possible we provide unit test for it?

[SuperQ]

You can add a test file to the fixtures. This way it will be tested by the end-to-end-test.sh script.

[zxhdaze]

We're happy to provide tests for this if you'd like, but we'll leave it up to you (upstream) to decide whether you want to include them or keep it as is.

[discordianfish]

LGTM, thanks!

[aieri]

Thanks for reviewing! Could someone with write access run the workflows, please?

[Deezzir]

Thanks for reviewing! Could someone with write access run the workflows, please?

+1