Official GIGABYTE Forum

NVMe SSD power management (APST) and thermal control

dadinn

  • 1
  • 0
NVMe SSD power management (APST) and thermal control
« on: August 22, 2023, 06:04:46 pm »
I have a Gigabyte R272-Z34 server with an AMD EPYC 75F3 CPU, and 8x 2TB SAMSUNG MZWLJ1T9HBJR-00007

The storage is mostly idle at the moment, yet the SSD temperatures reported by smartctl say its around 51 celsius.
I would like to put SSDs into some lower power states most of the time, and only power them up to max, when there is a heavy workload coming.

I've looked into this an ArchWiki says I should be using nvme get-feature /dev/nvme0 -f 0x0c -H to check whether APST is supported by the SSD:
https://wiki.archlinux.org/title/Solid_state_drive/NVMe#Power_Saving_(APST)

With this, I get an Invalid Field in Commad error.

Instead if I run nvme get-feature /dev/nvme0 -f 0x3 -H, then I get the following output:

get-feature:0x02 (Power Management), Current value:00000000
        Workload Hint (WH): 0 - No Workload
        Power State   (PS): 0

Not sure what I can deduce from this output, or in case this means that APST feature is supported, then how I could adjust it so that the drives go idle when there is no serious workload.

Does this require some bios adjustment in the casis, or am I using the wrong SSDs?