Jump to content

Johnny159

Members
  • Posts

    6
  • Joined

  • Last visited

Johnny159's Achievements

Noob

Noob (1/14)

2

Reputation

  1. @chchiyan Seems like you have (like me) no device connected to this pci bridge and therefore no endpoint. You should access register 0x50 of this brigde and set it to L1 (0x02) instead of L0sL1 (0x03). Following command should work for you: sudo setpci -s 00:1c.0 0x50.B=0x02 Based on my setup, this is not improving the consumption. I have now stripped down my system to the bare minimum: Motherboard CPU + Fan Power Supply USB live OS (Unraid and Debian 12 standard) on USB2.0 port All unused peripherals deactivated in BIOS ASPM and other measures enabled Played around with cpu governor and other settings I don't see any impact from the settings and am stuck at 12-13 W idle consumption with 80-90% in C10 state. Does anyone has further proposals?
  2. I have not yet created a script for the workaround, but just did a line by line execution in the console. https://wireless.wiki.kernel.org/en/users/Documentation/ASPM You can follow the explanation in above link to adjust your ASPM settings for any device with just its root complex and no endpoint. Hope this solves your problem.
  3. @bagican Thanks a lot for your help! It took me a bit to workaround the script since I have no end point connected to this bridge, but finally set it too L1 only. Unfortunately it has no impact on the power consumption. I have not found any explanation, why spinning up the HDDs blocks the package to reach C10, but only C8. Does anyone has a glue why this is happening? Any other idea what to change for further idle reduction?
  4. @bagican: Do you have any hint how to do this? My bios settings are all set to ASPM L1 only and ASPM power policy is either powersave or powersupersave, but I don't find any option to avoid the L0sL1 mode for the second PCI bridge.
  5. Thanks for the quick reply. ASPM is enabled on all devices as far as I can see: ~# lspci -vvPPDq | awk '/ASPM/{print $0}' RS= | grep --color -P '(^[a-z0-9:./]+|:\sASPM (\w+)? ?((En|Dis)abled)?)'; 0000:00:1a.0 PCI bridge: Intel Corporation Device 7a48 (rev 11) (prog-if 00 [Normal decode]) LnkCtl: ASPM L1 Enabled; RCB 64 bytes, Disabled- CommClk+ 0000:00:1c.0 PCI bridge: Intel Corporation Device 7a38 (rev 11) (prog-if 00 [Normal decode]) LnkCtl: ASPM L0s L1 Enabled; RCB 64 bytes, Disabled- CommClk- 0000:00:1a.0/01:00.0 Non-Volatile memory controller: Samsung Electronics Co Ltd NVMe SSD Controller SM981/PM981/PM983 (prog-if 02 [NVM Express]) LnkCtl: ASPM L1 Enabled; RCB 64 bytes, Disabled- CommClk+ The power supply is the only one I have available at the moment. Since it should have a quite high efficiency even in low load condition, I hope it to be a minor optimization factor. I have equipped a single RAM module and have now performed an additional test with Unraid in safe mode without any plugin, docker etc. active and disconnected HDDs and fans (beside CPU fan). Display and USB devices are plugged off, only Unraid USB stick and Ethernet are connected. Due to disconnected HDDs, the array can of course not be started. But since powertop shows C10 state (checked while not in Unraid safe mode) I assume, that system is still controlled and power safe modes are active. Power consumption shows: 12-13 Watt Both measurement devices are cross-validated with several classical bulbs and LED lamps for which they show accurate results. Any proposal for further reduction measures?
  6. Hi guys, hope someone might have an answer for my system behavior. My setup: - Intel i3-13100 - ASUSTeK COMPUTER INC. Pro B760M-C - Samsung 970 Evo Plus - 2x WD HDD - Corsair RM550x - Only onboard controllers for SATA, Ethernet, ... - Unraid 6.12.6 - Powertop 2.15 I am able to achieve approx. 15-17 W in Idle with HDD spinned down. With spinned up HDD consumption is around 20-22 W. (No dockers or VMs active) Do you feel there is space for further improvements? What is strange to me in powertop: With spinned up HDDs C-states are reaching up to C10 With spinned down HDD, only C8 is possible Do you have any idea what could be the reason for this? Any further info needed to identify the root cause? Looking forward to your advices and ideas :-)
×
×
  • Create New...