4 hours ago, jit-010101 said:

have you looked up standby power consumption for those?

It's 5x spinning rust + 7x ssd so 12 physical devices.

There is a limit to what kind of power consumption you can save from standby and spindown.

however what you are describing sounds like something is constantly writing to your array.

do you have apps on a separate nvme/ssd only pool? If not there's your answer as to why the hdds constantly spin up.

i have all my config etc on the nvme for plex and the arrs..

i just disabled my bazarr docker.. trying to see if that changes anything..
edit: disk 1 just started: Dec 6 13:20:19 emhttpd: read SMART /dev/sdk
disk activity under tools shows nothing..

Edited December 6, 2025Dec 6 by furian

Hi, I'm building a new system with unRAID and I'm having a small problem with the power saving states in package C. Has anyone here with an MSI Tomahawk Z790 WiFi motherboard been able to get past the C3 package? I have one installed with a 14500t processor, but it's stuck on the C3 package.

The BIOS configuration looks like this:

Powertop --auto-tune running

I've also tried Ubuntu Live USB with the same result.

Nothing else is connected besides RAM, CPU, and mouse/keyboard (apart from the USB drive with Ubuntu Live USB or unRAID).

Edited January 19Jan 19 by Nozle

@mgutt , i was wondering if there is any chance to compile an new version of powertop and publish it for Unraid?

I saw there are some updates to powertop in the commit history, for example, support for Intel Arrow Lake, which helps me to see the C states

On the chance this helps anyone else: I've had to completely roll back all my ASPM settings and uninstall powertop; at some point, I will see what I can safely enable again, as I've lost ~10+W of efficiency. Why remove? Hard system locks and USB corruption; it's a USB 2.0 port, USB 2.0 flash drive, which has previously been recommended as a way to avoid this. Why did it start now? No idea! In fact, I got all new hardware (except drives) after I started having this issue late last year and it actually killed a Samsung BAR Plus (which was in a 2.0 port). It was an AMD system, and I shifted to Intel. The new hardware went 3 months stable, then 38 days, then 2 days, 8hr, 2hr... and from there it was mere minutes. Of course, at that point I was actively troubleshooting. Went days on memtest, all clean; It was locking up even with Windows (Hirens USB), but disabling just Native ASPM seemed to resolve that. It may be that's the only Bios setting I need, + some powertop configuration changes (it was scripted in go file).

ASUS w680-ace ipmi, Intel i5-14500 for reference.

Any tips on what settings I might need to avoid would also be appreciated - I need to go through the thread again too looking for tips.

1 hour ago, _cjd_ said:

Any tips on what settings I might need to avoid would also be appreciated

On the first page you find the custom commands, which would be executed by powertop. You could now remove the commands, which is related to USB devices.

I had to do the same as my Zigbee USB stick became unresponsive:

```
# Autosuspend for USB device
#echo auto | tee /sys/bus/usb/devices/*/power/control
#echo on > /sys/bus/usb/devices/1-4/power/control

```

Regarding Zigbee, there may be a better option (or an alternative) to the USB stick. You can use an Ethernet (PoE) version of a Zigbee coordinator instead. One such device, which has good reviews, is the SMLIGHT SLZB-06 (though it belongs to another thread).

1 minute ago, bagican said:

You can use an Ethernet (PoE) version

Yeah, I even bought it, but finally I got in love with Shelly ^^ So finally I will completely abolish Zigbee.

I've made some weird observations and need to know if more people have the same issue.

I have 2 NVME ports on my ITX board with 11th gen cpu.

Before I upgraded from 10th to 11th gen I could only have the back m.2 port running thru my PCH chip leaving me with only C3-state.

Now after I've added a second m.2 to the front port, I went from ~17w to 25-29w stuck in C2.

Some people mentioned that by having the m.2 directly sitting in the 20 lanes to the CPU, keeps the SSD active for no reason. But I've found out that someone else with the same board, has almost the same setup and no trouble reaching C10 with Windows or Debian.

After some investigation on my system, I found some interesting stuff.

My ASPM and BIOS settings are fine.

[    3.207150] nvme 0000:02:00.0: platform quirk: setting simple suspend <-- That could mean it could bypass deeper sleep
[    3.207243] nvme nvme1: pci function 0000:02:00.0
[    3.207255] nvme nvme0: pci function 0000:08:00.0
[    3.216209] nvme nvme0: 16/0/0 default/read/poll queues
[    3.218728]  nvme0n1: p1
[    3.221735] nvme nvme1: D3 entry latency set to 8 seconds  <-- It should enter D3cold or hot after 8 seconds, but it never does.
[    3.231602] nvme nvme1: allocated 64 MiB host memory buffer.
[    3.279170] nvme nvme1: 12/0/0 default/read/poll queues
[    3.286783]  nvme1n1: p1
[   79.541831] nvme nvme0: using unchecked data buffer

This tells me that the NVME is consuming more thant he virtual power state reflects. If it would enter d3 after 8 seconds, it should be here, right?

root@NASty:~# cat /sys/class/nvme/nvme1/device/power_state
D0                                                                                                                                  

root@NASty:~# nvme get-feature /dev/nvme0 -f 0xc -H
get-feature:0x0c (Autonomous Power State Transition), Current value:0x00000001
    Autonomous Power State Transition Enable (APSTE): Enabled

Does the power state below reflect that the consumption comes on top of the already drained energy, or print the real time use in the WebUI? Because the PS4 of 0,005w seems to be wrong.

root@NASty:~# nvme get-feature -f 2 -H /dev/nvme1
get-feature:0x02 (Power Management), Current value:0x00000004
    Workload Hint (WH): 0 - No Workload                                                              <--- no workloads, tells me that it idles (with containers active/inactive)
    Power State   (PS): 4                                                                                        

root@NASty:~# smartctl -c /dev/nvme1n1
smartctl 7.5 2025-04-30 r5714 [x86_64-linux-6.12.54-Unraid] (local build)
=== START OF INFORMATION SECTION ===
Firmware Updates (0x16):            3 Slots, no Reset required
Optional Admin Commands (0x0017):   Security Format Frmw_DL Self_Test
Optional NVM Commands (0x0055):     Comp DS_Mngmt Sav/Sel_Feat Timestmp
Log Page Attributes (0x0f):         S/H_per_NS Cmd_Eff_Lg Ext_Get_Lg Telmtry_Lg
Maximum Data Transfer Size:         512 Pages
Warning  Comp. Temp. Threshold:     82 Celsius
Critical Comp. Temp. Threshold:     85 Celsius
Namespace 1 Features (0x10):        NP_Fields
Supported Power States
St Op     Max   Active     Idle   RL RT WL WT  Ent_Lat  Ex_Lat
 0 +     5.24W       -        -    0  0  0  0        0       0
 1 +     4.49W       -        -    1  1  1  1        0       0
 2 +     2.19W       -        -    2  2  2  2        0    6000
 3 -   0.0500W       -        -    3  3  3  3     6000    1500
 4 -   0.0050W       -        -    4  4  4  4     4000    9000

ASPM is enabled.

08:00.0 Non-Volatile memory controller: Sandisk Corp SanDisk Extreme Pro / WD Black 2018/SN750/PC SN720 NVMe SSD (prog-if 02 [NVM Express])
        LnkCap:    Port #0, Speed 8GT/s, Width x4, ASPM L1, Exit Latency L1 <8us
        LnkCtl:    ASPM L1 Enabled; RCB 64 bytes, LnkDisable- CommClk+
02:00.0 Non-Volatile memory controller: Samsung Electronics Co Ltd NVMe SSD Controller 980 (DRAM-less) (prog-if 02 [NVM Express])
        LnkCap:    Port #0, Speed 8GT/s, Width x4, ASPM L1, Exit Latency L1 <64us
        LnkCtl:    ASPM L1 Enabled; RCB 64 bytes, LnkDisable- CommClk+

I thought that it might be a SSD firmware issue, but I have tested another 1TB drive in place of the Samsung and it behaves the same.

Shouldn't I be able to "spin down" the cache too if it aint in use? That would have been nice to find out if the system might be able to get past c2 after I turn em off via the WebUI and check values.

Edited January 19Jan 19 by W0nderW0lf

14 hours ago, mgutt said:

Yeah, I even bought it, but finally I got in love with Shelly ^^ So finally I will completely abolish Zigbee.

hehe, me too! I bought it more than a year ago, but until yesterday I hadn't used it because I already have most things covered with Shelly. But I started adding zigbee devices I already have too ...

Edited January 20Jan 20 by bagican

On 1/19/2026 at 12:52 PM, mgutt said:

On the first page you find the custom commands, which would be executed by powertop. You could now remove the commands, which is related to USB devices.

I had to do the same as my Zigbee USB stick became unresponsive:

```
# Autosuspend for USB device
#echo auto | tee /sys/bus/usb/devices/*/power/control
#echo on > /sys/bus/usb/devices/1-4/power/control

```

Already in my sights. Thanks.

I have an orangepi pi zero running pihole and my zwave dongle tucked away central in the house, zwave-js and home assistant still on unraid/docker. Not sure if zigbee has a similar remote approach. 1.5w poe powered.

Edited January 23Jan 23 by _cjd_

On 12/19/2024 at 12:48 PM, lordprotector said:

After almost three weeks of researching and playing, I was able to achieve pkg C10 with W680D4ID-2T/G5/X550 board (full story there: https://forum.level1techs.com/t/asrock-rack-w680d4id-2t-g5-x550-c-states-aspm-problem/221103), unfortunately only without one NVMe and dGPU installed.

 

As far as I understand C8 and C10 requires ASPM L1 substates (more particularity L1.1 or L1.2), but for some unknown reason CPU PCIe lines don't support it (at least Linux's lspci doesn't show those capabilities), but BIOS has L1 substate option enabled (though I cannot change it).

 

Is it a BIOS bug or modern Intel CPU doesn't support L1 substates on those lines?

 

UPD. I was able to enable L1 substates for CPU PCIe lines, but unfortunately Linux for some reason doesn't use them, moreover it limits the system by pkg C3 instead of C6.

Old post but can also confirm after turning off L1 substates I got pkg C6 instead of C3 when it was turned on.

Community ... help needed with ARC A380 GPU, which is preventing C10

System:

Gigabyte B760M DS3H DDR4

Intel i3-12100

32GB Ram

4 x Toshiba N300 8TB (onboard sata ports)

1 x ASM1166 with flashed Bios

1 x Nvme (in chipset slot)

3 x Sata SSD

2 x BD Disc drives

3 x Fans controlled by Bios (can´t get them controlled reliably in unraid)

PSU: Coolermaster NWE 400

Idle Wattage measured with a Tapo P110 plug: 27 Watt / C10 / spindown / Docker running

I´d say not too shabby. A Watt or two remain to be shaved off by disableling the BD drives when not needed with a switch. And when I can I´ll swap the drives for bigger ones and reduce the amount of drives needed.

It was a fight going for a week or two in between work and sleep. I´ve been trying to get my system past C3. Tried numerous clean unraid sticks and ubuntu and even went into SCEWIN, searched through all settings related to ASPM, PEG, DMI and whatnot. Willing to destroy my Bios and board flashed the editied bios file and was about to give up. Then took my USB unraid stick, removed all that was in go file and syslinux (only went with the usual "force" and suddenly my system went into C10. For no apparent reason, or at least for no reason I could point to.

Lets say I´m quite happy with the outcome of 27 Watts.

For the final battle I´ll need your help. I would like to install a Sparkle A380, but it won´t let the HW Package go past C3 again. It adds 15 Watts of idle.

From what I´ve read it might be that 15 Watts is the lowest it will go anyway (no idea how Wolfgang did it). But if it prevents C10 I´ll be ditching the card and move it into my main computer for encoding tasks.

Up until now I´ve tried:

# ASPM Global Policy setzen

echo powersave > /sys/module/pcie_aspm/parameters/policy

echo performance > /sys/module/pcie_aspm/parameters/policy

echo default > /sys/module/pcie_aspm/parameters/policy

# Bridge 01:00.0: L1.2 Substates erzwingen (Register 1e4)

setpci -s 01:00.0 1e4.l=425a000f

# ARC 03:00.0: L1 Substates manuell aktivieren (Register 158 / 1a8)

setpci -s 03:00.0 158.l=f

setpci -s 03:00.0 1a8.l=f

setpci -s 03:00.0 158.l

# Common Clock (CommClk+) für Bridge und ARC erzwingen

setpci -s 01:00.0 CAP_EXP+0x10.w=0040:0040

setpci -s 03:00.0 CAP_EXP+0x10.w=0040:0040

# PCIe Link Retrain

setpci -s 01:00.0 CAP_EXP+0x10.w=0020:0020

# Link Speed auf Gen 3 (8GT/s) forcieren

setpci -s 01:00.0 CAP_EXP+0x30.w=0003:000f

# LTR (Latency Tolerance Reporting) deaktivieren

setpci -s 01:00.0 0x400.l=0 setpci -s 03:00.0 0x400.l=0

# Root Port 00:01.0 ASPM L1 aktivieren

setpci -s 00:01.0 CAP_EXP+0x10.w=0002:0003

setpci -s 00:01.0 1e4.l=425a000f

Our AI friends tell me because the card is falsely reporting 1x speeds it can´t be done.

What more information you need?

taruga-diagnostics-20260126-0846.zip

Ok I think I am actually cursed. 4th server and still havent seen anything above C3 (that is usable).

This latest one has:

Asus Pro H610T D4-CSM motherboard

Intel i3 12100 cpu

Kingston KC3000 1TB nvme

SK Hynix 2x16GB DDR4 SO-DIMM ram

19v 6.32a (120W) power brick

2x Noctua 40mm fans

Thermalright axp90-x47 full copper cpu fan

With just proxmox running, idles at 13W measured from my Tuya Zigbee smart plug.

Tunables are all good and WakeUp are all disabled.

ASPM are all enabled.

CPU governor set to powersave

All bios settings I adjusted:

Advanced:

Platform Misc Configuration:

Native ASPM Disabled to Auto

DMI Link ASPM Control Disabled to L1

ASPM auto to L1

L1 Substates Disabled (was disabled from the start)

DMI ASPM Disabled to “ASPM L1”

DMI Gen3 ASPM Disabled to “ASPM L1”

CPU Configuration:

CPU - Power Management Control:

CPU C-states Auto to Enabled

Package C State Limit Auto to C10

APM Configuration:

Max Power Saving Disabled to Enabled

ErP Ready Disabled to Enabled(S4+S5)

Onboard Devices Configuration:

HD Audio Enabled to Disabled

USB power delivery in soft off state (S5) Enabled to Disabled

Connectivity mode (Wi-Fi & Bluetooth) Enabled to Disabled

Serial Port Configuration:

Serial Port Enabled to Disabled

Monitor:

Q-Fan Configuration:

CPU Fan Profile Standard to Silent

Chassis Fan(s) Configuration:

Chassis Fan Profile Standard to Silent

On 2/3/2026 at 12:31 AM, silicon4079 said:

Ok I think I am actually cursed. 4th server and still havent seen anything above C3 (that is usable).

This latest one has:

Asus Pro H610T D4-CSM motherboard

Intel i3 12100 cpu

Kingston KC3000 1TB nvme

SK Hynix 2x16GB DDR4 SO-DIMM ram

19v 6.32a (120W) power brick

2x Noctua 40mm fans

Thermalright axp90-x47 full copper cpu fan

With just proxmox running, idles at 13W measured from my Tuya Zigbee smart plug.

Tunables are all good and WakeUp are all disabled.

ASPM are all enabled.

CPU governor set to powersave

All bios settings I adjusted:

Advanced:

Platform Misc Configuration:

Native ASPM Disabled to Auto

DMI Link ASPM Control Disabled to L1

ASPM auto to L1

L1 Substates Disabled (was disabled from the start)

DMI ASPM Disabled to “ASPM L1”

DMI Gen3 ASPM Disabled to “ASPM L1”

CPU Configuration:

CPU - Power Management Control:

CPU C-states Auto to Enabled

Package C State Limit Auto to C10

APM Configuration:

Max Power Saving Disabled to Enabled

ErP Ready Disabled to Enabled(S4+S5)

Onboard Devices Configuration:

HD Audio Enabled to Disabled

USB power delivery in soft off state (S5) Enabled to Disabled

Connectivity mode (Wi-Fi & Bluetooth) Enabled to Disabled

Serial Port Configuration:

Serial Port Enabled to Disabled

Monitor:

Q-Fan Configuration:

CPU Fan Profile Standard to Silent

Chassis Fan(s) Configuration:

Chassis Fan Profile Standard to Silent

I know the feeling. I have been doing some more testing and I´ve abandoned using the ARC for the moment. My C10 comes at some more cost though: the 2.5gbe speed I get from my Realtek onboard NIC is abysmal. Because I run on the stock kernel drivers and not the recommended ich777 drivers. When I install the RTL drivers I get sweet 2.5gbe speeds, but the system won´t go below C3, no matter what I do with any of the usual power saving commands. It shows as ASPM Enabled, but just won´t enter below C3. So getting into C10, now with two nvme, one even on the CPU lanes, comes with 20-20mb/s on the network. Not ideal I´d say.

Maybe you should try the kernel network driver and try enabling ASPM with ASPM Helper plugin. At least you can rule out another culpirt, if it does not work.

1 hour ago, nassauer said:

I know the feeling. I have been doing some more testing and I´ve abandoned using the ARC for the moment. My C10 comes at some more cost though: the 2.5gbe speed I get from my Realtek onboard NIC is abysmal. Because I run on the stock kernel drivers and not the recommended ich777 drivers. When I install the RTL drivers I get sweet 2.5gbe speeds, but the system won´t go below C3, no matter what I do with any of the usual power saving commands. It shows as ASPM Enabled, but just won´t enter below C3. So getting into C10, now with two nvme, one even on the CPU lanes, comes with 20-20mb/s on the network. Not ideal I´d say.

Maybe you should try the kernel network driver and try enabling ASPM with ASPM Helper plugin. At least you can rule out another culpirt, if it does not work.

Note that this setup above is with proxmox and the motherboard only has a single 1gbe ethernet port and not 2.5gbe.

Others with this identical setup with just a different nvme has gotten it down to 3W in idle and is in C10. I am 100% sure it is due to the Kingston nvme (KC3000) that prevents anything more than C3. As the KC3000 also did this in my other unraid server and when I swapped it out to a samsung 990 pro I instantly got 7w less drain in idle (I had two kc3000 and swapped with two 990 pro). Just extremely frustrating how much work is needed to get low power consumption on x86 based hardware....my orange pi 5 plus (ARM) idles at 5.4W and this is including an UPS! (This ups: https://www.waveshare.com/ups-module-3s.htm ) and I havent touched any other settings in it.....

46 minutes ago, silicon4079 said:

Note that this setup above is with proxmox and the motherboard only has a single 1gbe ethernet port and not 2.5gbe.

Ah sorry. Somehow I did not expect Proxmox in here ;) I can relate to the Kingston problem. Found out the one I had does not do any ASPM. Now running two Samsung 980pro.

59 minutes ago, nassauer said:

Ah sorry. Somehow I did not expect Proxmox in here ;) I can relate to the Kingston problem. Found out the one I had does not do any ASPM. Now running two Samsung 980pro.

This thread probably the most popular place for any c-state/powertop/low-idle discussion so...

Which Kingston do you have? I know another guy with an NV1 that also prevents deeper than C3, I dunno if it supports ASPM. The KC3000 actually DOES support ASPM though.

However I have tried the KC3000 in another proxmox server and another nvme (JUHOR J3500) which does NOT have ASPM but they both idled at around the same anyway....

Although a quick search ( https://forums.unraid.net/topic/98070-reduce-power-consumption-with-powertop/page/35/#findComment-1424783 ) seems like the NV2 does not have ASPM so I suppose the same with the NV1.

https://forums.unraid.net/topic/98070-reduce-power-consumption-with-powertop/page/13/#comment-1255922

Same with the A2000.

Edited February 5Feb 5 by silicon4079

6 hours ago, silicon4079 said:

Which Kingston do you have?

Kingston SM2267XT. An older drive I guess. Can´t remember in which of my old computers I found it.

The first post instructs on how to install powertop but I noticed it references Unraid version 6.11.0, can I still install powertop on v7.2.3 the same way?

Hi everyone,

I updated my server and paid attention to energy-saving aspects when selecting the hardware.

I got myself an HP Elitedesk 800 G6.

I replaced the CPU with an i9-12900 and added 16GB of RAM.

Initial tests with an SSD showed very good values of approx. 5-6 watts.

Of course, an SSD doesn't do much on its own. So I expanded my setup to:

- 4x Sata HDD 2x 18TB1, 1x20TB Toshiba and 22TB Seagate Parity

- Plex 8747 4-slot SSD card (equipped with 4 SSDs)

- 10GB ASUS XG-C100F network card

- 3 fans for CPU, HDD, and SSD

Unfortunately, I can't achieve any C-states at all now. Consumption is now 55W.

I find that to be a lot when all HDDs are idle.

Through the thread, I found out that the chipset of the 10GB NT card outputs ASPM Disabled.

I then removed the card, but that didn't make much difference.

Still C-State 0.

I read here that Intel cards enable better C-states, but that wouldn't help because I can't achieve better states even without a card.

Can someone please help me? I'm desperate.

On 3/2/2026 at 9:53 PM, Benni_ said:

Hi everyone,

I updated my server and paid attention to energy-saving aspects when selecting the hardware.

I got myself an HP Elitedesk 800 G6.

I replaced the CPU with an i9-12900 and added 16GB of RAM.

Initial tests with an SSD showed very good values of approx. 5-6 watts.

Of course, an SSD doesn't do much on its own. So I expanded my setup to:

- 4x Sata HDD 2x 18TB1, 1x20TB Toshiba and 22TB Seagate Parity

- Plex 8747 4-slot SSD card (equipped with 4 SSDs)

- 10GB ASUS XG-C100F network card

- 3 fans for CPU, HDD, and SSD

Unfortunately, I can't achieve any C-states at all now. Consumption is now 55W.

I find that to be a lot when all HDDs are idle.

Through the thread, I found out that the chipset of the 10GB NT card outputs ASPM Disabled.

I then removed the card, but that didn't make much difference.

Still C-State 0.

I read here that Intel cards enable better C-states, but that wouldn't help because I can't achieve better states even without a card.

Can someone please help me? I'm desperate.

Yeah my own unraid server with the 12600K (TDP lowered to 63W) and 64gb ram, with 2x2tb nvme (samsung 990 pro) and 3x hdds (2x 20tb and 1x18tb) with 2.5gb nic, idles around 40W.

Adding the hdds is what made it go up a ton, without the array running it idles around 32W but I am in C6 c-state. I also have no clue how to improve this further as when my array is on and my 100+ docker containers are on I am out of any c-state as well.

Have you checked your bios settings? Here is my current settings:

Advanced tab:
RC ACPI Settings:
    Native APSPM auto to enabled
    BDAT ACPI Table Support Disabled to Enabled
    D3 Settings for Storage D3Hot to D3Cold
    Low Power S0 Idle Capability Disabled to Enabled
    Sensor Standby Disabled to Enabled
    PEP Constraints Configuration:
        PEP PCIe Storage D3 to D0/F1
        PEP PCIe WLAN D0/F1 to D3
        PEP PCIe Other No Constraint to D0/F1
Connectivity Configuration:
    CNVi Mode Auto Detection to Disable Integrated
    Discrete Bluetooth Interface USB to Disabled
    
Power & Performance:
    CPU - Power Management Control:
        Boot Performance Mode Turbo Performance to Max Battery
        Turbo Mode Enabled to Disabled
        View/Configure Turbo Options:
            Power Limit 1 Ovveride Disabled to Enabled
	        Power Limit 1 0 to 63000
	        Power Limit 2 0 to 65000
	    "Platform PL1 Enable" Disabled to Enabled
	    Platform PL1 Power 0 to 63000
	    "Platform PL2 Enable" Disabled to Enabled
	    Platform PL2 Power 0 to 65000
        Package C State Limit Auto to C10
Hardware Monitor:
    Smart Fan Function:
        Temperature Limit of CPU Fan OFF 0 to 25
        Temperature Limit of CPU Fan start 20 to 26
        CPU Fan start PWM 72 to 10
        Fan Monitor Temperature Source Select "System Temperature" to "CPU Temperature"
        System Fan start PWM 15 to 25
        Temperature Limit of System fan OFF 0 to 20
        Temperature Limit of System Fan1 Start 20 to 21
        CPU Fan CurveB Slope PWM 2 to 4
        System Fan CurveB Slope PWM 2 to 5
Power Management Configuration:
    Wakeup on USB Enabled to Disabled
    Wakeup on LAN Enabled to Disabled
PCI Subsystem Settings:
    SR-IOV Support Disabled to Enabled

Chipset tab:PCH-IO Configuration:PCI Express Configuration:DMI Link ASPM Control L1 to AutoPCI Express Root Port 1,2,3,4: 
ASPM Disabled to Auto
L1 Substates Disabled
L1 Low Disabled to Enabled
PCI Express Root Port 5, 21, 25:
ASPM Disabled to Auto
L1 substates disabled
L1 low disabled

HD Audio Configuration:HD Audio Enabled to Disabled
Foxville I225 Wake on LAN Support Enabled to disabled
SATA Configuration:SATA Port (all of them) DevSlp Disabled to Enabled
Foxville I225 LAN Controller Enabled to Disabled
DeepSx Power Policies Disabled to Enabled in S4-S5
C10 Dynamic threshold adjustment Disabled to Enabled
OS IDLE Mode Disabled to Enabled
S0ix Auto Demotion Disabled to Enabled
SerialIo Configuration:I2C0,I2C1,I2C2,I2C3,I2C4,I2C5 Controller Enabled to Disabled
SPI1 Controller Enabled to Disabled

System agent (SA) Configuration:DMI/OPI Configuration:
DMI ASPM "ASPM L1" to "ASPM L0sL1"
PCI Express Configuration:
PCI Express Root Port 1,2,3:ASPM Disabled to L0sL1
L1 Substates Disabled to L1.1 & L1.2

Guys, finally unraid 7.3.0-rc1 fixed the performance issue on the i226-v.
See here for details: https://forums.unraid.net/topic/182957-is-i226-v-nic-supported-by-unraid/#findComment-1589155

So i can recommend this card now for an low power (aspm enabled) unraid build.
Without have to deal with plugins for realtek drivers 😀

Details incl. the commit see here: https://forums.unraid.net/topic/141349-plugin-realtek-r8125-r8126-r8168-and-r815267-drivers/page/18/#findComment-1620966

My power saving settings have been working fine for years but recently I see this for all scsi hosts

root@myunraid:~# echo med_power_with_dipm | tee /sys/class/scsi_host/host*/link_power_management_policy
med_power_with_dipm
tee: /sys/class/scsi_host/host0/link_power_management_policy: Operation not supported
tee: /sys/class/scsi_host/host1/link_power_management_policy: Operation not supported
tee: /sys/class/scsi_host/host10/link_power_management_policy: Operation not supported
tee: /sys/class/scsi_host/host11/link_power_management_policy: Operation not supported
tee: /sys/class/scsi_host/host12/link_power_management_policy: Operation not supported
tee: /sys/class/scsi_host/host13/link_power_management_policy: Operation not supported
tee: /sys/class/scsi_host/host14/link_power_management_policy: Operation not supported
tee: /sys/class/scsi_host/host15/link_power_management_policy: Operation not supported
tee: /sys/class/scsi_host/host16/link_power_management_policy: Operation not supported
tee: /sys/class/scsi_host/host17/link_power_management_policy: Operation not supported
tee: /sys/class/scsi_host/host18/link_power_management_policy: Operation not supported
tee: /sys/class/scsi_host/host19/link_power_management_policy: Operation not supported
tee: /sys/class/scsi_host/host2/link_power_management_policy: Operation not supported
tee: /sys/class/scsi_host/host20/link_power_management_policy: Operation not supported
tee: /sys/class/scsi_host/host21/link_power_management_policy: Operation not supported
tee: /sys/class/scsi_host/host22/link_power_management_policy: Operation not supported
tee: /sys/class/scsi_host/host23/link_power_management_policy: Operation not supported
tee: /sys/class/scsi_host/host24/link_power_management_policy: Operation not supported
tee: /sys/class/scsi_host/host25/link_power_management_policy: Operation not supported
tee: /sys/class/scsi_host/host26/link_power_management_policy: Operation not supported
tee: /sys/class/scsi_host/host27/link_power_management_policy: Operation not supported
tee: /sys/class/scsi_host/host28/link_power_management_policy: Operation not supported
tee: /sys/class/scsi_host/host29/link_power_management_policy: Operation not supported
tee: /sys/class/scsi_host/host3/link_power_management_policy: Operation not supported
tee: /sys/class/scsi_host/host30/link_power_management_policy: Operation not supported
tee: /sys/class/scsi_host/host31/link_power_management_policy: Operation not supported
tee: /sys/class/scsi_host/host32/link_power_management_policy: Operation not supported
tee: /sys/class/scsi_host/host33/link_power_management_policy: Operation not supported
tee: /sys/class/scsi_host/host34/link_power_management_policy: Operation not supported
tee: /sys/class/scsi_host/host35/link_power_management_policy: Operation not supported
tee: /sys/class/scsi_host/host36/link_power_management_policy: Operation not supported
tee: /sys/class/scsi_host/host37/link_power_management_policy: Operation not supported
tee: /sys/class/scsi_host/host38/link_power_management_policy: Operation not supported
tee: /sys/class/scsi_host/host39/link_power_management_policy: Operation not supported
tee: /sys/class/scsi_host/host4/link_power_management_policy: Operation not supported
tee: /sys/class/scsi_host/host5/link_power_management_policy: Operation not supported
tee: /sys/class/scsi_host/host6/link_power_management_policy: Operation not supported
tee: /sys/class/scsi_host/host7/link_power_management_policy: Operation not supported
tee: /sys/class/scsi_host/host8/link_power_management_policy: Operation not supported
tee: /sys/class/scsi_host/host9/link_power_management_policy: Operation not supported

Bar OS updates, currently 7.3.0, nothing has changed

#!/bin/bash
# Start the Management Utility
/usr/local/sbin/emhttp
cp /boot/config/sensor-quirk /etc/sensors.d

#Start nvidia-persistenced
nvidia-persistenced

#ASPM fix for Aquantia AQC100
sudo setpci -s 00:1c.0 0x50.B=0x43
sudo setpci -s 04:00.0 0x50.B=0x43

#ASPM fix for Realtek RTL8125
sudo setpci -s 00:1d.0 0x50.B=0x43
sudo setpci -s 05:00.0 0x80.B=0x43

#powertop auto tune
#powertop --auto-tune &>/dev/null

# -------------------------------------------------
# powertop tweaks
# -------------------------------------------------

# Enable SATA link power management
echo med_power_with_dipm | tee /sys/class/scsi_host/host*/link_power_management_policy

# Runtime PM for I2C Adapter (i915 gmbus dpb)
#echo auto | tee /sys/bus/i2c/devices/i2c-*/device/power/control

# Autosuspend for USB device
echo auto | tee /sys/bus/usb/devices/*/power/control

# Runtime PM for disk
echo auto | tee /sys/block/sd*/device/power/control

# Runtime PM for PCI devices
echo auto | tee /sys/bus/pci/devices/????:??:??.?/power/control

# Runtime PM for ATA devices
echo auto | tee /sys/bus/pci/devices/????:??:??.?/ata*/power/control

ASPM is all good for my ASM1166 sata controller

lspci -vv | awk '/ASPM/{print $0}' RS= | grep --color -P '(^[a-z0-9:.]+|ASPM )'
00:01.0 PCI bridge: Intel Corporation 12th Gen Core Processor PCI Express x16 Controller #1 (rev 05) (prog-if 00 [Normal decode])
                LnkCap: Port #2, Speed 32GT/s, Width x16, ASPM L0s L1, Exit Latency L0s <4us, L1 <16us
                LnkCtl: ASPM L1 Enabled; RCB 64 bytes, LnkDisable- CommClk+
00:06.0 PCI bridge: Intel Corporation 12th Gen Core Processor PCI Express x4 Controller #0 (rev 05) (prog-if 00 [Normal decode])
                LnkCap: Port #5, Speed 16GT/s, Width x4, ASPM L0s L1, Exit Latency L0s <4us, L1 <16us
                LnkCtl: ASPM L0s L1 Enabled; RCB 64 bytes, LnkDisable- CommClk+
00:1a.0 PCI bridge: Intel Corporation Raptor Lake PCI Express Root Port #25 (rev 11) (prog-if 00 [Normal decode])
                LnkCap: Port #25, Speed 16GT/s, Width x4, ASPM L1, Exit Latency L1 <64us
                LnkCtl: ASPM L1 Enabled; RCB 64 bytes, LnkDisable- CommClk+
00:1c.0 PCI bridge: Intel Corporation Raptor Lake PCI Express Root Port #1 (rev 11) (prog-if 00 [Normal decode])
                LnkCap: Port #1, Speed 8GT/s, Width x4, ASPM L1, Exit Latency L1 <64us
                LnkCtl: ASPM L0s L1 Enabled; RCB 64 bytes, LnkDisable- CommClk+
00:1d.0 PCI bridge: Intel Corporation Raptor Lake PCI Express Root Port #15 (rev 11) (prog-if 00 [Normal decode])
                LnkCap: Port #15, Speed 16GT/s, Width x1, ASPM L1, Exit Latency L1 <64us
                LnkCtl: ASPM L0s L1 Enabled; RCB 64 bytes, LnkDisable- CommClk+
01:00.0 VGA compatible controller: NVIDIA Corporation AD107 [GeForce RTX 4060] (rev a1) (prog-if 00 [VGA controller])
                LnkCap: Port #0, Speed 16GT/s, Width x8, ASPM L1, Exit Latency L1 <4us
                LnkCtl: ASPM L1 Enabled; RCB 64 bytes, LnkDisable- CommClk+
01:00.1 Audio device: NVIDIA Corporation AD107 High Definition Audio Controller (rev a1) (prog-if 00 [HDA compatible])
                LnkCap: Port #0, Speed 16GT/s, Width x8, ASPM L1, Exit Latency L1 <4us
                LnkCtl: ASPM L1 Enabled; RCB 64 bytes, LnkDisable- CommClk+
02:00.0 SATA controller: ASMedia Technology Inc. ASM1166 Serial ATA Controller (rev 02) (prog-if 01 [AHCI 1.0])
                LnkCap: Port #0, Speed 8GT/s, Width x2, ASPM L0s L1, Exit Latency L0s <4us, L1 <64us
                LnkCtl: ASPM L0s L1 Enabled; RCB 64 bytes, LnkDisable- CommClk+
03:00.0 Non-Volatile memory controller: Phison Electronics Corporation E18 PCIe4 NVMe Controller (rev 01) (prog-if 02 [NVM Express])
                LnkCap: Port #0, Speed 16GT/s, Width x4, ASPM L1, Exit Latency L1 <64us
                LnkCtl: ASPM L1 Enabled; RCB 64 bytes, LnkDisable- CommClk+
04:00.0 Ethernet controller: Aquantia Corp. AQtion AQC100 NBase-T/IEEE 802.3an Ethernet Controller [Atlantic 10G] (rev 02)
pcilib:                 LnkCap: Port #0, Speed 8GT/s, Width x4, ASPM L0s L1, Exit Latency L0s unlimited, L1 unlimited
sysfs_read_vpd: read failed: No such device             LnkCtl: ASPM L0s L1 Enabled; RCB 64 bytes, LnkDisable- CommClk+

05:00.0 Ethernet controller: Realtek Semiconductor Co., Ltd. RTL8125 2.5GbE Controller (rev 05)
                LnkCap: Port #0, Speed 5GT/s, Width x1, ASPM L0s L1, Exit Latency L0s unlimited, L1 <64us
                LnkCtl: ASPM L0s L1 Enabled; RCB 64 bytes, LnkDisable- CommClk+

Any ideas would be appreciated

Thanks

myunraid-diagnostics-20260513-2317.zip

Edited May 13May 13 by fka

Used the search in this topic, but did not find my answer.

What should i choose in "Power Mode" in the Unraid settings?

Motherboard: ASRock B365M Pro4

CPU: i9-9900

Power supply: be quiet! Pure Power 11 FM Netzteil 550 Watt 80+ Gold

HDD: 6x (down)

NVME: 1x

USB-Boot-Stick: 1x

I reach C-State 9

Power consumption keeps fluctuating between 35 and 65 W.

Daily power consumption: 1,453 kWh

Edited May 15May 15 by p5lukas

On 5/13/2026 at 5:27 PM, fka said:

My power saving settings have been working fine for years but recently I see this for all scsi hosts

root@myunraid:~# echo med_power_with_dipm | tee /sys/class/scsi_host/host*/link_power_management_policy
med_power_with_dipm
tee: /sys/class/scsi_host/host0/link_power_management_policy: Operation not supported
tee: /sys/class/scsi_host/host1/link_power_management_policy: Operation not supported
tee: /sys/class/scsi_host/host10/link_power_management_policy: Operation not supported
tee: /sys/class/scsi_host/host11/link_power_management_policy: Operation not supported
tee: /sys/class/scsi_host/host12/link_power_management_policy: Operation not supported
tee: /sys/class/scsi_host/host13/link_power_management_policy: Operation not supported
tee: /sys/class/scsi_host/host14/link_power_management_policy: Operation not supported
tee: /sys/class/scsi_host/host15/link_power_management_policy: Operation not supported
tee: /sys/class/scsi_host/host16/link_power_management_policy: Operation not supported
tee: /sys/class/scsi_host/host17/link_power_management_policy: Operation not supported
tee: /sys/class/scsi_host/host18/link_power_management_policy: Operation not supported
tee: /sys/class/scsi_host/host19/link_power_management_policy: Operation not supported
tee: /sys/class/scsi_host/host2/link_power_management_policy: Operation not supported
tee: /sys/class/scsi_host/host20/link_power_management_policy: Operation not supported
tee: /sys/class/scsi_host/host21/link_power_management_policy: Operation not supported
tee: /sys/class/scsi_host/host22/link_power_management_policy: Operation not supported
tee: /sys/class/scsi_host/host23/link_power_management_policy: Operation not supported
tee: /sys/class/scsi_host/host24/link_power_management_policy: Operation not supported
tee: /sys/class/scsi_host/host25/link_power_management_policy: Operation not supported
tee: /sys/class/scsi_host/host26/link_power_management_policy: Operation not supported
tee: /sys/class/scsi_host/host27/link_power_management_policy: Operation not supported
tee: /sys/class/scsi_host/host28/link_power_management_policy: Operation not supported
tee: /sys/class/scsi_host/host29/link_power_management_policy: Operation not supported
tee: /sys/class/scsi_host/host3/link_power_management_policy: Operation not supported
tee: /sys/class/scsi_host/host30/link_power_management_policy: Operation not supported
tee: /sys/class/scsi_host/host31/link_power_management_policy: Operation not supported
tee: /sys/class/scsi_host/host32/link_power_management_policy: Operation not supported
tee: /sys/class/scsi_host/host33/link_power_management_policy: Operation not supported
tee: /sys/class/scsi_host/host34/link_power_management_policy: Operation not supported
tee: /sys/class/scsi_host/host35/link_power_management_policy: Operation not supported
tee: /sys/class/scsi_host/host36/link_power_management_policy: Operation not supported
tee: /sys/class/scsi_host/host37/link_power_management_policy: Operation not supported
tee: /sys/class/scsi_host/host38/link_power_management_policy: Operation not supported
tee: /sys/class/scsi_host/host39/link_power_management_policy: Operation not supported
tee: /sys/class/scsi_host/host4/link_power_management_policy: Operation not supported
tee: /sys/class/scsi_host/host5/link_power_management_policy: Operation not supported
tee: /sys/class/scsi_host/host6/link_power_management_policy: Operation not supported
tee: /sys/class/scsi_host/host7/link_power_management_policy: Operation not supported
tee: /sys/class/scsi_host/host8/link_power_management_policy: Operation not supported
tee: /sys/class/scsi_host/host9/link_power_management_policy: Operation not supported

Bar OS updates, currently 7.3.0, nothing has changed

#!/bin/bash
# Start the Management Utility
/usr/local/sbin/emhttp
cp /boot/config/sensor-quirk /etc/sensors.d

#Start nvidia-persistenced
nvidia-persistenced

#ASPM fix for Aquantia AQC100
sudo setpci -s 00:1c.0 0x50.B=0x43
sudo setpci -s 04:00.0 0x50.B=0x43

#ASPM fix for Realtek RTL8125
sudo setpci -s 00:1d.0 0x50.B=0x43
sudo setpci -s 05:00.0 0x80.B=0x43

#powertop auto tune
#powertop --auto-tune &>/dev/null

# -------------------------------------------------
# powertop tweaks
# -------------------------------------------------

# Enable SATA link power management
echo med_power_with_dipm | tee /sys/class/scsi_host/host*/link_power_management_policy

# Runtime PM for I2C Adapter (i915 gmbus dpb)
#echo auto | tee /sys/bus/i2c/devices/i2c-*/device/power/control

# Autosuspend for USB device
echo auto | tee /sys/bus/usb/devices/*/power/control

# Runtime PM for disk
echo auto | tee /sys/block/sd*/device/power/control

# Runtime PM for PCI devices
echo auto | tee /sys/bus/pci/devices/????:??:??.?/power/control

# Runtime PM for ATA devices
echo auto | tee /sys/bus/pci/devices/????:??:??.?/ata*/power/control

ASPM is all good for my ASM1166 sata controller

lspci -vv | awk '/ASPM/{print $0}' RS= | grep --color -P '(^[a-z0-9:.]+|ASPM )'
00:01.0 PCI bridge: Intel Corporation 12th Gen Core Processor PCI Express x16 Controller #1 (rev 05) (prog-if 00 [Normal decode])
                LnkCap: Port #2, Speed 32GT/s, Width x16, ASPM L0s L1, Exit Latency L0s <4us, L1 <16us
                LnkCtl: ASPM L1 Enabled; RCB 64 bytes, LnkDisable- CommClk+
00:06.0 PCI bridge: Intel Corporation 12th Gen Core Processor PCI Express x4 Controller #0 (rev 05) (prog-if 00 [Normal decode])
                LnkCap: Port #5, Speed 16GT/s, Width x4, ASPM L0s L1, Exit Latency L0s <4us, L1 <16us
                LnkCtl: ASPM L0s L1 Enabled; RCB 64 bytes, LnkDisable- CommClk+
00:1a.0 PCI bridge: Intel Corporation Raptor Lake PCI Express Root Port #25 (rev 11) (prog-if 00 [Normal decode])
                LnkCap: Port #25, Speed 16GT/s, Width x4, ASPM L1, Exit Latency L1 <64us
                LnkCtl: ASPM L1 Enabled; RCB 64 bytes, LnkDisable- CommClk+
00:1c.0 PCI bridge: Intel Corporation Raptor Lake PCI Express Root Port #1 (rev 11) (prog-if 00 [Normal decode])
                LnkCap: Port #1, Speed 8GT/s, Width x4, ASPM L1, Exit Latency L1 <64us
                LnkCtl: ASPM L0s L1 Enabled; RCB 64 bytes, LnkDisable- CommClk+
00:1d.0 PCI bridge: Intel Corporation Raptor Lake PCI Express Root Port #15 (rev 11) (prog-if 00 [Normal decode])
                LnkCap: Port #15, Speed 16GT/s, Width x1, ASPM L1, Exit Latency L1 <64us
                LnkCtl: ASPM L0s L1 Enabled; RCB 64 bytes, LnkDisable- CommClk+
01:00.0 VGA compatible controller: NVIDIA Corporation AD107 [GeForce RTX 4060] (rev a1) (prog-if 00 [VGA controller])
                LnkCap: Port #0, Speed 16GT/s, Width x8, ASPM L1, Exit Latency L1 <4us
                LnkCtl: ASPM L1 Enabled; RCB 64 bytes, LnkDisable- CommClk+
01:00.1 Audio device: NVIDIA Corporation AD107 High Definition Audio Controller (rev a1) (prog-if 00 [HDA compatible])
                LnkCap: Port #0, Speed 16GT/s, Width x8, ASPM L1, Exit Latency L1 <4us
                LnkCtl: ASPM L1 Enabled; RCB 64 bytes, LnkDisable- CommClk+
02:00.0 SATA controller: ASMedia Technology Inc. ASM1166 Serial ATA Controller (rev 02) (prog-if 01 [AHCI 1.0])
                LnkCap: Port #0, Speed 8GT/s, Width x2, ASPM L0s L1, Exit Latency L0s <4us, L1 <64us
                LnkCtl: ASPM L0s L1 Enabled; RCB 64 bytes, LnkDisable- CommClk+
03:00.0 Non-Volatile memory controller: Phison Electronics Corporation E18 PCIe4 NVMe Controller (rev 01) (prog-if 02 [NVM Express])
                LnkCap: Port #0, Speed 16GT/s, Width x4, ASPM L1, Exit Latency L1 <64us
                LnkCtl: ASPM L1 Enabled; RCB 64 bytes, LnkDisable- CommClk+
04:00.0 Ethernet controller: Aquantia Corp. AQtion AQC100 NBase-T/IEEE 802.3an Ethernet Controller [Atlantic 10G] (rev 02)
pcilib:                 LnkCap: Port #0, Speed 8GT/s, Width x4, ASPM L0s L1, Exit Latency L0s unlimited, L1 unlimited
sysfs_read_vpd: read failed: No such device             LnkCtl: ASPM L0s L1 Enabled; RCB 64 bytes, LnkDisable- CommClk+

05:00.0 Ethernet controller: Realtek Semiconductor Co., Ltd. RTL8125 2.5GbE Controller (rev 05)
                LnkCap: Port #0, Speed 5GT/s, Width x1, ASPM L0s L1, Exit Latency L0s unlimited, L1 <64us
                LnkCtl: ASPM L0s L1 Enabled; RCB 64 bytes, LnkDisable- CommClk+

Any ideas would be appreciated

Thanks

myunraid-diagnostics-20260513-2317.zip

I have an ASM1166 on the Silverstone firmware and just upgraded to Unraid 7.3.

I have the same issue. I think I first saw it on Unraid 7.2.5 or 7.2.6 tho.

I found this link: https://forum.odroid.com/viewtopic.php?p=403149#p403149

I ended up adding this to syslinux.cfg:

ahci.mask_port_ext=0000:02:00.0=0xffffffff

0000:02:00.0 being my ASM1166 PCI address.

BTW, I noticed Unraid has a new Boot Parameters menu under the Boot device to modify syslinux.cfg.

I rebooted and that seemed to fix issues with the ASM1166 ports. I haven't seen any issues pop up yet.

Running echo med_power_with_dipm | tee /sys/class/scsi_host/host*/link_power_management_policy still shows a few Operation not supported errors for me.

root@UNS01:~# echo med_power_with_dipm | tee /sys/class/scsi_host/host*/link_power_management_policy
med_power_with_dipm
tee: /sys/class/scsi_host/host0/link_power_management_policy: Operation not supported
tee: /sys/class/scsi_host/host1/link_power_management_policy: Operation not supported
tee: /sys/class/scsi_host/host14/link_power_management_policy: Operation not supported
tee: /sys/class/scsi_host/host15/link_power_management_policy: Operation not supported
tee: /sys/class/scsi_host/host2/link_power_management_policy: Operation not supported
tee: /sys/class/scsi_host/host3/link_power_management_policy: Operation not supported

If I run dmesg | grep -E "ATA-[0-9]|ata[0-9]" it shows those as listed as being DUMMY. I don't think those matter if you can change those or not.