Need help reaching Package C8 on Intel N100

[Che0063]

I also have a old laptop from Chuwi Corebook X with 8th Intel CPU, it's also stuck at package C2 when boot with default M.2 Sata SSD (already installed Intel 300 AHCI driver, notice that it doesn't have any HIPM/DIPM/DevSleep feature from SSD firmware), package C3 with nvme ssd. The bios also unlocked all advanced settings but nothing help to enable package C8 until I've removed all SSD from M.2 sata and M.2 nvme slots, boot windows from USB and package C8 appear normally.
View attachment 350495

I remember having this issue (NVMe + C3 limited). I made a post years ago about enabling the 'Lowest' ACHI Power Management but that post got deleted when NotebookReview got shut down. Can you try setting https://www.tenforums.com/tutorials/72971-add-ahci-link-power-management-power-options-windows.html ACHI Link power options to Lowest?

I do specifically remember this to fix the issue on my 6th gen U-series platform with a SATA SSD.

At least you've isolated the issue to the NVME subsystem though...

 

[nhl59623]

I remember having this issue (NVMe + C3 limited). I made a post years ago about enabling the 'Lowest' ACHI Power Management but that post got deleted when NotebookReview got shut down. Can you try setting https://www.tenforums.com/tutorials/72971-add-ahci-link-power-management-power-options-windows.html ACHI Link power options to Lowest?

I do specifically remember this to fix the issue on my 6th gen U-series platform with a SATA SSD.

At least you've isolated the issue to the NVME subsystem though...

I already enable HIPM+DIPM tweak but my m2 sata ssd (NGFF form) doesn't support any these features and stuck at C2. For nvme issue I guess Chuwi just use nvme lane from cpu and it make cpu idle stuck at package C3, I've tested on my desktop, move nvme ssd from first slot (cpu lane) to second slot (pch lane) also help enable package C6.

 

[tmkhiem]

@tmkhiem
Just curious. When you put your computer to sleep for a few minutes and then resume, does ThrottleStop show anything in the Sleep Residency section of the C States window for package C8, C9 or C10?

I just tested this and there's still nothing in Package C8, C9 or C10.

The only bad thing is that this forum thread was originally about the Intel N100. ThrottleStop and HWiNFO may not be reporting the C state activity of the N100 correctly because Intel has not released updated documentation that covers the N100.

I agree.
In the first post, I tried to follow the latest documentation from Intel (that's how I came up with the MSR addresses), but I did not notice that it does not apply to N100 family.
I'm trying to replace the AirDisk inside with an Intel nvme SSD disk to see if it works.

UPDATE:
I tried replacing the AirDisk (controller MAS1102B) with an Intel SSD 760P. However, I couldn't get the 760p to work, despite enabling back all PCIe ports and later loading the BIOS defaults. The NVMe menu still shows "no NVMe drives found". I'm not sure if it's because of the BIOS or my 760p is really old and possibly broken.

I ran HWInfo on the drive and there is some interesting info: many power saving features are not supported. I think this could aggravate the high power idling problem.

 

[Che0063]

I just tested this and there's still nothing in Package C8, C9 or C10.

I agree.
In the first post, I tried to follow the latest documentation from Intel (that's how I came up with the MSR addresses), but I did not notice that it does not apply to N100 family.
I'm trying to replace the AirDisk inside with an Intel nvme SSD disk to see if it works.

UPDATE:
I tried replacing the AirDisk (controller MAS1102B) with an Intel SSD 760P. However, I couldn't get the 760p to work, despite enabling back all PCIe ports and later loading the BIOS defaults. The NVMe menu still shows "no NVMe drives found". I'm not sure if it's because of the BIOS or my 760p is really old and possibly broken.

I ran HWInfo on the drive and there is some interesting info: many power saving features are not supported. I think this could aggravate the high power idling problem.
View attachment 350548

I am not 100% sure this is correct but your existing AirDisk SSD might be a M.2 SATA SSD. (and forgive me if you've already double checked - does AirDisk show up under Drives > NVMe or (S)ATA/ATAPI Drives in HWInfo64?)

M.2 refers to the physical slot form factor, whereas SATA or NVMe refers to the protocol by which the disk communicates the system. If your AirDisk uses the SATA protocol, you might consider trying a SATA SSD. NVMe and SATA are often not interchangeable in laptops.

Also you've been very brave changing BIOS settings ahaha, especially on a fully unlocked BIOS. Just be aware that some settings can irreversibly brick your laptop so try to be a bit careful!

 

[tmkhiem]

I am not 100% sure this is correct but your existing AirDisk SSD might be a M.2 SATA SSD. (and forgive me if you've already double checked - does AirDisk show up under Drives > NVMe or (S)ATA/ATAPI Drives in HWInfo64?)

M.2 refers to the physical slot form factor, whereas SATA or NVMe refers to the protocol by which the disk communicates the system. If your AirDisk uses the SATA protocol, you might consider trying a SATA SSD. NVMe and SATA are often not interchangeable in laptops.

Also you've been very brave changing BIOS settings ahaha, especially on a fully unlocked BIOS. Just be aware that some settings can irreversibly brick your laptop so try to be a bit careful!

Since I am able to desolder the BIOS and have the tooling to backup/flash its content and reset the EEPROM, I think I can be more aggressive towards tweaking the settings. But anyway, thank you for the heads up.

The AirDisk has the form factor of M2 2280 key B+M, and indeed a SATA SSD, indicated by:

• Empty (non-routing) PCIe pins, observed on the disk.
• The controller being Maxio MAS1102B, a confirmed SATA controller.
• The controller ("parent") node is Intel AHCI controller instead of any kind of PCIe root port.

However, the slot that the motherboard exposes:

• Is a key M only connector.
• Has routing of at least 1 PCIe lane visible (only the first transmit pair visible, however), apart from 2 SATA signal pairs. To fully verify this, a boardview or desoldering of the slot would be required.

During the testing, I found out Others posted in the Chuwi forum that it did not support NVMe. However, with the traces already there, I think it could be possible to have a functional NVMe slot by populate missing components.

However, this could be tangent to the main topic that I started -- I originally thought the N100 was the culprit. It may not be, but right now between the disk issue and the cpu issue, I think this thread would be better focused on the N100.

 

[nhl59623]

However, this could be tangent to the main topic that I started -- I originally thought the N100 was the culprit. It may not be, but right now between the disk issue and the cpu issue, I think this thread would be better focused on the N100.

I just read a topic on Reddit about N305 (use same Alder Lake N core as N100) and it can reach to C10 state, so I think there is no problem with cpu.
https://www.reddit.com/r/homelab/comments/1bd892o

 

[Marctraider]

I recently acquired a Surface Go 4 with N200 CPU, C10 package residency hits 92% with some optimization.

I also had a Razer blade stealth 2020 a while back, also managed C10 state (75% ish). (Entirely different and older CPU)

In my experience, when assuming BIOS/UEFI supports C10, its always the drivers when C10 aint working. You should never ever disable devices in Windows. In some events you can disable them in BIOS but it depends on the method of disabling (or simply hiding from OS. In which case you will also see regression)

By disabling drivers, the device will remain in ACPI D0 state after a reboot.

A lot of devices come with driver packs, which should include support for all devices connected to the CPU I/O. Sometimes the packs are outdated, have old drivers, sometimes the microsoft drivers from updates are generic or also outdated.

Very seldom software (load) is to blame. Wost case there is a screwed bios/uefi implementation, in which case it can mostly only be fixed by modding the uefi. But this should really be fixed by the vendor itself.

Last but not least, there (was?) a C10 reporting bug, but this usually occured after waking up from old skool standby, the CPU stopped reporting C10 but it was in fact still active. (Measured from laptop discharge rate).

Anyway, to sum it up, ALL recent mobile intel CPU's should get into C10 state. If not, something is amiss.

 

[mkdr]

Last resort - try booting from an external HDD/SDD or any boot medium that is NOT your AirDisk SSD - once I had a Teclast F6 m3-7Y30 laptop with a no-name Chinese SSD, which likely didn't support the power states that the CPU needed to enter anything below C3, resulting in a 1.1W idle. After changing it to a Transcent SSD, suddenly C8 became available

That was already what I said above in my posts and my #1 theory, that the SSD is not using low power state, causing the PCH/PCIe link be in a higher state/link all the time, causing C8/C10 not to work, resulting in a permanent ~2w drain.

Is it possible that there is a bug or some other limitation that prevents the E cores from using core C7? The N100 has 0 P cores and 4 E cores. Maybe the limited C states has something to do with the N100 being an E core only processor.

It might also be, that all these cheap N100 mini PCs use cheap SSDs which dont support low energy state.

Intel SSD 760P.

I read somewhere before that some Intel SSDs have some bug or false pin layout making them compatible with a lot of M2 slots. Might also be, your N100 device just uses M2 with no NVMe support just SATA M2.

By disabling drivers, the device will remain in ACPI D0 state after a reboot.

This is correct. I have a script because of this on my Dell XPS 15, disabling the Nvidia dGPU causes it to stay in high energy state, always fans on, and 5w drain. Fix is to enable the device and disable it again. But you need to do this after each reboot. I noticed though, that this seems to be fixed since I updated to Windows 11 24h2, need to test again in the next days.

 

[tmkhiem]

I have finally cracked the code.

The problem was the 512GB AirDisk.

I ordered a 1TB WD SA510, transferred the contents from the AirDisk over (I really do not want to experience the process of installing drivers for the Chuwi for the second time). Upon the first boot, there are some C8 residency immediately. At this point, I am happy. I am writing this in the first boot of the new SATA drive, with at least C8 residency.

In the previous weeks, I ordered another Chuwi, this time an Xpro laptop with essentially the same motherboard and more importantly, proper M2 nvme support (it came with a cheap 256gb nvme ssd). I was able to highlight the differences in the pcb, and will post a snapshot of the pcb with some measurements soon that may help someone figure out the process of adding nvme support to Gemibook. I do believe the differences come down to some SMD resistors, caps and mosfets.