Arch Linux User Repository

NPU has now been tested & works! https://wiki.archlinux.org/title/ASUS_Zenbook_UM5606#NPU

@VanKurt you might also need to do install/enable these services: https://wiki.archlinux.org/title/ASUS_Zenbook_UM5606#Power_management

P.S. it's normal/expected behavior for linux-firmware-git to include amd-ucode-git :)

A reminder to please take discussions about these patches on other distros to an appropriate channel :)

@Boulate With MAKEFLAGS="-j$(nproc)", my system took exact 30 mins to compile the kernel.

Meanwhile, I found my system becomes not stable with the kernel 6.11. I encountered quite a few system hangs on EndeavourOS Gnome 47 (also on Fedora 41 beta, but less frequent).

@Boulate:

Thanks for the tip on the flags. Tomorrow I'll replace the SSD and have chance to try it.

My current BIOS version is 309. It certainly makes my system cooler and quieter.

@bob418 :

First : Are you certain that you have activated the flag MAKEFLAGS="-j$(nproc)" (instead of the default MAKEFLAGS=1) so that all your cores work on this compilation? Because with the same machine, it takes me about 20 minutes to compile the kernel.

Second : I think they have limited the maximum TDP of the Zenbook, because I have a Vivobook M5606 (same processor, etc.) and by limiting the processor to 15W, I get exactly the same results as on the Zenbook (which I also have).

If that's the case, it's very weird : they let the tech world do benchmarks, extremely positive feedbacks about performance... and then they restrict it.

I own both and made lots of comparison between Zenbook and Vivobook of this series (graphics, proc, battery, etc.)

With this limitation the Zenbook have no sense at all : more performance if needed in the vivobook (by putting the TDP to 28w or more), same battery life by aligning the TDP (15w), less heat, a numpad, and a cheeper price...

@Boulate I think the new BIOS in S16 limits the performance to keep the laptop cool. I built this package on a new installation (Kernel 6.10.7) and it took 4 hours to complete. But the laptop's fan was never loud. Last time it needs just over 3 hours but the fan was really loud.

I've seen a dramatic drop in performance since kernel 6.11.1 (after switching from 6.11-rc7 to 6.11.1).

Examples: https://openbenchmarking.org/result/2409160-BOUL-ZENBOOK31,2409196-BOUL-ZENBOOK08

@That1Calculator: Have you seen this too?

PS: I've also updated the bios in the meantime, from version 305 to 308. Could this be related? Could they have lowered the machine's performance following the various feedbacks related to the device heating up?

Hey @Boulate @thomasgassmann -- sorry about that! I accidentally used the checksum from the output of b4 to fetch the patch instead of the direct link. Fixed now!

Having the same issue as @Boulate. Updating the hash fixes it.

For information :

The sha256sum of the file keyboard_bl.patch downloaded via yay is "c56b692379403d493d6455400c868b1b7441a5c382a2002fc5019b379b108c9a". When I try a wget https://patchwork.kernel.org/project/platform-driver-x86/patch/20240909223503.1445779-1-bas@basnieuwenhuizen.nl/mbox/ , I have the same file (so same sha256sum).

The sha256sum you seem to seek on your PBGBUILD for keyboard_bl.patch is 3330dfc9c8dc49c470041f69f9558af69b7b6de2de5230a274e606754cef645a .

I think (maybe i'm wrong) that there is a problem in the PKGBUILD file (line 50).

Hi @That1Calculator !

First : Congratulations again for your work.

I have trouble this morning with the 6.11rc7-1. Are you aware about that ?

The problem : The sha256sum seems to be incorrect for the keyboard_bl.patch. I tried multiple times, even with a "ALL CleanBuild" so i downloaded it again (3 times).

--- LOGS ---

``` ==> Validating source files with sha256sums...

linux-mainline ... Skipped

config ... Passed

bluetooth.patch ... Passed

keyboard_bl.patch ... FAILED

npu.patch ... Passed

zsc_control.patch ... Passed

==> ERROR: One or more files did not pass the validity check!

-> error making: linux-mainline-um5606-exit status 1 ```

Nevermind, I reinstalled VirtualBox (after already reinstalling multiple times) and it just decided to work this time.