Weird graphics slowdowns triggered by YouTube

Framework Laptop 16 Linux
1

Which Linux distro are you using? Arch Linux

Which release version?
(if rolling release without a release version, skip this question)

(If rolling release, last date updated?) Last week (~25 september 2024)

Which kernel are you using? 6.10.10

Which BIOS version are you using? 3.03

Which Framework Laptop 16 model are you using? (AMD Ryzen™ 7040 Series)
7940HS, 96 GB RAM, no dGPU.

Laptop was received last week.

Sometimes, when watching YouTube (it’s possible that the issue triggers in other situations too, but I have not seen that happen yet), my FW laptop’s display output becomes incredibly laggy / slow at refreshing (we’re talking on the order of a few refreshes per minute). This persists if I close Chrome, switch to the console, or restart the graphical user interface; the laptop recovers after a reboot.

Neither dmesg nor journalctl show anything obviously suspicious.

I am running wayland, with google-chrome-beta running directly on wayland too (but the issue has happened with ozone-platform set to x11 also)

2

In have had this with the stock kernels. Seems that it switches from accelerated mode to non accelerated.
Make sure you update the BIOS to 3.0.4 and enable “gaming” mode in BIOS (GPU assigned memory).
Of course, have the latest AMD drivers installed.

That made it for me.

3

Isn’t the latest BIOS 3.0.3?

I already have the GPU memory set to the highest value.

4

For the FW16, it is 3.0.4.
Try out: sudo fwupdmgr update
It should update it for you.

Handle 0x0000, DMI type 0, 26 bytes
BIOS Information
        Vendor: INSYDE Corp.
        Version: 03.04 <= BIOS Version
        Release Date: 07/09/2024
        Address: 0xE0000
        Runtime Size: 128 kB
        ROM Size: 32 MB
5

Interesting, fwupdmgr and dmidecode claim to be at 3.3 and fwupdmgr claims to be up to date. I guess I’ll try out the beta BIOS :wink:

6

I thought it was in the GA release already. I have it running

7

Sadly, this is a known bug. in the Radeon 600M (Ryzen 6000 / 7035 APUs) and 700M (Ryzen 7040 APUs). There are several reports on freedesktop gitlab, such as this one.

It appears that AMD has been trying to mitigate this issue in newer versions of the AGESA and PSP firmware, but the cadence of firmware updates on the Framework 16 is not exactly leaving me hopeful for a timely fix. A friend’s HP Elitebook 845 has received several AMD firmware upgrades in the time frame where my 16 has received none, and he claims that one of the more recent AGESA versions seems to have fixed the issue for now on his 7840U.

Could anybody running the new BETA BIOS go into the BIOS Setup settings and report back on what versions of the AMD AGESA, PSP and GPU firmware are running on it? One of the above comments suggests the error “seems to be gone” on 3.04, it would be helpful to see if this perceived improvement is also connected to an update in AMD’s firmware, as it seems to be on the HP laptop I am referencing here.

8

Well, looking at dmidecode AGESA!V9 PhoenixPI-FP8-FP7 1.1.0.2a, but nothing for PSP and GPU.

9

The issue has happened again after the upgrade to the beta BIOS, unfortunately.

10

Not amazing. This is the same ancient AGESA version that is in 03.03.

11

It is bound to. The new beta BIOS still does not upgrade the ancient AGESA version, while almost every other laptop vendor with this chip has moved away from this version. Until Framework gets it together with firmware upgrades, the crashes will continue.

12

You opened a case to Framework Support for that?
Maybe they are just not aware of that.