Hardware Transcoding issues - ApolloLake & GeminiLake CPUs

Tested on RocketLake. working fine :slight_smile:

1 Like

Not sure what my CPU code name is (DS920) but still not working for me. I have peeps watching now but can gather logs tomorrow if needed.

To confirm - I no longer have i965 quoted anywhere in my preferences either, I put a copy of the driver in the Cache folder after I saw it wasn’t working and restarted and it’s fine now. Hopefully my CPU wasn’t one of the ones that was fixed yet.

DS920+ Intel Celeron J4125 4 4 ✓ Geminilake DDR4 4GB

If I understand everything correctly, the J4xxx CPUs are tagged for the driver to download anyway. It should be there in case it’s needed.

We worked on fixing the ApolloLake (J3355 / J3455) CPUs as they are widely used on QNAP and Synology

1 Like

So if I understand correctly, I shouldn’t be having an issue?

nope. it should all be good.

I’ve not been able to follow it as i normally do but i trust.

In any case, keep the SPK file you have just in case.
If it does go sideways, you still have the drive and SPK.

With your logs before you roll back, we’ll get the data

Hi all,
Any news about a potential delivery date ?

Tested the version 1.32.4.7164 in my Synology DS218+ (J3355), the problem persists.

Hi !
Tested the Version 1.32.3.7162 in my Synology DS918+ (J3455), the problem persists.

For those expecting the current Plex Pass to fix the issue -

  1. The corrections were not merged into the master CM until last night.
  2. What you see in PlexPass predates the correction by a few days.

You’re welcome to use the Engineer’s build until the next PlexPass build

2 Likes

Hi @ChuckPa Thank you for your answer.
I use PLEX through docker. How do I get the latest image?

Just do a ‘pull’ of the image you want or plexinc/pms-docker:latest

You can also pull any version which still exists upstream

(this one isn’t guaranteed)

plexinc/pms-docker:1.29.2.6364-6d72b0cf6
1 Like

Thanks ChuckPa, I’m now using this build and it is working flawless.

1 Like

is the version with the fix 1.32.3.7162-b0a36929b?

No. See @ChuckPA’s next to last post and the one above (scroll up 4 - 5 posts).

Also, fix is not in 1.32.4.7164 (which does fix Intel Core ix-3xxx to -5xxx).

Wait for the next beta.

Do you know when the new beta with corrections comes for Plex Pass?

Maybe as suggestion, a proper beta should come with proper release notes as per feature changes integrated, instead of the package of changes for the final release…

Some additional work but brings clarity to clients:-)

The new release will be ‘cut’ on Friday (9-Jun-2023).
I’ll be able to grab an alpha-test copy at that point.

It will then be in test for 10 days, scheduled PlexPass on 19-Jun-2023 if there’s nothing requiring rework.

This build will be scrutinized (I’ll be sharing with folks so we know we have HW transcoding solid again… it did get a lot of attention since last build)

4 Likes

Thank you @ChuckPa. The Engineer’s build Version 1.32.5.7145 fixed the hardware transcoding on my DS218+.

1 Like

Is it possible for anyone to grab this engineering build? I just noticed this myself and had to back off all the way back to 1.32.1.6999 to get hw transcoding working again… I hadn’t really noticed until this morning and saw a user transcoding and no (hw)… I had been noticing my plex cpu running higher but hadn’t spent any time putting 2 and 2 together…

Or since today is the 9th, will we be able to test this new release cut that @ChuckPa mentions would be released today?

edit: doh, scrolled up and saw the link to the eng build - that seems to be working on ds918+… Thanks!!

edit2: Sure glad I keep old versions going way back :wink:

Synology NAS DS920+ here.

Just discovered the issue today when my device started alarming for temps.

Transcoding was all being done on CPU.

Years ago I did a setup of my system using the Synology Docker Client (PMS, and a few other tools)

I had to tinker with the i965 lines and also make sure some /dev/pri lines were somewhere (it was years ago and part of my one time setup, I don’t tinker in the setup, drivers, and logs like many of the power users here.)

I update my PMS by going to the Synology Docker GUI, stopping my PMS container, click the button to clear the container, clicking the repository tab, search for Plex, grab the linuxserver.io PMS latest, then restart the container from the GUI.

PMS updates perfectly this way. Been doing it for years now.

I’m now at 1.32.3.7162 according to my general setting page.

I can deal with telling my family and friends that they can’t use Plex for a little while until it gets fixed, but with all of today’s reading and discoveries I’m a little concerned I’m going to need to relearn all the tiny details and tweaks I used to get it working perfectly a few years ago.

I’d really hope it’s just a matter of waiting until another new release drops that fixes this stuff.

Can I leave my system alone and wait for a fix? Or do I need to go in and start playing with drivers, config file settings, and learning how to compose new containers that I’ve completely forgotten how to do.

Does anyone have any easy ways to tell a container to use an old image? I downloaded the .6999 image through the Synology Docker GUI but I don’t see any easy way to edit the PMS Container to point to that image instead of the one marked “latest”.

As you can see in the picture, two streams completely maxed out my CPU.


I have not seen the new Docker GUI. (My DSM 7 system can’t have docker – ARMv8)

Unless you specifically see in the new GUI where it allows you to PASS THROUGH devices then it won’t.

Upgrading to a newer image in container is easy (docker pull)
Downgrading requires:

  1. All your data is maintained outside the container mechanism (like in a shared folder somewhere)

  2. You delete the container and image

  3. You create the new container,
    with device(s) mapped in,
    and with your existing /config path,
    specifying the version you want, perhaps 1.32.1.6999-91e1e2e2c

  4. This is best done via the command line.
    ( you can also do this in a “Task Scheduler” → User-script → which runs as ‘root’ user )

make sense?

As for the thermal alarms…

  1. Shut it down
  2. Take off the cover
  3. Blow out all the dust.

(It takes a lot to overheat a syno but it happens quickly when there isn’t enough air flow – OR – fan speed was manually turned down)

3 Likes