Server Version#: 1.20.2.3402
Player Version#: 1.20.0.1406
Synology DSM 6.2.3-25426u2
DS920+
When will we get a fix for the Goldmont and Goldmont Plus (Apollo Lake, Gemini Lake, and Gemini Lake Refresh) Intel CPUs and hardware transcoding? It’s been 10 months since this issue first cropped up, and so far, there has been no official fix, other than to log on to the Syno command line and rename/delete the file:
/var/packages/Plex Media Server/target/lib/dri/iHD_drv_video.so
While the i9xx driver works, it’s a pain in the butt to do this every time I update the server. At the very least, the faulty driver could be removed from the package until a full fix has been developed.
OK, I understood based on original postings that this first cropped up with the Apollo Lake processors in December 2019 (long before the Gemini Lake based Synologys were released). That’s where I found the solution to rename iHD_drv_video.so. I stand corrected.
I added the delete to my automatic plex package script on my synology ds920+, but would the override survive a package update or is the delete of package just as good? (easier with the auto update script).
The problem with this override is that I have no clue when its actually fixed as hw transcoding working fine without that file installed.
So what method SHOULD we use. Is it a patch in the works, should we override (if this is not overwritten each time) or delete the file and what performance hit do we get by disabling this file?
My point was just I hope it was some kind of dialogue between the Plex developers and Intel so every party can enjoy plex (and probably other things using that driver) without having to modify the software. Its of course not plex fault intel has not fixed it, but not is it the users responsibility to flag and work with intel to fix the issue, that would be the plex dev team
You do understand it was not a hostile comment, just saying I hope this can be solved? I have no issue with the workaround, but a workaround is not a long term solution. Remember people are paying for pass and transcoding and expecting it to work, not having to issue workarounds
Can we have a build of the Plex server with the changes specified in the above link to see if it resolves the issue? I think gmmlib may also need to be updated.
Thanks.
Edit: Both of these changes are included in the latest releases for the media-driver and SDK (2020.03). The Plex Transcoder appears to be compiled with 2019.04.
As a longtime Plex user, and new DS920+ owner, I was reading this thread with interest. Until I read this abrupt “don’t like this country then leave” comment. We can do better than that.
It’s naive to suggest that one person should “talk to Intel” as if that were equal to a big company (Plex) raising the same issue with Intel. Plex has an infinitely bigger stick. Obviously.
May I ask the source of this quote? I’m having difficulty putting the statement in context.
As for the one below it, about one person versus Plex reaching out to Intel, it’s obvious that Plex, as a single entity, has more ‘weight’. That said, if a mere 10,000 individuals reached out to Intel, the impact of their current decisions would be obvious.
@trumpy81, the quote @Plexi-in-OK referenced is correct. I had no intention of creating a lengthy, irrelevant detour off the OP. But it’s worth remembering that we can do better than “like it or leave it,” especially a group of people literally discussing a bug fix.
@RockStar sure, fine. I’m a snowflake. I’m a SJW for saying we can do better. Sorry for treading on you. #boring.
@Plexi-in-OK I also hope Plex and/or Synology communicate the desire for a driver fix, because it’s important for their shared customer base, and that’s important to successful companies.
Attempting to interpret what others are meaning – Contributes nothing.
I asked one user. I did not invite community interpretation
“don’t like the country then leave” is exactly how I interpreted trumpy81’s comment as well, and it’s the only comment on the whole thread that includes this sort of context. The wording of trumpy’s statement too was quite harsh even by the standards of emotionless text forums. Trumpy could have easily responded by simply directing attention to FordGuy61’s response earlier (with the link to your FAQ) and left it at that (or even not responded at all, since the conversation was off of the proposed workaround, and on to Intel’s driver issues), rather than start implying that editing the preferences file was “too much” for us.
Well. I can see what triggered Trumpy, because there is a sense of entitlement in the statement he replied to that borders on ignoranse.
You buy a NAS from a maker that ships it with broken drivers, and a third party software company is somehow expected to take action to fix this? Id say providing a workaround like they have, is going above and beyond.
Plex charges for Plex pass in order to facilitate hardware transcoding, but a pre-requisite is having hardware that actually supports it. It seems these complaints would be better directed towards the NAS makers for selling a half-assed product.
I completely disagree. I’m not blaming Plex for Intel can’t fix the driver on that cpu, but synology is an officially supported nas. As a developer myself I would never say to my users, yes I know we have a problem, but go yourself and report it to Intel. It would be my job to report it and try to find a solution.
We might disagree and it might have been misunderstood. I’m not sitting here like a kid and saying I want Plex to fix this now, because I’m entitled and I don’t care if Intel is the reason. I’m just saying if you are developing a software that require a driver that is known broken is more professional that company go to Intel than 200 users sending random emails to Intel.
Take it as you want, ignorant or not, but I don’t see why it’s ignorant saying it’s not the users job to fix a driver that is used by a paid software
