The CPU in this instance is a Ryzen 7 1700 with 8 Cores + 8 virtual cores. I kept a running log of CPU activity at time of freeze, and none of the cores were ever near 100%, so it didn’t look like a CPU loading problem, although CPU hardware faults are difficult to trace from logs. The CPU seems to perform really well under heavy AI loads, with no freezing using other softwares, so I think it’s ok.
Same with RAM. System load never went about 10%, and RAM checks were GOOD on every test.
Thanks for the note on the RX-460. Do you mean that PMS doesn’t use it for hardware acceleration by choice? It seems to be a capable, but not muscular hardware accelerator. It works perfectly doing Python Neural Network calculations, so I thought it would be a good candidate for PMS hardware acceleration. Guess not.
The RTX-2070 is indeed a chore to install drivers on, although I tried several different recent drivers to try to stop the random machine freezes. Again, the hardware acceleration worked for other demanding AI Neural Network softwares, so I was running PMS with transcoder hardware acceleration ENABLED, and the machine was freezing randomly.
Right now, my machine has the RX-460 installed, with transcoder hardware acceleration DISABLED, and for the first time in a long while, the system is stable using two Silicon Dust units (4 tuners total) that frequently have three tuners active simultaneously.
I haven’t tried reinstalling the RTX-2070 under Ubuntu 20.04 LTS running PMS with transcoder hardware acceleration ENABLED, because I don’t have confidence in the hardware acceleration feature as of yet. For now, system stability is good with the RX-460 doing CPU transcoding, so I’m staying with it.
Have been a PLEX user since inception, and it’s frankly my go-to media platform. It’s been running 24/7 in my house for years now.
