Server Version#: 1.20.0.3181
Player Version#: (not relevant)
I’m testing two servers, both on the latest Plex Beta channel. One is a 2018 Mac mini 6C/12T i7 with 32GB of RAM and the built-in Intel UHD Graphics 630, the other is a custom Windows machine with a 16C/32T 1950X Threadripper with 64GB of RAM and a 1080 Ti. I’m testing both with the same (locally accessed for each machine) 4K MKV file and both have similar RAID-0 arrays with a couple of 7200RPM drives each. The Mac mini has the disks attached via a Thunderbolt 3 enclosure, the PC has the disks attached to the native SATA 6G ports on the motherboard. Both machines are set to use hardware transcoding in the Plex server settings, which I know is working on both because during a transcode, neither are barely tapping the CPUs.
The interesting thing I’m noticing is:
The Mac mini is WAY more responsive when starting a transcode and skipping around the timeline is WAY faster to start the stream again. This is across all the clients - Chrome-based browsers, iOS apps, tvOS app, etc.
I’m just wondering how a little Mac mini using Quick Sync can outperform a beastly PC with a 1080 Ti for transcoding. Is Intel’s Quick Sync actually faster (and/or more efficient) than NVENC on the GPU?
Would love to hear and insights from people who have had similar results or know of any tips or tricks I may be missing.
Thanks
