Server Upgrade Questions


Recommended Posts

Hi folks, 

 

I've finished my first ever server, and everything is working quite well (thanks, entirely, to Spaceinvader One's amazing videos and the great community here!!!). Of course, now that it's up and running, I've immediately started thinking of the ways in which I could improve it (I just had so much fun with the build that I don't want the fun to be over--I'm sure y'all can relate). My use case: Plex (for me and around 5 pals, probably only a few streams at once), Nextcloud, possibly some VMs down the line, maybe more as my knowledge of what I can do with Plex increases.  Anyways, the possible weak points, it seems to me, are the GPU (GeForce 710) and the RAM (8GB Patriot DDR4). So, here are my questions: 

 

Is it worth upgrading to, say, 32GB ECC RAM or is 8GB sufficient for my use case (I'm really new to all this and so cannot gauge what I might need in the future).  

Provided that HW accelerated transcoding is compatible with my AMD cpu (I've found no clear answers online), is it worth upgrading the GPU to, say, a GTX 1060? Something else? 

 

Thanks in advance for your responses (and apologies in advance if this topic has been covered elsewhere).

 

Onward, 

 

pb 

Link to comment
19 hours ago, pnbrooks said:

Provided that HW accelerated transcoding is compatible with my AMD cpu (I've found no clear answers online), is it worth upgrading the GPU to, say, a GTX 1060? Something else? 

Hardware accelerated transcoding is not really CPU dependent unless you want to use an integrated GPU to do it.  This is what I do and requires an Intel processor.  The GPU in the AMD APUs are not supported by Plex for hardware transcoding, although some claim they have been successful in getting them to work.

 

If you have an AMD CPU, you need an Nvidia GPU for hardware transcoding.  The GTX 1060 is supported but is limited to two transcoding streams.  For "unrestricted" transcoding streams, you need a Quadro P2000 or better.   This will require you to use the Unraid Nvidia plugin and associated UnRaid build with Nvidia drivers.

 

Software (CPU)-based transcoding capability is measured by passmark scores and requires no GPU.  You need 2000 passmarks per 1080p stream (you need 17,000 for a single 4K HEVC 10-bit stream).  My CPU (Xeon E-2288G) has a 17,000 passmark score, so, theoretically, I could do about 7 simultaneous 1080p streams with the CPU (unRAID needs about 2000 passmarks for its overhead as well).  Since it has an iGPU, I use that for hardware transcoding and the CPU doesn't break a sweat.

 

If you want to know what your CPU can do, check out its passmark score.  Here is a link to the passmark score for my E-2288G.

Edited by Hoopster
Link to comment
54 minutes ago, pnbrooks said:

Is it worth upgrading to, say, 32GB ECC RAM or is 8GB sufficient for my use case (I'm really new to all this and so cannot gauge what I might need in the future).  

8GB is the minimum I would recommend for running docker containers.  It all depends on what you plan to run.  For multiple dockers (some of which can use a fair bit of RAM) you may be more comfortable with 16GB RAM. 

 

When you start introducing VMs, depending on what they are, you may want 32GB, especially if they are VMs that are "always on or if you want to do transcoding in RAM.

 

Although it is completely different hardware than what you have, my prior system, before my last upgrade, did some things similar to what you want to do.  It supported ~20 docker containers, Plex hardware transcoding for 5 households, some light use VMs, backups of my PCs/laptops, etc.

 

I had a 4-core/8-thread Xeon with iGPU and 32GB ECC RAM in that system.

Link to comment

Join the conversation

You can post now and register later. If you have an account, sign in now to post with your account.
Note: Your post will require moderator approval before it will be visible.

Guest
Reply to this topic...

×   Pasted as rich text.   Restore formatting

  Only 75 emoji are allowed.

×   Your link has been automatically embedded.   Display as a link instead

×   Your previous content has been restored.   Clear editor

×   You cannot paste images directly. Upload or insert images from URL.