Do i need to upgrade the full build CPU Xeon 3204

Hello all,

Do I need to upgrade the full build CPU Xeon 3204 to run proxmox? The reason I am asking is that the Intel Arc website states Intel® Hyper-Threading Technology is No

I am not sure how much of a performance impact it will be for a home lab set up running about 5 VMs

Welcome to the forums @Isotope2295!

While you shouldn’t need to upgrade just to run Proxmox whether there will be performance benefits really depends on how resource intensive the VMs are and if they are using the same resources at the same time. The stock Xeon processors still have multiple cores, but not the extra pseudo-cores HT would bring for 12 instead of 6 threads.

2 Likes

I mean at the very least you can give the pre-built CPU a try and see how well it does for your workloads. If it seems to struggle or you find that the performance is not up to what your needs are you can always just drop a new one in.

The hardware that is chosen for the HL15 I would guess is ideal for storage first applications since that is what 45 Drives is all about. However if you are going to be loading up TrueNAS, Proxmox, or really any other hypervisor - you can change out the CPU with no issues. (most of the time)

I am sure if you are using TrueNAS or something with a few Docker containers, you would be more than fine.

I’m running Proxmox on an N6005 Pentium Silver (an over-amped 4 core Celeron) and it’s handling several LXC containers just fine; I’m guessing the 3204 has signficantly more oomph than that, and the more powerful Xeons that will fit in that socket are cheap on eBay. :slight_smile:

Hyper threading is not a requirement. In fact, it’s mostly disabled in quite a few Intel generations for various hypervisors due to security concerns.

Run your workload on the system as it comes. If you find you need more performance then look into upgrades.

Or… keep the HL15 as a dedicated storage appliance and let it be good at what it’s built for. Then get yourself a couple SFF PC’s or rackmount servers, cluster them and consume the HL15 as a shared datastore.

You’d get HA, redundancy, and ease of maintenance.

2 Likes