I have 5 servers with several Windows VMs. However, some of them use Nextcloud, and from what I've noticed, it heavily utilizes graphical processing, approximately 20% of a GT 1030, for example (quite significant for a web server). However, on each server, I have a dedicated GPU for passthrough, but I use them on other machines. Is it possible to share my GPU among Windows machines without attaching it to a specific machine? I need all of them to use it without any performance issues. or.......... maybe i need virt GL or Virt IO in windows vm(i never tested before).
I need help with gpu share
- Thread starter gabrielrene
- Start date
Hello,
there are two technologies you could use:
Are you sure you need a GPU for a Nextcloud server? Since most of the GPU utilization is on the client side.
there are two technologies you could use:
- Intel GVT-g with Intel APUs (5th, 6th, 7th generation Intel Core & E3 v4, E3v5, E3 v6 Xeon) https://pve.proxmox.com/wiki/PCI(e)_Passthrough#_mediated_devices_vgpu_gvt_g
- NVIDIA vGPU which only works on specific enterprise GPUs https://pve.proxmox.com/wiki/NVIDIA_vGPU_on_Proxmox_VE_7.x#cite_note-2
Are you sure you need a GPU for a Nextcloud server? Since most of the GPU utilization is on the client side.
Last edited:
Hello Maskus, good morning, maybe it wasn't clear but it's on the client side, I need the gpu on the client side, as the resource usage occurs when the client is browsing nextcloud, it's a slightly insignificant usage but it affects the customer experience and this is a problem because when this happens, nextcloud has lags and non-smooth transactions when I add the gpu it uses about 20% of the same but it is 70% more fluid