I have an Intel CPU with embedded iGPU (i5 1335u) on my proxmox host and 16Gb of ddr5.
In host bios, i can chose 128-256-512-1024 memory aperture for the iGPU. This amount of memory is later no more available in proxmox host (reserved)
I wonder, if the GPU is shared between several VMs ( vGPU ) how ram is assigned/allocated.
As far as I understand, inside a single VMs, the vGPU ask for "v-ram" according to its needs,
but is it limited to what is specified in BIOS ? Probably "no" each VM can allocated v-ram up to max-ram assigned to VM is reached: but, if it is so, what about the reserved memory specified in BIOS ? Which VMs will take it or how they share it ? Is it wasted ? It would be better to keep this size at minimum ?
In host bios, i can chose 128-256-512-1024 memory aperture for the iGPU. This amount of memory is later no more available in proxmox host (reserved)
I wonder, if the GPU is shared between several VMs ( vGPU ) how ram is assigned/allocated.
As far as I understand, inside a single VMs, the vGPU ask for "v-ram" according to its needs,
but is it limited to what is specified in BIOS ? Probably "no" each VM can allocated v-ram up to max-ram assigned to VM is reached: but, if it is so, what about the reserved memory specified in BIOS ? Which VMs will take it or how they share it ? Is it wasted ? It would be better to keep this size at minimum ?