I don't think that vector would work, as only a single tenant's VM is scheduled on any particular core. That's why the minimum number of vCPUs for instances on

I don't think that vector would work, as only a single tenant's VM is scheduled on any particular core. That's why the minimum number of vCPUs for instances on hyperthreaded hosts is 2.