Scenario: IT management has decided which server hardware model to use for new XenApp and XenDesktop Sites and needs guidance on sizing the Hardware
Layer.
Click the Exhibit button to view the host details.
A Citrix Architect is planning to install XenServer 7.x on this host which will be used to host only XenApp servers.
How many vCPUs should be consumed to deploy XenApp servers with high performance, considering the over-subscription ratio?
Given the host details, the server has 2 sockets with 8 cores per socket, amounting to 16 physical CPUs (pCPUs) without considering hyperthreading. The specified CPU over-subscription ratio is 1.5:1, indicating that for each physical CPU, 1.5 virtual CPUs (vCPUs) can be allocated. Calculating based on physical CPUs: 16 pCPUs * 1.5 vCPUs per pCPU equals 24 vCPUs. Therefore, the appropriate number of vCPUs that should be provisioned for XenApp servers in this scenario is 24.
1,5:1 = 1,5 means 1,5 vCPU to 1 pCPU. 2 sockets x 8 cores = 16 cores with HT enabled = 32 pCPU / 1.5 = 24, so the answer should be B and not C
I think B is correct but not with that calculation, because: 1. some articles descrive: You should consider pCPUs (including Cores) only. Hyper-threading does not actually double the available pCPU. Hyper-threading works by providing a second execution thread to a processor core 2. 1,5:1 = 1,5 vCPU's to 1 pCPI (2x8) x2 = 32 1,5:1 so you need to do 32x1,5 = 48 (But this is no correct because hyperthreading does not multiply pCpu Correct Calculation 2 sockets x 8 cores = 16 pCPU's 1,5 vCPU per 1 pCPU, so 16x1,5 = 24.
Correct Calculation 2 sockets x 8 cores = 16 pCPU's 1,5 vCPU per 1 pCPU, so 16x1,5 = 24.
B. 24 is right