HP ProLiant BL660c AMD Opteron™ and Intel® Xeon® x - Page 12

Virtualization

Page 12 highlights

Power Cap Manager Power Cap Manager allows you to set a fixed limit on a server's processor power consumption by controlling the P-state of individual cores. The caveat is that a single voltage is supplied to all cores in the processor even if you request different P-states. In this instance, the actual P-state will equal that of the highest voltage required of the selected P-states. You can control this through HP Dynamic Power Capping. You can access Dynamic Power Capping from iLO Advanced menus or the HP Insight Control (ICE) management suite. Virtualization In a virtual machine (VM) environment, a software layer known as the hypervisor makes it appear that each guest OS has full control over the server's processor, memory, and I/O devices. Actually, each guest OS has its own virtual processors, virtual drives, virtual NICs, and virtual storage controllers. Several categories of software-layer abstraction exist. AMD-V and Intel VT technologies fall in the hardware-assisted-virtualization category. These technologies assist the hypervisor by removing the overhead associated with software-only virtualization. AMD Virtualization Technology (AMD-V) The VM's view of physical memory is different from that of the memory controller. The hypervisor dynamically translates the VM's memory addresses sent to and received from the memory controller. As a result, each VM application remains unaware that memory is being virtualized. This translation process, known as shadow paging, increases memory latency and degrades performance. Rapid Virtualization Indexing Rapid Virtualization Indexing, an AMD-V innovation, reduces software virtualization overhead. It introduces hardware support for a second (nested) level of address translation. Rapid Virtualization Indexing, also known as nested paging, uses nested page tables that the hypervisor sets up to improve guest OS performance, when compared to shadow page tables (Figure 6). Figure 6. Hardware-based management using nested page tables improves the speed of the guest OS. AMD-Vi for I/O virtualization The chipset feature AMD-Vi, or IOMMU, speeds up I/O functions in VMs. Without IOMMU, a VM could program an I/O device in a way that corrupts another VM's memory. With IOMMU, I/O devices access only memory addresses that belong to a given VM. This lets a VM perform more like a physical machine, increasing performance, reliability, and security. 12

  • 1
  • 2
  • 3
  • 4
  • 5
  • 6
  • 7
  • 8
  • 9
  • 10
  • 11
  • 12
  • 13
  • 14
  • 15
  • 16

Power Cap Manager
Power Cap Manager allows you to set a fixed limit on a server's processor power consumption by
controlling the P-state of individual cores. The caveat is that a single voltage is supplied to all cores in
the processor even if you request different P-states. In this instance, the actual P-state will equal that of
the highest voltage required of the selected P-states. You can control this through HP Dynamic Power
Capping. You can access Dynamic Power Capping from iLO Advanced menus or the HP Insight
Control (ICE) management suite.
Virtualization
In a virtual machine (VM) environment, a software layer known as the hypervisor makes it appear that
each guest OS has full control over the server’s processor, memory, and I/O devices. Actually, each
guest OS has its own virtual processors, virtual drives, virtual NICs, and virtual storage controllers.
Several categories of software-layer abstraction exist. AMD-V and Intel VT technologies fall in the
hardware-assisted-virtualization category. These technologies assist the hypervisor by removing the
overhead associated with software-only virtualization.
AMD Virtualization Technology (AMD-V)
The VM’s view of physical memory is different from that of the memory controller. The hypervisor
dynamically translates the VM’s memory addresses sent to and received from the memory controller.
As a result, each VM application remains unaware that memory is being virtualized. This translation
process, known as shadow paging, increases memory latency and degrades performance.
Rapid Virtualization Indexing
Rapid Virtualization Indexing, an AMD-V innovation, reduces software virtualization overhead. It
introduces hardware support for a second (nested) level of address translation. Rapid Virtualization
Indexing, also known as nested paging, uses nested page tables that the hypervisor sets up to
improve guest OS performance, when compared to shadow page tables (Figure 6).
Figure 6.
Hardware-based management using nested page tables improves the speed of the guest OS.
AMD-Vi for I/O virtualization
The chipset feature AMD-Vi, or IOMMU, speeds up I/O functions in VMs. Without IOMMU, a VM
could program an I/O device in a way that corrupts another VM’s memory. With IOMMU, I/O
devices access only memory addresses that belong to a given VM. This lets a VM perform more like a
physical machine, increasing performance, reliability, and security.
12