White Paper - Intel Intelligent Power Node Manager

ManualsBrandsIntel ManualsOtherIntel Server Board S5520UR

White Paper Intel® Intelligent Power

Node Manager

Power consumption is the biggest area of concern in the server domain. Between 2000 and 2005,

server electricity use more than doubled, growing at a rate of 14 percent each year. The power

consumption cost of servers is rapidly growing and management costs have gone higher forcing IDC

to identify power management as the critical need of the hour.

By 2007, Gartner had predicted that power and cooling spending will exceed server spending. By

2008, 50% of all data centers will have insufficient power and cooling capacity to meet the demand

of high density equipment. Power and cooling costs will increase to more than one-third of the total IT

budget

Enhancements in processor performance and density of servers have been the prime contributors to

an increase in power consumption. The Power and cooling spend on server is supposed to increase to

$30 billion approx in 2010 from $10 billion in 1996 as indicated in the following Table 1.

Summary of content (7 pages)

PAGE 1
White Paper Intel® Intelligent Power Node Manager Power consumption is the biggest area of concern in the server domain. Between 2000 and 2005, server electricity use more than doubled, growing at a rate of 14 percent each year. The power consumption cost of servers is rapidly growing and management costs have gone higher forcing IDC to identify power management as the critical need of the hour. By 2007, Gartner had predicted that power and cooling spending will exceed server spending.
PAGE 2
While the components of a server system - processor, operating system, and so on - employ different techniques to check power consumption, areas that could potentially help to reduce power consumption are largely untapped. 1. Intel® processors and power management Intel® processors support many features for power management. Prominent among them are demand based switching using the Intel® SpeedStep® Technology and clock modulation.
PAGE 3
5500 series chipset systems and their hardware support needed for Node Manager 3.
PAGE 4
Intel® Server Chassis SC5600BASE Chassis-specific Server Management Features • • SC5600BRP • • • • • Non- redundant power: One power supply module which is not manageable Non-redundant cooling: Five fixed chassis fans which are not hot swappable Chassis Intrusion Redundant power: Two PMBus enabled power supply modules which are manageable Power Distribution Board (PDB) Non-redundant cooling: Five fixed chassis fans which are not hot swappable Chassis Intrusion Chassis-dependent Sensors Managed by Ma
PAGE 5
4. PMBus and Power Supply sensors to enable NM support The Power Management Bus (“PMBus”) is an open standard protocol that defines a means of communicating with power conversion and other devices. For more information, please see the System Management Interface Forum Web site: www.powerSIG.org. To support NM capabilities the server should have the power supply device and power supply unit compliant with PMBus specification. 5. Node Manager FAQs for finer control of the server system 1.
PAGE 6
Possible Setting Friendly Name: Off Possible Setting Index: 001 Possible Setting Friendly Name: On Current AC Power Setting Index: 0x00000000 Current DC Power Setting Index: 0x00000000 c) Copy the globally unique identifier (GUID) for the “Allow Throttle States” processor power policy. d) Set the “Allow Throttle States” policy value to 1 by using the following command: powercfg -setacvalueindex scheme_current sub_processor 3b04d4fd-1cc7-4f23ab1c-d1337819c4bb 1 3.
PAGE 7
6.