Technologies in HP ProLiant Gen8 c-Class server blades

You can read more about HP SmartMemory at http://h18004.www1.hp.com/products/servers/docs/4AA3-

9210ENW.pdf

HP Advanced Memory Error Detection technology

Uncorrectable memory errors can cause applications and operating systems to crash, so they are costly in terms of

downtime and repairs. The best way to prevent unnecessary DIMM replacements is to filter out superfluous errors and

identify critical errors that can lead to a shutdown. You can no longer rely on simple error event counts on systems

containing up to 14 trillion memory transistors. With HP Advanced Memory Error Detection Technology, we re-invented

a precision system that pinpoints errors that cause downtime.

HP Advanced Memory Error Detection Technology seeks out specific defects that either cause performance degradation

or significantly increase the probability of an uncorrectable (non-recoverable) memory condition. By improving the

prediction of non-recoverable memory events, this technology prevents unnecessary DIMM replacements and increases

server uptime.

ProLiant Gen8 server blades include these advanced memory protection features:

 Advanced ECC / SDDC - Continued memory operation in the event of a single memory device failure. Allows removal of

a single DRAM from the memory map if it exhibits a failure and recovers its data into a new device. Support for both x4

and x8 SDDC.

 Rank Sparring (OnLine Spare) - Dynamic fail-over to a spare DIMM rank or spare rank pair behind the same memory

controller. No OS involvement. You cannot enable this feature concurrently with memory mirroring. HP offers the rank

sparing rather than DIMM sparing as rank sparing uses less spare memory resulting in less overhead.

 Demand Scrubbing - Writes corrected data back to the memory once a correctable error is detected on a read

transaction.

 Patrol Scrubbing - Proactively searches the system memory repairing correctable errors. Patrol and Demand

scrubbing work together to prevent accumulation of correctable errors and reducing the likelihood of unplanned

downtime.

 Failed DIMM Isolation - Identifies a specific failing DIMM lockstep pair thereby enabling the user to replace only the

failed DIMM pair. Identifies a single DIMM for correctable errors and DIMM pair for uncorrectable errors.

 Memory Thermal Control - Used to prevent DIMMs from overheating. Achieved by slowing down the memory access

rate. The temperature is monitored by a DIMM sensor.

 DIMM Address/Control Bus Parity Protection - Methods to detect and protect command and address errors.

 Lockstep Memory Mode - Corrects a single x8 DRAM device failure on a DIMM. The DIMMs in each paired memory

channel must have identical HP part numbers.

For more information, read the technology brief “HP Advanced Memory Error Detection Technology.” Find it at

http://h20000.www2.hp.com/bc/docs/support/SupportManual/c02878598/c02878598.pdf.

I/O Interconnect and convergence technologies

HP ProLiant Gen8 server blades support PCI Express (PCIe), serial attached SCSI (SAS), serial ATA (SATA) I/O technologies,

1 Gb and 10 Gb Ethernet, 10 Gb FlexFabric, 10 Gb Flex-10, 8 Gb Fibre Channel, and 4X DDR (20 Gb) InfiniBand. Beyond the

I/O technology and performance characteristics described in this section, I/O convergence is an increasingly important

factor in current and future data center infrastructure. HP Converged Infrastructure and Virtual Connect technology

address the need for a common, virtualized network fabric. HP ProLiant Gen8 server blades also support I/O

convergence with HP optional mezzanine cards, and FlexibleLOMs.

PCI Express technology

The PCI Express (PCIe) serial interface provides point-to-point connections between the chipset I/O controller hub and

I/O devices. Each PCIe serial link consists of one or more dual-simplex lanes. Each lane contains a send pair and a receive