beautypg.com

Memory cartridge population guidelines – HP ProLiant DL580 G7 Server User Manual

Page 46

background image

Hardware options installation 46

o

If the configuration is two quad-rank 8-GB DIMMs and six dual-rank 8-GB DIMMs in a single

cartridge configuration, the quad-rank DIMMs can be located in either of the corresponding white
DIMM pairs (1A/8A or 3B/6B) as long as they are installed at the end point of the DDR3 channel.

AMP modes Advanced ECC, Online Spare, and Mirrored Memory have further requirements beyond
the ones listed here. For additional memory configuration requirements, see the corresponding AMP

sections:

o

Advanced ECC memory population guidelines (on page

50

)

o

Online Spare memory population guidelines (on page

50

)

o

Mirrored Memory population guidelines (on page

51

)

Memory cartridge population guidelines

This server contains eight memory cartridge slots in the processor memory drawer.
Observe the following guidelines:

Do not mix standard memory cartridges and (E7) memory cartridges.

Do not use (E7) memory cartridges in servers that shipped with standard memory cartridges.

Populate memory cartridge slots for a processor only if the corresponding processor is installed:

o

Processor 1 connects to the memory cartridge in slots 1 and 2.

o

Processor 2 connects to the memory cartridge in slots 3 and 4.

o

Processor 3 connects to the memory cartridge in slots 5 and 6.

o

Processor 4 connects to the memory cartridge in slots 7 and 8.

Memory is only accessible to the system if the associated processor is installed. Do not install memory

cartridges in cartridge slots without the corresponding processor installed.

The minimum memory cartridge population is one per processor. For best performance, two memory

cartridges per processor are recommended.

Due to higher memory access latencies, HP does not recommend installing a processor without its

corresponding memory cartridges.

To maximize performance in multi-processor configurations, distribute the total memory capacity evenly

across all processors.