Configuring and using DDR3 memory with HP ProLiant Gen8 Servers

15
• Use the same HP SmartMemory part number in each memory channel.
Optimizing memory configurations
By taking advantage of the different DIMM types, sizes and speeds available for HP ProLiant Gen8 servers, you can
optimize server memory configuration to meet different application or datacenter requirements.
Optimizing for capacity
You can maximize memory capacity on ProLiant Gen8 servers using the new 32 GB LRDIMMs. With LRDIMMs you can
install up to three quad-ranked DIMMs in a memory channel, which was not possible with earlier ProLiant G6 or G7
ProLiant servers. On 24 slot servers, you can configure the system with up to 768 GB of total memory.
Table 10 shows the maximum memory capacities for ProLiant Gen8 servers using each of the 3 DIMM types.
Table 10. Maximum memory capacities for 2P ProLiant Gen8 servers using different DIMM types
Number of DIMM Slots DIMM Type Maximum Capacity Configuration
24 UDIMM 128 GB 16 x 8GB 2R
RDIMM 384 GB 24 x 16GB 2R
LRDIMM 768 GB 24 x 32GB 4R
16 UDIMM 128 GB 16 x 8GB 2R
RDIMM 256 GB 16 x 16GB 2R
LRDIMM 512 GB 16 x 32GB 4R
12 UDIMM 96 GB 12 x 8GB 2R
RDIMM 192 GB 12 x 16GB 2R
LRDIMM 384 GB 12 x 32GB 4R
Optimizing for performance
The two primary measurements of memory subsystem performance are throughput and latency. Latency is a measure
of the time it takes for the memory subsystem to begin to deliver data to the processor core after the processor makes a
request. Throughput measures the total amount of data that the memory subsystem can transfer to the system
processor(s) during a given period.
Factors influencing latency
Unloaded and loaded latencies are a measure of the efficiency of the memory sub-section in a server. Memory latency in
servers is usually measured from the time of a read request in the core of a processor until the data is supplied to that
core. This is also called load-to-use. Unloaded latency measures the latency when the system is idle and represents the
lowest latency that the system can achieve for memory requests for a given processor/memory combination. Loaded
latency is the latency when the memory subsystem is saturated with memory requests. Loaded latency will always be
greater than unloaded latency.
There are a number of factors that influence memory latency in a system.
• DIMM Speed. Faster DIMM speeds deliver lower latency, particularly loaded latency. Under loaded conditions, the
primary contributor to latency is the time memory requests spend in a queue waiting to be executed. The faster the
DIMM speed, the more quickly the memory controller can process the queued commands. For example, Memory
running at 1600 MT/s has about 20% lower loaded latency than memory running at 1333 MT/s.
• Ranks. For the same memory speed and DIMM type, more ranks will result in lower loaded latency. More ranks give
the memory controller a greater capability to parallelize the processing of memory requests. This results in shorter
request queues and therefore lower latency.