DIMM sparing provides a mechanism on SPARC T8 servers to unconfigure a single failed DIMM with minimal performance loss. This feature allows deferred maintenance for DIMMs that are faulted. DIMM sparing is supported only on systems whose memory slots are fully-populated with DIMMs.
If a DIMM is diagnosed to be faulty during boot time or run time, the memory dynamically switches from 16-way to 15-way interleave by remapping all of the physical addresses to the remaining 15 DIMMs. To enable this remapping, the platform firmware must initially reserve space for the contents of one DIMM. Consequently, only 15 DIMMs worth of physical address space is made available to the system, even when 16 DIMMs are functioning.
Keep in mind that, due to the requirements of DIMM sparing, the amount of available memory reported might be less than the amount you had estimated based on the quantity and capacity of the server's DIMMs.
When DIMM sparing is enabled, if the server must unconfigure a DIMM, the associated fault is treated as a nonserviceable fault so a service notification is not issued. If one DIMM in a group of 16 is unconfigured, you do not need to replace that DIMM until another DIMM in that group is unconfigured.
DIMM sparing is enabled by default on CPU nodes that are fully populated with DIMMs.
The DIMM sparing feature is enabled when the optional memory risers are installed and fully occupied.
See Memory Riser and DIMM Configuration in SPARC T8-1 Server Service Manual.
The DIMM sparing feature is enabled in fully-populated configurations.
See Memory Riser and DIMM Configuration in SPARC T8-2 Server Service Manual
The DIMM sparing feature is enabled in fully-populated configurations.
See Understanding DIMM Configurations in SPARC T8-4 Server Service Manual
My Oracle Support article SPARC T7 / M7 Servers : DIMM sparing FAQ (Doc ID 2037793.1)