Locality-Optimized Resource Alignment for Superdome 2

7
Glossary
Term Definition
Socket
A socket on Superdome 2 represents the field-replaceable processing unit on a blade.
Each socket is associated with a set of memory DIMMs and local I/O resources.
Processor
core
When a socket contains multiple independent processing entities, each of them is
called a processor core.
Core
Same as processor core. The keyword ‘cpu’ in the vPars commands refers to a
core.
CPU Acronym for Central Processing Unit. The term ‘processor core’ is preferred.
Cache line
The quantum of data transferred between a processor core and memory. A typical
size is 128 bytes.
Complex An entire Superdome 2 system taken as a whole is called a complex.
Blade
The physical building block of a complex and the unit of assignment for forming
nPartitions. A blade contains processor sockets, memory, and I/O components.
Crossbar
A component of the interconnect fabric that allows the blades in a complex to
communicate with each other.
Locality
A set of processors, memory, and I/O system bus adapters identified by the
operating system for resource alignment purposes.
Locality
domain
A locality that contains at least one processor core for scheduling processes.
SMP
Acronym for Symmetric Multiprocessor. A model in which all of the cores in a
system are equivalent to and interchangeable with each other.
NUMA
Acronym for Non--Uniform Memory Architecture. A hardware platform in which
system memory is separated into localities based on memory access latency times
for processors.
ILM
Acronym for Interleaved Memory. A technique in which successive cache lines of
memory are drawn from different localities.
Technical details
The appendices that follow contain the technical details to supplement the general information
presented in the opening sections of this paper. The appendix
Configuring nPartitions for LORA
explains the steps needed for deployment of LORA directly on an nPartition. The appendix
Configuring vPars for LORA describes the steps needed when vPars is used in a LORA nPartition.
Recommendations for fine-tuning workloads are given in the appendix
Advanced tuning.