Non-Uniform Memory Access (NUMA)

GPTKB entity

Predicate	Object
gptkbp:instanceOf	gptkb:architecture
gptkbp:abbreviation	gptkb:NUMA
gptkbp:characteristic	memory access time depends on memory location relative to processor
gptkbp:contrastsWith	gptkb:Uniform_Memory_Access_(UMA)
gptkbp:documentedIn	gptkb:ACM_Digital_Library gptkb:IEEE_Computer_Society_publications
gptkbp:enables	scalability in multi-CPU systems
gptkbp:hasComponent	node interconnect local memory remote memory
gptkbp:hasFeature	gptkb:ACPI_SLIT gptkb:ACPI_SRAT
https://www.w3.org/2000/01/rdf-schema#label	Non-Uniform Memory Access (NUMA)
gptkbp:improves	memory bandwidth
gptkbp:introducedIn	gptkb:Honeywell_Multics gptkb:Sequent_Balance_8000
gptkbp:limitation	performance penalty for remote memory access
gptkbp:reduces	memory latency for local memory access
gptkbp:relatedTo	gptkb:cache-coherent_NUMA_(ccNUMA) gptkb:symmetric_multiprocessing_(SMP) distributed shared memory
gptkbp:requires	NUMA-aware operating system
gptkbp:supportedBy	gptkb:Windows gptkb:FreeBSD gptkb:Linux
gptkbp:usedBy	high-performance computing database servers virtualization platforms
gptkbp:usedIn	multiprocessor systems
gptkbp:bfsParent	gptkb:NUMAlink_interconnect gptkb:Symmetric_Multiprocessing_(SMP)
gptkbp:bfsLayer	7