{ "mode": "man", "parameter": "numa", "section": "7", "url": "https://www.chedong.com/phpMan.php/man/numa/7/json", "generated": "2026-10-09T17:00:48Z", "sections": { "NAME": { "content": "numa - overview of Non-Uniform Memory Architecture\n", "subsections": [] }, "DESCRIPTION": { "content": "Non-Uniform Memory Access (NUMA) refers to multiprocessor systems whose memory is divided\ninto multiple memory nodes. The access time of a memory node depends on the relative loca‐\ntions of the accessing CPU and the accessed node. (This contrasts with a symmetric multi‐\nprocessor system, where the access time for all of the memory is the same for all CPUs.)\nNormally, each CPU on a NUMA system has a local memory node whose contents can be accessed\nfaster than the memory in the node local to another CPU or the memory on a bus shared by all\nCPUs.\n", "subsections": [ { "name": "NUMA system calls", "content": "The Linux kernel implements the following NUMA-related system calls: getmempolicy(2),\nmbind(2), migratepages(2), movepages(2), and setmempolicy(2). However, applications\nshould normally use the interface provided by libnuma; see \"Library Support\" below.\n\n/proc/pid/numamaps (since Linux 2.6.14)\nThis file displays information about a process's NUMA memory policy and allocation.\n\nEach line contains information about a memory range used by the process, displaying—among\nother information—the effective memory policy for that memory range and on which nodes the\npages have been allocated.\n\nnumamaps is a read-only file. When /proc/pid/numamaps is read, the kernel will scan the\nvirtual address space of the process and report how memory is used. One line is displayed\nfor each unique memory range of the process.\n\nThe first field of each line shows the starting address of the memory range. This field al‐\nlows a correlation with the contents of the /proc/pid/maps file, which contains the end ad‐\ndress of the range and other information, such as the access permissions and sharing.\n\nThe second field shows the memory policy currently in effect for the memory range. Note that\nthe effective policy is not necessarily the policy installed by the process for that memory\nrange. Specifically, if the process installed a \"default\" policy for that range, the effec‐\ntive policy for that range will be the process policy, which may or may not be \"default\".\n\nThe rest of the line contains information about the pages allocated in the memory range, as\nfollows:\n\nN=\nThe number of pages allocated on . includes only pages currently\nmapped by the process. Page migration and memory reclaim may have temporarily un‐\nmapped pages associated with this memory range. These pages may show up again only\nafter the process has attempted to reference them. If the memory range represents a\nshared memory area or file mapping, other processes may currently have additional\npages mapped in a corresponding memory range.\n\nfile=\nThe file backing the memory range. If the file is mapped as private, write accesses\nmay have generated COW (Copy-On-Write) pages in this memory range. These pages are\ndisplayed as anonymous pages.\n\nheap Memory range is used for the heap.\n\nstack Memory range is used for the stack.\n\nhuge Huge memory range. The page counts shown are huge pages and not regular sized pages.\n\nanon=\nThe number of anonymous page in the range.\n\ndirty=\nNumber of dirty pages.\n\nmapped=\nTotal number of mapped pages, if different from dirty and anon pages.\n\nmapmax=\nMaximum mapcount (number of processes mapping a single page) encountered during the\nscan. This may be used as an indicator of the degree of sharing occurring in a given\nmemory range.\n\nswapcache=\nNumber of pages that have an associated entry on a swap device.\n\nactive=\nThe number of pages on the active list. This field is shown only if different from\nthe number of pages in this range. This means that some inactive pages exist in the\nmemory range that may be removed from memory by the swapper soon.\n\nwriteback=\nNumber of pages that are currently being written out to disk.\n" } ] }, "STANDARDS": { "content": "None.\n", "subsections": [] }, "NOTES": { "content": "The Linux NUMA system calls and /proc interface are available only if the kernel was config‐\nured and built with the CONFIGNUMA option.\n", "subsections": [ { "name": "Library support", "content": "Link with -lnuma to get the system call definitions. libnuma and the required \nheader are available in the numactl package.\n\nHowever, applications should not use these system calls directly. Instead, the higher level\ninterface provided by the numa(3) functions in the numactl package is recommended. The nu‐\nmactl package is available at ftp://oss.sgi.com/www/projects/libnuma/download/. The package\nis also included in some Linux distributions. Some distributions include the development li‐\nbrary and header in the separate numactl-devel package.\n" } ] }, "SEE ALSO": { "content": "getmempolicy(2), mbind(2), movepages(2), setmempolicy(2), numa(3), cpuset(7), numactl(8)\n\nLinux man-pages 6.7 2023-10-31 numa(7)", "subsections": [] } }, "summary": "numa - overview of Non-Uniform Memory Architecture", "flags": [], "examples": [], "see_also": [ { "name": "getmempolicy", "section": "2", "url": "https://www.chedong.com/phpMan.php/man/getmempolicy/2/json" }, { "name": "mbind", "section": "2", "url": "https://www.chedong.com/phpMan.php/man/mbind/2/json" }, { "name": "movepages", "section": "2", "url": "https://www.chedong.com/phpMan.php/man/movepages/2/json" }, { "name": "setmempolicy", "section": "2", "url": "https://www.chedong.com/phpMan.php/man/setmempolicy/2/json" }, { "name": "cpuset", "section": "7", "url": "https://www.chedong.com/phpMan.php/man/cpuset/7/json" }, { "name": "numactl", "section": "8", "url": "https://www.chedong.com/phpMan.php/man/numactl/8/json" } ] }