NUMA_MAPS

Section: Linux Programmer's Manual (5)
Updated: 06 March 2006
Index Return to Main Contents
 

NAME

numa_maps - information about numa memory  

DESCRIPTION

The file /proc/<pid>/numa_maps contrains information about each memory area used by a given process allowing--among other information--the determination of which nodes were used for the pages.

numa_maps is only read by programs or by shell commands, and not written to. A read operation on the file will make the kernel scan the memory area of a process to determine how memory is used. One line is displayed for each memory area of the process. The numa_maps format is very simple. The first field shown on each line is the starting address that allows a correlation with output from /proc/<pid/maps. The second field shows the memory policy currently in effect for this particular memory area. The rest of the line contains series of informationi items about the pages in the memory area.  

Possible information items

N<node>=<pages>
The number of pages used on a node. The number of pages includes only pages mapped by this process. If this is a shared mapping then other processes may have mapped different pages in a memory area. Note also that page migration may have temporarily unmapped pages. These may only show up again after the process has attempted to reference these pages.
file=<filename>
The file backing the memory area. If COW (Copy-On-Write) pages were generated then the memory area may contain additional anonymous pages.
heap
Memory area is used for the heap.
stack
Memory area is used for the stack.
huge
Huge memory area. The page counts shown are huge pages and not regular sized pages.
anon=<pages>
The number of anonymous page in the range.
dirty=<pages>
Number of dirty pages
mapped=<pages>
Number of mapped pages for file backed files. Only shown if different from the count of anonymous and dirty pages.
mapmax=<count>
Maximum mapcount (number of processes mapping a single page) encountered during scan. This may be used as an indicator as to the sharing occuring in a given memory area.
swapcache=<count>
Number of pages that have an associated entry on the swap device.
active=<pages>
The number of pages on the active list (only shown if different from the number of pages in this area). If this item is shown then some inactive pages exist in that particular memory area that may be removed from memory by the swapper soon.
writeback=<pages>
Number of pages that are currently being written out to disk.
 

FILES

/proc/<pid>/numa_maps  

SEE ALSO

migratepages(8), numactl(8), cpuset(8), numactl(5)


 

Index

NAME
DESCRIPTION
Possible information items
FILES
SEE ALSO

linux.jgfs.net manual pages