Lines Matching defs:local
35 * one or more CPUs and some local memory. The CPUs in each node can access
37 * local memory. Typically, a system with only one node has Uniform Memory
50 * local memory from each node as needed. The common lgroup framework uses the
65 * CPUs and memory are local to a given proximity domain (NUMA node). The SLIT
446 * Forward declarations of local routines
1223 * close to each other and local latencies are made the same. Lastly, the
1225 * local < remote, latency(a, b) == latency(b, a), etc.).
1228 * configuration with the local CPUs and memory given by the SRAT or PCI config
1314 * - Fix up latencies such that local latencies are same,
1319 * - Fallback to just optimizing for local and remote if
1393 * local APIC ID in SRAT table.
1497 * Determine which CPUs and memory are local to each other and number
1677 * any nodes that are within some tolerance to be same, and make local
1854 * - Find min and max local latencies
1855 * - Make all local latencies be minimum
1872 int local;
1877 local = lat_stats->latencies[i][i];
1878 if (local == 0)
1883 * tolerance of minimum local latency and how much
1886 if (local - min > min >> lgrp_plat_probe_lt_shift)
1889 probe_stats->probe_errors[i][i] += local - min;
1892 * Make local latencies be minimum
2040 * (ie. local and remote only) if they haven't been set yet
2801 * and memory are local to each other in the same NUMA node and return number
3273 "MPO only optimizing for local and remote\n");