numa

  1. D

    Why memory RSS is low but throughput is high

    I'm working on a NUMA-related benchmark and got an issue troubling me for a week. I use numactl to pin a metrics multiplication workload to use node0's CPU and node1's memory like this: However, I observe tons of memory access still goes to node0 (throughput is high), either using pcm or...
  2. Y

    vmstat numa on perf-top takes up to 13% in Kernel-4.14

    I'm started with NUMA on Kernel-4.14 arch-arm64 and see that vnstat.c::refresh_cpu_vm_stats is the most cpu-consumptive on the "perf top" and takes 13% in the IPERF loopback test described below. The NUMA is adding several atomic-counters in VMSTAT which obviously cause for this performance...
  3. K

    Any automated tool or balancer for Resource affinity domain setting for CPU in Linux?

    Anyone can share on CPU groupings techniques
Top