I just ran powertop on a server which has a lot of small HTTP requests coming in. I'm wondering if there's anything here that can be optimized (other than powertop's suggestion of optimizing disk access... the number of interrupts there is so low in comparison that it doesn't seem worth the time). Top causes for wakeups: 30.9% (10350.7) <interrupt> : eth0 26.9% (9009.8) <kernel core> : hrtimer_start_range_ns (tick_sched_timer) 21.5% (7194.3) <kernel IPI> : Rescheduling interrupts 14.4% (4813.7) <interrupt> : eth1 1.3% (428.0) python : hrtimer_start_range_ns (hrtimer_wakeup) 1.2% (387.2) python : mod_timer (tcp_delack_timer) 1.0% (349.7) mysqld : mod_timer (tcp_delack_timer) 0.9% (298.3) python : mod_timer (tcp_write_timer) 0.8% (276.2) mysqld : mod_timer (tcp_write_timer) 0.5% (158.7) <kernel IPI> : TLB shootdowns "Rescheduling interrupts" seems interesting. Can anything be done to get rid of those? (e.g. binding a process to a specific CPU/core) -- fedora-list mailing list fedora-list@xxxxxxxxxx To unsubscribe: https://www.redhat.com/mailman/listinfo/fedora-list Guidelines: http://fedoraproject.org/wiki/Communicate/MailingListGuidelines
