> At Facebook we have a pretty heavily multi-threaded application that is > sensitive to latency. We have been pulling forward the old SD_WAKE_IDLE code > because it gives us a pretty significant performance gain (like 20%). It > turns > out this is because there are cases where the scheduler puts our task on a > busy > CPU when there are idle CPU's in the system. We verify this by reading the > cpu_delay_req_avg_us from the scheduler netlink stuff. With our crappy patch > we > get much lower numbers vs baseline. >
Was this application run under cpu cgroup. Because we were seeing bursty workloads exhibiting this behaviour esp when run under cpu cgroups. http://mid.gmane.org/53a11a89.5000...@linux.vnet.ibm.com -- Thansk and Regards Srikar -- To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to majord...@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html Please read the FAQ at http://www.tux.org/lkml/
