| Subject | Re: [PATCH 18/19] sched/numa: Reset scan rate whenever task moves across nodes | From | Rik van Riel <> | Date | Mon, 04 Jun 2018 16:08:12 -0400 |
| |
On Mon, 2018-06-04 at 15:30 +0530, Srikar Dronamraju wrote: > Currently task scan rate is reset when numa balancer migrates the > task > to a different node. If numa balancer initiates a swap, reset is only > applicable to the task that initiates the swap. Similarly no scan > rate > reset is done if the task is migrated across nodes by traditional > load > balancer. > > Instead move the scan reset to the migrate_task_rq. This ensures the > task moved out of its preferred node, either gets back to its > preferred > node quickly or finds a new preferred node. Doing so, would be fair > to > all tasks migrating across nodes.
How does this impact performance of benchmarks closer to real world workloads?
Not just the test cases measuring NUMA convergence.
-- All Rights Reversed.[unhandled content-type:application/pgp-signature] |