[PATCH v2 0/6] numabalancing patches

From: Srikar Dronamraju
Date: Fri Sep 21 2018 - 13:49:38 EST


This patchset provides left out patches from the previous series.
http://lore.kernel.org/lkml/1529514181-9842-1-git-send-email-srikar@xxxxxxxxxxxxxxxxxx

Changes from previous posting:
Rebased the patchset to 4.19-rc4.
Handled comments from Ingo Molnar,Peter Peter Zijlstra and Mel Gorman.
Reorder to keep the most discussed patches at the end.

Some of the Ack/Reviewed-by carried from
http://lore.kernel.org/lkml/1528106428-19992-1-git-send-email-srikar@xxxxxxxxxxxxxxxxxx

Please refer to previous postings
http://lore.kernel.org/lkml/1533276841-16341-1-git-send-email-srikar@xxxxxxxxxxxxxxxxxx
http://lore.kernel.org/lkml/1529514181-9842-1-git-send-email-srikar@xxxxxxxxxxxxxxxxxx
http://lore.kernel.org/lkml/1528106428-19992-1-git-send-email-srikar@xxxxxxxxxxxxxxxxxx

Specjbb2005 results (8 warehouses)
Higher bops are better

2 Socket - 2 Node Haswell - X86
JVMS Prev Current %Change
4 200194 205910 2.85523
1 311331 318491 2.2998

2 Socket - 4 Node Power8 - PowerNV
JVMS Prev Current %Change
1 197654 226796 14.7439

4 Socket - 4 Node Power7 - PowerVM
JVMS Prev Current %Change
8 52227.1 60370 15.5913
1 102529 108100 5.43358

Mel Gorman (1):
sched/numa: Limit the conditions where scan period is reset

Srikar Dronamraju (5):
sched/numa: Stop multiple tasks from moving to the CPU at the same
time
sched/numa: Pass destination CPU as a parameter to migrate_task_rq
sched/numa: Reset scan rate whenever task moves across nodes
mm/migrate: Use trylock while resetting rate limit
sched/numa: Avoid task migration for small NUMA improvement

kernel/sched/core.c | 2 +-
kernel/sched/deadline.c | 2 +-
kernel/sched/fair.c | 92 ++++++++++++++++++++++++++++++++++++++++++-------
kernel/sched/sched.h | 3 +-
mm/migrate.c | 16 ++++++---
5 files changed, 96 insertions(+), 19 deletions(-)

--
2.7.4