mirror of
https://github.com/AuxXxilium/linux_dsm_epyc7002.git
synced 2024-12-03 01:36:44 +07:00
317f394160
Now that we've removed the rq->lock requirement from the first part of ttwu() and can compute placement without holding any rq->lock, ensure we execute the second half of ttwu() on the actual cpu we want the task to run on. This avoids having to take rq->lock and doing the task enqueue remotely, saving lots on cacheline transfers. As measured using: http://oss.oracle.com/~mason/sembench.c $ for i in /sys/devices/system/cpu/cpu*/cpufreq/scaling_governor ; do echo performance > $i; done $ echo 4096 32000 64 128 > /proc/sys/kernel/sem $ ./sembench -t 2048 -w 1900 -o 0 unpatched: run time 30 seconds 647278 worker burns per second patched: run time 30 seconds 816715 worker burns per second Reviewed-by: Frank Rowand <frank.rowand@am.sony.com> Cc: Mike Galbraith <efault@gmx.de> Cc: Nick Piggin <npiggin@kernel.dk> Cc: Linus Torvalds <torvalds@linux-foundation.org> Cc: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Ingo Molnar <mingo@elte.hu> Signed-off-by: Peter Zijlstra <a.p.zijlstra@chello.nl> Link: http://lkml.kernel.org/r/20110405152729.515897185@chello.nl
73 lines
1.8 KiB
C
73 lines
1.8 KiB
C
/*
|
|
* Only give sleepers 50% of their service deficit. This allows
|
|
* them to run sooner, but does not allow tons of sleepers to
|
|
* rip the spread apart.
|
|
*/
|
|
SCHED_FEAT(GENTLE_FAIR_SLEEPERS, 1)
|
|
|
|
/*
|
|
* Place new tasks ahead so that they do not starve already running
|
|
* tasks
|
|
*/
|
|
SCHED_FEAT(START_DEBIT, 1)
|
|
|
|
/*
|
|
* Should wakeups try to preempt running tasks.
|
|
*/
|
|
SCHED_FEAT(WAKEUP_PREEMPT, 1)
|
|
|
|
/*
|
|
* Based on load and program behaviour, see if it makes sense to place
|
|
* a newly woken task on the same cpu as the task that woke it --
|
|
* improve cache locality. Typically used with SYNC wakeups as
|
|
* generated by pipes and the like, see also SYNC_WAKEUPS.
|
|
*/
|
|
SCHED_FEAT(AFFINE_WAKEUPS, 1)
|
|
|
|
/*
|
|
* Prefer to schedule the task we woke last (assuming it failed
|
|
* wakeup-preemption), since its likely going to consume data we
|
|
* touched, increases cache locality.
|
|
*/
|
|
SCHED_FEAT(NEXT_BUDDY, 0)
|
|
|
|
/*
|
|
* Prefer to schedule the task that ran last (when we did
|
|
* wake-preempt) as that likely will touch the same data, increases
|
|
* cache locality.
|
|
*/
|
|
SCHED_FEAT(LAST_BUDDY, 1)
|
|
|
|
/*
|
|
* Consider buddies to be cache hot, decreases the likelyness of a
|
|
* cache buddy being migrated away, increases cache locality.
|
|
*/
|
|
SCHED_FEAT(CACHE_HOT_BUDDY, 1)
|
|
|
|
/*
|
|
* Use arch dependent cpu power functions
|
|
*/
|
|
SCHED_FEAT(ARCH_POWER, 0)
|
|
|
|
SCHED_FEAT(HRTICK, 0)
|
|
SCHED_FEAT(DOUBLE_TICK, 0)
|
|
SCHED_FEAT(LB_BIAS, 1)
|
|
|
|
/*
|
|
* Spin-wait on mutex acquisition when the mutex owner is running on
|
|
* another cpu -- assumes that when the owner is running, it will soon
|
|
* release the lock. Decreases scheduling overhead.
|
|
*/
|
|
SCHED_FEAT(OWNER_SPIN, 1)
|
|
|
|
/*
|
|
* Decrement CPU power based on irq activity
|
|
*/
|
|
SCHED_FEAT(NONIRQ_POWER, 1)
|
|
|
|
/*
|
|
* Queue remote wakeups on the target CPU and process them
|
|
* using the scheduler IPI. Reduces rq->lock contention/bounces.
|
|
*/
|
|
SCHED_FEAT(TTWU_QUEUE, 1)
|