lkml.org 
[lkml]   [2019]   [Apr]   [3]   [last100]   RSS Feed
Views: [wrap][no wrap]   [headers]  [forward] 
 
Messages in this thread
/
Date
From
SubjectRe: [RFC][PATCH 03/16] sched: Wrap rq::lock access
> >>>Is the core wide lock primarily responsible for the regression? I ran
> >>>upto patch
> >>>12 which also has the core wide lock for tagged cgroups and also calls
> >>>newidle_balance() from pick_next_task(). I don't see any regression. 
> >>>Of
> >>>course
> >>>the core sched version of pick_next_task() may be doing more but
> >>>comparing with
> >>>the __pick_next_task() it doesn't look too horrible.
> >>On further testing and investigation, we also agree that spinlock
> >>contention
> >>is not the major cause for the regression, but we feel that it should be
> >>one
> >>of the major contributing factors to this performance loss.
> >>
> >>
> >I finally did some code bisection and found the following lines are
> >basically responsible for the regression. Commenting them out I don't see
> >the regressions. Can you confirm? I am yet to figure if this is needed for
> >the correctness of core scheduling and if so can we do this better?
> >
> >-------->8-------------
> >
> >diff --git a/kernel/sched/core.c b/kernel/sched/core.c
> >index fe3918c..3b3388a 100644
> >--- a/kernel/sched/core.c
> >+++ b/kernel/sched/core.c
> >@@ -3741,8 +3741,8 @@ pick_next_task(struct rq *rq, struct task_struct
> >*prev, struct rq_flags *rf)
> >                                 * If there weren't no cookies; we don't
> >need
> >                                 * to bother with the other siblings.
> >*/
> >-                               if (i == cpu && !rq->core->core_cookie)
> >-                                       goto next_class;
> >+                               //if (i == cpu && !rq->core->core_cookie)
> >+                                       //goto next_class;
> >
> >continue;
> >                        }
> AFAICT this condition is not needed for correctness as cookie matching will
> sill be enforced. Peter any thoughts? I get the following numbers with 1 DB
> and 2 DB instance.
>
> 1 DB instance
> users  baseline   %idle  core_sched %idle
> 16     1          84     -5.5% 84
> 24     1          76     -5% 76
> 32     1          69     -0.45% 69
>
> 2 DB instance
> users  baseline   %idle  core_sched %idle
> 16     1          66     -23.8% 69
> 24     1          54     -3.1% 57
> 32     1          42     -21.1%      48

We tried to comment those lines and it doesn’t seem to get rid of the
performance regression we are seeing.
Can you elaborate a bit more about the test you are performing, what kind of
resources it uses ?
Can you also try to replicate our test and see if you see the same problem ?

cgcreate -g cpu,cpuset:set1
cat /sys/devices/system/cpu/cpu{0,2,4,6}/topology/thread_siblings_list
0,36
2,38
4,40
6,42

echo "0,2,4,6,36,38,40,42" | sudo tee /sys/fs/cgroup/cpuset/set1/cpuset.cpus
echo 0 | sudo tee /sys/fs/cgroup/cpuset/set1/cpuset.mems

echo 1 | sudo tee /sys/fs/cgroup/cpu,cpuacct/set1/cpu.tag

sysbench --test=fileio prepare
cgexec -g cpu,cpuset:set1 sysbench --threads=4 --test=fileio \
--file-test-mode=seqwr run

The reason we create a cpuset is to narrow down the investigation to just 4
cores on a highly powerful machine. It might not be needed if testing on a
smaller machine.

Julien

\
 
 \ /
  Last update: 2019-04-03 22:17    [W:0.098 / U:28.552 seconds]
©2003-2020 Jasper Spaans|hosted at Digital Ocean and TransIP|Read the blog|Advertise on this site