[<prev] [next>] [thread-next>] [day] [month] [year] [list]
Message-Id: <1484745662-15928-1-git-send-email-kan.liang@intel.com>
Date: Wed, 18 Jan 2017 08:21:00 -0500
From: kan.liang@...el.com
To: linux-kernel@...r.kernel.org, peterz@...radead.org,
mingo@...hat.com
Cc: alexander.shishkin@...ux.intel.com, eranian@...gle.com,
ak@...ux.intel.com, Kan Liang <kan.liang@...el.com>
Subject: [PATCH 0/2] patches to reduce perf overhead
From: Kan Liang <kan.liang@...el.com>
perf brings additional overhead when monitoring the task which
frequently generates child task. Two of the related issues are fixed
in the patch series.
Here is some data from the overhead test on Broadwell server
perf record -e $TEST_EVENTS -- ./loop.sh 50000
loop.sh
start=$(date +%s%N)
i=0
while [ "$i" -le "$1" ]
do
date > /dev/null
i=`expr $i + 1`
done
end=$(date +%s%N)
elapsed=`expr $end - $start`
Event# Original elapsed time Elapsed time with two patches delta
1 196,573,192,397 187,781,884,240 -4.47%
2 257,567,753,013 238,954,990,569 -7.23%
4 398,730,726,971 354,101,552,792 -11.19%
8 824,983,761,120 673,623,574,792 -18.35%
16 1,883,411,923,498 1,456,452,391,357 -22.67%
Kan Liang (2):
perf,core: try parent pmu first when init child event
perf,core: use parent avg sample period as child initial period
include/linux/perf_event.h | 3 +++
kernel/events/core.c | 28 ++++++++++++++++++++++++++--
2 files changed, 29 insertions(+), 2 deletions(-)
--
2.4.3
Powered by blists - more mailing lists