[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Date: Tue, 8 Aug 2017 09:19:23 +0800
From: kernel test robot <xiaolong.ye@...el.com>
To: Nadav Amit <namit@...are.com>
Cc: linux-mm@...ck.org, nadav.amit@...il.com,
linux-kernel@...r.kernel.org, akpm@...ux-foundation.org,
Minchan Kim <minchan@...nel.org>,
Ingo Molnar <mingo@...hat.com>,
Russell King <linux@...linux.org.uk>,
Tony Luck <tony.luck@...el.com>,
Martin Schwidefsky <schwidefsky@...ibm.com>,
"David S. Miller" <davem@...emloft.net>,
Heiko Carstens <heiko.carstens@...ibm.com>,
Yoshinori Sato <ysato@...rs.sourceforge.jp>,
Jeff Dike <jdike@...toit.com>, linux-arch@...r.kernel.org,
Nadav Amit <namit@...are.com>, lkp@...org
Subject: [lkp-robot] [mm] 7674270022: will-it-scale.per_process_ops -19.3%
regression
Greeting,
FYI, we noticed a -19.3% regression of will-it-scale.per_process_ops due to commit:
commit: 76742700225cad9df49f05399381ac3f1ec3dc60 ("mm: fix MADV_[FREE|DONTNEED] TLB flush miss problem")
url: https://github.com/0day-ci/linux/commits/Nadav-Amit/mm-migrate-prevent-racy-access-to-tlb_flush_pending/20170802-205715
in testcase: will-it-scale
on test machine: 88 threads Intel(R) Xeon(R) CPU E5-2699 v4 @ 2.20GHz with 64G memory
with following parameters:
nr_task: 16
mode: process
test: brk1
cpufreq_governor: performance
test-description: Will It Scale takes a testcase and runs it from 1 through to n parallel copies to see if the testcase will scale. It builds both a process and threads based test in order to see any differences between the two.
test-url: https://github.com/antonblanchard/will-it-scale
Details are as below:
-------------------------------------------------------------------------------------------------->
To reproduce:
git clone https://github.com/01org/lkp-tests.git
cd lkp-tests
bin/lkp install job.yaml # job file is attached in this email
bin/lkp run job.yaml
testcase/path_params/tbox_group/run: will-it-scale/16-process-brk1-performance/lkp-bdw-ep3d
378005bdbac0a2ec 76742700225cad9df49f053993
---------------- --------------------------
%stddev change %stddev
\ | \
3405093 -19% 2747088 will-it-scale.per_process_ops
2.396e+08 8283% 2.008e+10 perf-stat.iTLB-loads
0.03 26% 0.04 ± 10% perf-stat.branch-miss-rate%
0.00 25% 0.00 ± 11% perf-stat.dTLB-store-miss-rate%
0.56 18% 0.66 perf-stat.cpi
2.859e+13 -13% 2.479e+13 perf-stat.instructions
5.906e+12 -14% 5.087e+12 perf-stat.branch-instructions
4.831e+12 -14% 4.154e+12 perf-stat.dTLB-stores
1.575e+10 ± 6% -15% 1.345e+10 perf-stat.iTLB-load-misses
1.77 -15% 1.51 perf-stat.ipc
8.476e+12 ± 5% -16% 7.079e+12 perf-stat.dTLB-loads
98.49 -59% 40.10 perf-stat.iTLB-load-miss-rate%
will-it-scale.per_process_ops
3.5e+06 ++----------------------------------------------------------------+
*..*.*..*.*.. .*.*..*.*..*. .* *..*.*..*.*..*.*..*. .*. .*
3.4e+06 ++ *.*. *. : : *. *. |
3.3e+06 ++ : : |
| : : |
3.2e+06 ++ :: |
3.1e+06 ++ * |
| |
3e+06 ++ |
2.9e+06 ++ |
| |
2.8e+06 ++ |
2.7e+06 O+ O O O O O O O O O O O O O O O O O O O O O O O O O
| O |
2.6e+06 ++----------------------------------------------------------------+
perf-stat.branch-instructions
6.8e+12 ++----------------------------------------------------------------+
6.6e+12 ++ * |
| :: |
6.4e+12 ++ : : |
6.2e+12 ++ : : |
| .*.*..*.*..* * |
6e+12 *+.*.*..*.*..*.*. + *..*.*..*.*..*.*..*.*..*.*..*
5.8e+12 ++ + + |
5.6e+12 ++ * |
| |
5.4e+12 ++ |
5.2e+12 O+ O O O O O O O O O O O O O |
| O O O O O O O O O O O
5e+12 ++ O O |
4.8e+12 ++----------------------------------------------------------------+
perf-stat.iTLB-loads
3.5e+10 ++----------------------------------------------------------------+
O O O O O |
3e+10 ++ |
| O O |
2.5e+10 ++ O |
| |
2e+10 ++ O O O O O O O O O O O O O O O O O O O
| |
1.5e+10 ++ |
| |
1e+10 ++ |
| |
5e+09 ++ |
| |
0 *+-*-*--*-*--*-*--*-*--*-*--*-*--*--*-*--*-*--*-*--*-*--*-*--*-*--*
perf-stat.iTLB-load-miss-rate_
100 *+-*-*--*--*-*-----*--*-*--*--*-*--*--*-*--*--*-*--*--*--*-*--*--*-*--*
| *. |
90 ++ |
80 ++ |
| |
70 ++ |
| |
60 ++ |
| |
50 ++ |
40 ++ O O O O O O
| O O O O O O O O O O O O O |
30 ++ O O O |
O O O O O |
20 ++--------------------------------------------------------------------+
[*] bisect-good sample
[O] bisect-bad sample
Disclaimer:
Results have been estimated based on internal Intel analysis and are provided
for informational purposes only. Any difference in system hardware or software
design or configuration may affect actual performance.
Thanks,
Xiaolong
View attachment "config-4.13.0-rc2-00028-g767427002" of type "text/plain" (160950 bytes)
View attachment "job-script" of type "text/plain" (6786 bytes)
View attachment "job.yaml" of type "text/plain" (4430 bytes)
View attachment "reproduce" of type "text/plain" (287 bytes)
Powered by blists - more mailing lists