lists.openwall.net   lists  /  announce  owl-users  owl-dev  john-users  john-dev  passwdqc-users  yescrypt  popa3d-users  /  oss-security  kernel-hardening  musl  sabotage  tlsify  passwords  /  crypt-dev  xvendor  /  Bugtraq  Full-Disclosure  linux-kernel  linux-netdev  linux-ext4  linux-hardening  linux-cve-announce  PHC 
Open Source and information security mailing list archives
 
Hash Suite: Windows password security audit tool. GUI, reports in PDF.
[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <1470834130.1954.5.camel@suse.cz>
Date:	Wed, 10 Aug 2016 15:02:10 +0200
From:	Giovanni Gherdovich <ggherdovich@...e.cz>
To:	Ingo Molnar <mingo@...nel.org>
Cc:	Ingo Molnar <mingo@...hat.com>,
	Peter Zijlstra <peterz@...radead.org>,
	Mike Galbraith <mgalbraith@...e.de>,
	Stanislaw Gruszka <sgruszka@...hat.com>,
	linux-kernel@...r.kernel.org,
	Mel Gorman <mgorman@...hsingularity.net>
Subject: Re: [PATCH 1/1] sched/cputime: Mitigate performance regression in
 times()/clock_gettime()

Hello Ingo,

thank you for your reply.

Ingo Molnar <mingo@...nel.org>
> Nice detective work! I'm wondering, where do we stand if compared with a 
> pre-6e998916dfe3 kernel?

The data follows. A considerable part of the performance loss is recovered;
something is still on the table.

"3.18-pre-bug" is the parent of 6e998916dfe3, i.e. 6e998916dfe3^1
"3.18-bug" is the revision 6e998916dfe3 itself.
Figures are in seconds. Percentages refer to 3.18-pre-bug, negative = worse.


times()

threads    3.18-pre-bug          3.18-bug              4.7.0-rc7             4.7.0-rc7-patched

2          2.27 (  0.00%)        3.73 (-64.71%)        3.65 (-61.14%)        3.06 (-35.16%)
5          2.78 (  0.00%)        3.77 (-35.56%)        3.45 (-23.98%)        3.25 (-16.79%)
8          2.79 (  0.00%)        4.41 (-57.71%)        3.52 (-26.05%)        3.26 (-16.53%)
12         3.02 (  0.00%)        3.56 (-17.94%)        3.29 ( -9.08%)        3.37 (-11.74%)
21         3.10 (  0.00%)        4.61 (-48.74%)        4.07 (-31.34%)        3.90 (-25.89%)
30         3.33 (  0.00%)        5.75 (-72.53%)        3.87 (-16.01%)        3.36 ( -0.81%)
48         2.96 (  0.00%)        6.06 (-105.04%)       3.79 (-28.10%)        3.11 ( -5.14%)
79         2.88 (  0.00%)        6.24 (-116.83%)       3.88 (-34.81%)        3.16 ( -9.84%)
110        2.98 (  0.00%)        6.37 (-114.08%)       3.90 (-31.12%)        3.25 ( -9.07%)
128        3.10 (  0.00%)        6.35 (-104.61%)       4.00 (-28.87%)        3.31 ( -6.57%)


clock_gettime()

threads    3.18-pre-bug          3.18-bug              4.7.0-rc7             4.7.0-rc7-patched

2          2.23 (  0.00%)        3.68 (-64.56%)        3.48 (-55.48%)        3.25 (-45.41%)
5          2.83 (  0.00%)        3.78 (-33.42%)        3.33 (-17.43%)        3.17 (-12.03%)
8          2.84 (  0.00%)        4.31 (-52.12%)        3.37 (-18.76%)        3.22 (-13.43%)
12         3.09 (  0.00%)        3.61 (-16.74%)        3.32 ( -7.17%)        3.36 ( -8.47%)
21         3.14 (  0.00%)        4.63 (-47.36%)        4.01 (-27.71%)        3.92 (-24.68%)
30         3.28 (  0.00%)        5.75 (-75.37%)        3.63 (-10.80%)        3.40 ( -3.69%)
48         3.02 (  0.00%)        6.05 (-100.56%)       3.71 (-22.99%)        3.16 ( -4.64%)
79         2.88 (  0.00%)        6.30 (-118.90%)       3.75 (-30.26%)        3.28 (-13.93%)
110        2.95 (  0.00%)        6.46 (-119.00%)       3.81 (-29.24%)        3.38 (-14.69%)
128        3.05 (  0.00%)        6.42 (-110.08%)       3.88 (-27.04%)        3.38 (-10.70%)


Regards,
Giovanni Gherdovich

Powered by blists - more mailing lists

Powered by Openwall GNU/*/Linux Powered by OpenVZ