Path: csiph.com!newsfeed.hal-mli.net!feeder3.hal-mli.net!newsfeed.hal-mli.net!feeder1.hal-mli.net!npeer01.iad.highwinds-media.com!news.highwinds-media.com!feed-me.highwinds-media.com!postnews.google.com!glegroupsg2000goo.googlegroups.com!not-for-mail From: John S Newsgroups: comp.os.linux.development.system Subject: Re: Howto - or - Feature Request - Process to Core locking Date: Sun, 11 Mar 2012 03:02:56 -0700 (PDT) Organization: http://groups.google.com Lines: 22 Message-ID: <20606109.96.1331460176200.JavaMail.geo-discussion-forums@vbze11> References: <25803231.739.1331157626474.JavaMail.geo-discussion-forums@vbbfv2> NNTP-Posting-Host: 24.126.95.12 Mime-Version: 1.0 Content-Type: text/plain; charset=ISO-8859-1 X-Trace: posting.google.com 1331460284 30100 127.0.0.1 (11 Mar 2012 10:04:44 GMT) X-Complaints-To: groups-abuse@google.com NNTP-Posting-Date: Sun, 11 Mar 2012 10:04:44 +0000 (UTC) In-Reply-To: Complaints-To: groups-abuse@google.com Injection-Info: glegroupsg2000goo.googlegroups.com; posting-host=24.126.95.12; posting-account=-NFKmwoAAAAl0IsBTNx2_eZpsoQrFUwM User-Agent: G2/1.0 X-Received-Bytes: 2033 Xref: csiph.com comp.os.linux.development.system:405 On Wednesday, March 7, 2012 7:17:57 PM UTC-5, Ersek, Laszlo wrote: > On Wed, 7 Mar 2012, John S wrote: > > > We have tried cpuset and taskset to control affinity, but these > > haven't fully resolved the issues. The CPU scheduler will still > > utilize an engine's core even though other cores are either unused or > > have low useage. > > I may be misunderstanding this: have you tried restricting everything else > to the complement CPU set? > > (Eg. find the one script that's used to start "everything" in the system > when it reaches the preferred multi-user runlevel. (On Debian > /etc/init.d/rc seems possible.) Use "taskset -p complement_mask $$" in it; > it's children will inherit the affinity. Use "taskset mask_i engine i" > (like now) to start a given engine instance.) > > Laszlo Will give it a try. Thanks John