Comments on: On guts and I/O schedulers http://dom.as/2007/12/13/io-scheduler-deadline-cfq/ Sat, 17 Dec 2011 20:14:59 +0000 hourly 1 http://wordpress.com/ By: domas mituzas: vaporware, inc. » Blog Archive » I/O schedulers seriously revisited http://dom.as/2007/12/13/io-scheduler-deadline-cfq/#comment-1209 Tue, 05 Feb 2008 10:41:03 +0000 http://dammit.lt/2007/12/13/io-scheduler-deadline-cfq/#comment-1209 [...] I/O scheduler problems have drawn my attention, and besides trusting empirical results, I tried to do more of benchmarking [...]

]]> By: Hashar http://dom.as/2007/12/13/io-scheduler-deadline-cfq/#comment-1208 Thu, 13 Dec 2007 17:23:07 +0000 http://dammit.lt/2007/12/13/io-scheduler-deadline-cfq/#comment-1208 Have a look at “Enhancements to Linux I/O Scheduling”:

http://www.linuxinsight.com/files/ols2005/seelam-reprint.pdf

]]> By: sapphirecat http://dom.as/2007/12/13/io-scheduler-deadline-cfq/#comment-1207 Thu, 13 Dec 2007 14:11:07 +0000 http://dammit.lt/2007/12/13/io-scheduler-deadline-cfq/#comment-1207 CFQ is meant to give a predictable, low latency to I/O requests, so that runnnig updatedb or starting Firefox doesn’t make your music skip. Deadline and anticipatory are oriented on the traditional measure of scheduler performance: raw, screaming throughput. Of the latter two, deadline is simpler, and the kernel config suggests it may be better for databases, but that’s something I’ve never bothered to benchmark.

]]>