Content-Type: multipart/mixed; boundary="===============3441113901273869493==" MIME-Version: 1.0 From: Jet Chen To: lkp@lists.01.org Subject: Re: [libata/ahci] 8a4aeec8d2d: +138.4% perf-stat.dTLB-store-misses, +37.2% perf-stat.dTLB-load-misses Date: Wed, 23 Apr 2014 16:21:06 +0800 Message-ID: <535777F2.1050501@intel.com> In-Reply-To: List-Id: --===============3441113901273869493== Content-Type: text/plain; charset="utf-8" MIME-Version: 1.0 Content-Transfer-Encoding: quoted-printable On 04/23/2014 01:11 AM, Dan Williams wrote: > On Mon, Apr 21, 2014 at 12:29 AM, Jet Chen wrote: >> HI Dan, >> >> we noticed the below changes on >> >> git://git.kernel.org/pub/scm/linux/kernel/git/tj/libata for-next >> commit 8a4aeec8d2d6a3edeffbdfae451cdf05cbf0fefd ("libata/ahci: accommoda= te >> tag ordered controllers") > = > Hi, was this on simulated hardware or a real AHCI controller and disk? > = Testing was on a physical machine with a real AHCI controller. root(a)bay ~# lspci | grep AHCI 00:1f.2 SATA controller: Intel Corporation 82801IR/IO/IH (ICH9R/DO/DH) 6 po= rt SATA Controller [AHCI mode] (rev 02) > It does appear this test noticed increased throughput: > = > 203893 ~ 0% +3.7% 211474 ~ 0% TOTAL iostat.sda.wkB/s > = > I wonder if ap->last_tag can be moved to a hotter cacheline, but if > throughput goes up I can imagine it throws off the cpu statistics > quite a bit. >=20 --===============3441113901273869493==--