From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1751768AbaIMEqi (ORCPT ); Sat, 13 Sep 2014 00:46:38 -0400 Received: from mail-pa0-f42.google.com ([209.85.220.42]:55160 "EHLO mail-pa0-f42.google.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1751123AbaIMEqh (ORCPT ); Sat, 13 Sep 2014 00:46:37 -0400 Date: Sat, 13 Sep 2014 13:46:30 +0900 From: Tejun Heo To: Alexander Gordeev Cc: linux-kernel@vger.kernel.org, linux-ide@vger.kernel.org Subject: Re: [PATCH RESEND 0/1] AHCI: Optimize interrupt processing Message-ID: <20140913044630.GB30477@mtj.dyndns.org> References: <20130521235003.GE6985@mtj.dyndns.org> <20140911124248.GA18490@dhcp-27-160.brq.redhat.com> MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline In-Reply-To: <20140911124248.GA18490@dhcp-27-160.brq.redhat.com> User-Agent: Mutt/1.5.23 (2014-03-12) Sender: linux-kernel-owner@vger.kernel.org List-ID: X-Mailing-List: linux-kernel@vger.kernel.org Hello, On Thu, Sep 11, 2014 at 02:42:49PM +0200, Alexander Gordeev wrote: > As conversion of libata to blk mq has long done I tried the change Hmmm? You mean scsi-mq? > against the recent version and the results still appear worthwhile. > > The numbers are taken by running 'dd if=/dev/sd{a,b} of=/dev/null' > in parallel. All time values are in us. > > Before this update host lock average holdtime was 2.45 and > average waittime was 1.24. After the update average holdtime > dropped to 0.29 (about eight times) while average waittime > decreased to 0.58 (about two times). > > Also, port events are handled with local interrupts enabled > and compete on individual per-port locks with average holdtime > 1.25 and average waittime 1.48. So combined average holdtime > spent while holding host and port locks decreased from 2.45 to > 0.29 + 1.25 = 1.54 (about 1.6 times). > > The downside of this change is introduction of a kernel thread. That shouldn't matter at all but can you please present the information in a more digestable form? e.g. CPU usage decreased from A to B when transferring N MB/s on certain setup. Thanks. -- tejun