From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: from 173-166-109-252-newengland.hfc.comcastbusiness.net ([173.166.109.252]:60067 "EHLO bombadil.infradead.org" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1751677Ab3BRIJz (ORCPT ); Mon, 18 Feb 2013 03:09:55 -0500 Date: Mon, 18 Feb 2013 09:08:29 +0100 From: Jens Axboe Subject: Re: How to get IOPS when using terse output Message-ID: <20130218080829.GC6003@kernel.dk> References: <511B6E9C.3000903@gmail.com> MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline In-Reply-To: <511B6E9C.3000903@gmail.com> Sender: fio-owner@vger.kernel.org List-Id: fio@vger.kernel.org To: Peter Xu Cc: fio@vger.kernel.org On Wed, Feb 13 2013, Peter Xu wrote: > Hi, all, > > I failed to find IOPS results for read/write under terse mode (with > --minimal param when invoking fio program), while IOPS does exist in normal > output without --minimal param. > > It seems that I even cannot calculate it out only with the tersed output > (since I don't know the total number of IO). Then, how should I get IOPS > results when using terse output? As Carl mentions, the IOPS are there. But let me make a suggestion to move over to the json format instead. It's easy to parse by computers and humans, and new performance metrics can be added without causing breakage in existing setups. The minimal/csv format is somewhat fragile for that. That said, IOPS is there in the minimal output :-) -- Jens Axboe