From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Subject: Re: fio 3.2 References: <903d418b-bac4-104b-28e5-3c529efab7f5@kernel.dk> From: Jens Axboe Message-ID: <40a7b698-e78f-4b82-cdb5-4629d19dff3f@kernel.dk> Date: Wed, 29 Nov 2017 21:04:13 -0700 MIME-Version: 1.0 In-Reply-To: Content-Type: text/plain; charset="utf-8" Content-Language: en-US Content-Transfer-Encoding: 8bit To: Rebecca Cran , "Elliott, Robert (Persistent Memory)" Cc: "Gavriliuk, Anton (HPS Ukraine)" , Sitsofe Wheeler , "fio@vger.kernel.org" , "Kani, Toshimitsu" List-ID: On 11/28/2017 09:44 PM, Rebecca Cran wrote: > I’ve not followed the whole discussion, but I did some tests a couple > of months ago replacing the calls to memcpy in the memmap engine with > a new fio_memcpy function, rebuilding fio with the -ftree-vectorize > flag along with -msse2, -mavx etc. I saw an improvement of a few > hundred MB/sec on my openSUSE system. Care to send a patch for that? I think it'd be nice for fio to pack it's own, especially when the performance varies so wildly and is platform (and even version) dependent. -- Jens Axboe