From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: from mail.free-electrons.com (top.free-electrons.com [176.31.233.9]) by yocto-www.yoctoproject.org (Postfix) with ESMTP id 54241E01493 for ; Thu, 29 Aug 2013 01:01:45 -0700 (PDT) Received: by mail.free-electrons.com (Postfix, from userid 106) id A45C7844; Thu, 29 Aug 2013 10:01:57 +0200 (CEST) X-Spam-Checker-Version: SpamAssassin 3.3.2 (2011-06-06) on mail.free-electrons.com X-Spam-Level: X-Spam-Status: No, score=-1.0 required=5.0 tests=ALL_TRUSTED,SHORTCIRCUIT, URIBL_BLOCKED shortcircuit=ham autolearn=disabled version=3.3.2 Received: from [192.168.1.53] (128-79-216-144.hfc.dyn.abo.bbox.fr [128.79.216.144]) by mail.free-electrons.com (Postfix) with ESMTPSA id 62614817; Thu, 29 Aug 2013 10:01:57 +0200 (CEST) Message-ID: <521EFFE7.8020304@free-electrons.com> Date: Thu, 29 Aug 2013 10:01:43 +0200 From: Alexandre Belloni Organization: Free Electrons User-Agent: Mozilla/5.0 (X11; Linux x86_64; rv:17.0) Gecko/20130803 Thunderbird/17.0.8 MIME-Version: 1.0 To: =?ISO-8859-1?Q?Eric_B=E9nard?= References: <521E28B7.7040105@freescale.com> <521E3159.7010908@freescale.com> <20130828200037.48571a17@e6520eb> <521E4189.9060005@freescale.com> <20130828222721.07cd5727@e6520eb> In-Reply-To: <20130828222721.07cd5727@e6520eb> X-Enigmail-Version: 1.4.6 Cc: meta-freescale@yoctoproject.org Subject: Re: GPU tests on master next - core-image-x11 - sfp X-BeenThere: meta-freescale@yoctoproject.org X-Mailman-Version: 2.1.13 Precedence: list List-Id: Usage and development list for the meta-fsl-* layers List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Thu, 29 Aug 2013 08:01:46 -0000 Content-Type: text/plain; charset=ISO-8859-1 Content-Transfer-Encoding: 8bit Hi, On 28/08/2013 22:27, Eric Bénard wrote: > > on the GPU point of view no. > > I find interesting that the hardfp gives lower results than the > softfp : maybe that would be interesting to have true real life > benchmarks on this side (not GPU centric) before defaulting hardfp in > meta-fsl-arm for i.MX6 platforms. I'm not sure we can actually find real life benchmarks, it always depends on how intensive are the floating point operation compared to the rest of the system. Also, I remember seeing that the difference between softfp and hardfp is not as big on cortex-a9 as on cortex-a8. Actually, softfp is already using the VFP but using the soft ABI. Then, it has to copy values from integer registers to float registers and that is the overhead. This overhead is lower on cortex-a9 because the vfp is pipelined, this was not the case on cortex -a8. You can find interesting benchs here: https://wiki.linaro.org/OfficeofCTO/HardFloat/Benchmarks https://wiki.linaro.org/OfficeofCTO/HardFloat/Benchmarks201205 -- Alexandre Belloni, Free Electrons Embedded Linux, Kernel and Android engineering http://free-electrons.com