From mboxrd@z Thu Jan 1 00:00:00 1970 From: Tobias Jakobi Subject: Re: drm/exynos: add async G2D execution to libdrm Date: Sat, 21 Mar 2015 15:03:37 +0100 Message-ID: <550D7A39.4080502@gmx.net> References: <1426890348-12807-1-git-send-email-tjakobi@math.uni-bielefeld.de> Mime-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 7bit Return-path: Received: from mout.gmx.net ([212.227.17.21]:64568 "EHLO mout.gmx.net" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1751349AbbCUODw (ORCPT ); Sat, 21 Mar 2015 10:03:52 -0400 In-Reply-To: <1426890348-12807-1-git-send-email-tjakobi@math.uni-bielefeld.de> Sender: linux-samsung-soc-owner@vger.kernel.org List-Id: linux-samsung-soc@vger.kernel.org To: Tobias Jakobi , linux-samsung-soc@vger.kernel.org Cc: dri-devel@lists.freedesktop.org, emil.l.velikov@gmail.com, jy0922.shim@samsung.com, m.szyprowski@samsung.com, inki.dae@samsung.com Tobias Jakobi wrote: > Hello, > > this series exposes async execution of G2D command buffers to userspace. Also includes is a small performance analysis test, which can also be used to > stress test the engine. The async operation is of course also tested. > > Please review and let me know what I can improve. > > With best wishes, > Tobias Almost forgot this. In case someone is interested in the solid fill performance, I've uploaded a plot here: http://www.math.uni-bielefeld.de/~tjakobi/exynos/g2d_clear_perf.pdf And yes, that should be _nano_seconds on the y-axis. What I find interesting that despite the performance being linear (well, affine linear, since we have a setup time) in the number of pixels to process, you can still see some kind of 'clustering'. I wonder where this comes from, maybe some small alignment penalty, or something? With best wishes, Tobias