From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: from mga14.intel.com (mga14.intel.com [192.55.52.115]) by gabe.freedesktop.org (Postfix) with ESMTPS id 1A88A10E039 for ; Sat, 21 Jan 2023 02:05:48 +0000 (UTC) Date: Fri, 20 Jan 2023 18:05:46 -0800 Message-ID: <87ilh0vdz9.wl-ashutosh.dixit@intel.com> From: "Dixit, Ashutosh" To: Kamil Konieczny , igt-dev@lists.freedesktop.org, Ashutosh Dixit , Badal Nilawar , Tvrtko Ursulin , Riana Tauro In-Reply-To: <20230120180009.kh7damckyobbigap@kamilkon-desk1> References: <20230120163831.229411-1-ashutosh.dixit@intel.com> <20230120180009.kh7damckyobbigap@kamilkon-desk1> MIME-Version: 1.0 (generated by SEMI-EPG 1.14.7 - "Harue") Content-Type: text/plain; charset=US-ASCII Subject: Re: [igt-dev] [PATCH i-g-t] tools/intel_gpu_power: Intel GPU idle/busy power measurement List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , Errors-To: igt-dev-bounces@lists.freedesktop.org Sender: "igt-dev" List-ID: On Fri, 20 Jan 2023 10:00:09 -0800, Kamil Konieczny wrote: > > Hi, > Hi Kamil, > On 2023-01-20 at 08:38:31 -0800, Ashutosh Dixit wrote: > > In several instances (e.g. when investigating GPU power limits) it is very > > useful to be able to measure GPU power easily. Since we already have all > > ingredients for doing so, add a tool to measure Intel GPU power when idle > > and power under load. > > > > v2: Use softpin/allocator (Riana) > > > > Signed-off-by: Ashutosh Dixit > > These looks like a test or benchmark, btw intel_gpu_top already > has printing used power. What we can have is standalone app which > shows only power consumption every few seconds or some other > configurable interval, plus maybe for all GPUs if there are more. > > +cc Tvrtko Ursulin You are absolutely right, I agree on all points. I completely missed that intel_gpu_top already shows power. At this point it only shows power for integrated, we need to switch it from the rapl interface to igt_power interface, igt_power supports both integrated and discrete (integrated using rapl and discrete using hwmon). I might resumbit this patch as a test, but tools wise I think the direction should be to extend intel_gpu_top to support discrete. Thanks. -- Ashutosh