From mboxrd@z Thu Jan 1 00:00:00 1970 From: Chris Wilson Subject: Re: [PATCH] drm/i915: set IDICOS to medium uncore resources Date: Mon, 16 Jul 2012 11:29:49 +0100 Message-ID: <1342434609_3829@CP5-2952> References: <1336183139-10286-1-git-send-email-ben@bwidawsk.net> <20120618083554.GB4773@phenom.ffwll.local> Mime-Version: 1.0 Content-Type: text/plain; charset="us-ascii" Content-Transfer-Encoding: 7bit Return-path: Received: from fireflyinternet.com (smtp.fireflyinternet.com [109.228.6.236]) by gabe.freedesktop.org (Postfix) with ESMTP id E80929E790 for ; Mon, 16 Jul 2012 03:30:15 -0700 (PDT) In-Reply-To: <20120618083554.GB4773@phenom.ffwll.local> List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , Sender: intel-gfx-bounces+gcfxdi-intel-gfx=m.gmane.org@lists.freedesktop.org Errors-To: intel-gfx-bounces+gcfxdi-intel-gfx=m.gmane.org@lists.freedesktop.org To: Daniel Vetter , Ben Widawsky Cc: intel-gfx@lists.freedesktop.org, "Kilarski, Bernard R" List-Id: intel-gfx@lists.freedesktop.org On Mon, 18 Jun 2012 10:35:54 +0200, Daniel Vetter wrote: > On Fri, May 04, 2012 at 06:58:59PM -0700, Ben Widawsky wrote: > > I'm seeing about a 5% FPS improvement across various benchmarks on my > > IVB i3. Rumor has it that the higher end parts show even more benefit. > > > > This derives from a patch originally given to me by Bernard. The docs > > are confusing about the definition names (ie. medium really seems like > > max), but it would seem it gives more cache to the GT at the expense of > > uncore. This configuration makes the split most in favor of the GT. I've > > not tried the other IDICOS values. > > > > Cc: "Kilarski, Bernard R" > > Signed-off-by: Ben Widawsky > Queued for -next, with Eric's ack added. Although I've had fix the patch, > it didn't compile :( So this introduces a 10% (205 to 185 fps on an i7-3720qm @10x7) performance regression on good old CPU bound padman, but one large win with alpha compositing the same texture over and over again (fishtank 7.8s to 6.8s). Oh well, there's plenty of overhead in mesa to recover -- reading the profiles it is rate limited by the cpu overhead in i965_dri.so. -Chris -- Chris Wilson, Intel Open Source Technology Centre