From mboxrd@z Thu Jan 1 00:00:00 1970 From: Marcin Slusarz Subject: Re: [kernel PATCH] drm/nouveau: add NOUVEAU_GEM_RELOC_HILO Date: Wed, 12 Jan 2011 22:41:24 +0100 Message-ID: <20110112214124.GA22901@joi.lan> References: <1294865031-22457-1-git-send-email-marcin.slusarz@gmail.com> <87d3o1j0d4.fsf@riseup.net> Mime-Version: 1.0 Content-Type: text/plain; charset="us-ascii" Content-Transfer-Encoding: 7bit Return-path: Content-Disposition: inline In-Reply-To: <87d3o1j0d4.fsf-sGOZH3hwPm2sTnJN9+BGXg@public.gmane.org> List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , Sender: nouveau-bounces+gcfxn-nouveau=m.gmane.org-PD4FTy7X32lNgt0PjOBp9y5qC8QIuHrW@public.gmane.org Errors-To: nouveau-bounces+gcfxn-nouveau=m.gmane.org-PD4FTy7X32lNgt0PjOBp9y5qC8QIuHrW@public.gmane.org To: Francisco Jerez Cc: nouveau-PD4FTy7X32lNgt0PjOBp9y5qC8QIuHrW@public.gmane.org List-Id: nouveau.vger.kernel.org On Wed, Jan 12, 2011 at 10:08:55PM +0100, Francisco Jerez wrote: > Marcin Slusarz writes: > > > add new relocation flag which allows to halving number of relocation > > entries on >= nv50 cards > > > I suspect that this change is unnecessary. If you're actually seeing a > performance improvement (I guess you are?), it's very likely to be I did it for performance reasons, but actual improvement is very small (0.3%). I sent these patches, because it's just better to do less and use less memory. > masking the actual root of the problem -- last time I checked the nv50 > gallium pipe driver was rather careless about reloc re-submission (it > re-submits the whole reloc chain once for every draw call instead of > once for every pushbuf flush), and nouveau_reloc_emit() and friends are > very unoptimized. > > I'm betting that using HILO or not will make pretty much no difference > once those two places are fixed :). It looks they are already fixed :). Well, you can treat these patches as data (relocations are not a bottleneck) and forget about them. Marcin