From mboxrd@z Thu Jan 1 00:00:00 1970 From: Tobias Doerffel Subject: Extreme memory usage when running USDX with Intel driver Date: Tue, 4 May 2010 01:19:01 +0200 Message-ID: <201005040119.08034.tobias.doerffel@gmail.com> Mime-Version: 1.0 Content-Type: multipart/mixed; boundary="===============1301613225==" Return-path: Received: from jessica.hrz.tu-chemnitz.de (jessica.hrz.tu-chemnitz.de [134.109.132.47]) by gabe.freedesktop.org (Postfix) with ESMTP id 3ED029E754 for ; Mon, 3 May 2010 16:19:11 -0700 (PDT) Received: from vpnclient-194-241.hrz.tu-chemnitz.de ([134.109.194.241] helo=nc10.localnet) by jessica.hrz.tu-chemnitz.de with esmtpsa (TLSv1:AES256-SHA:256) (Exim 4.69) (envelope-from ) id 1O94ur-0004kY-DP for intel-gfx@lists.freedesktop.org; Tue, 04 May 2010 01:19:09 +0200 List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , Mime-version: 1.0 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: intel-gfx@lists.freedesktop.org List-Id: intel-gfx@lists.freedesktop.org --===============1301613225== Content-Type: multipart/signed; boundary="nextPart1491484.Lvu9uK9pxZ"; protocol="application/pgp-signature"; micalg=pgp-sha1 Content-Transfer-Encoding: 7bit --nextPart1491484.Lvu9uK9pxZ Content-Type: Text/Plain; charset="us-ascii" Content-Transfer-Encoding: quoted-printable Hi, I have serious problems with running UltraStar Deluxe (a free OpenGL-based karaoke program) on an Intel graphic card (945GME). When starting the progr= am, the system starts to use more than 1,5 GB of RAM for (I guess) GEM stuff. I= =20 can't blame the program as it's running fine with both swrast (on the same= =20 computer and setup) and the nouveau driver on another computer and itself=20 consumes about 100 MB of RAM. I collected some data describing the system status after the program has be= en=20 started up (and switched to tty0 in order to get these values). With Mesa 7.9 /proc/dri/0/gem_object tells me 8996 objects =2D1422700544 object bytes 3 pinned 19365888 pin bytes 140005376 gtt bytes 260308992 gtt total With Mesa 7.8.1 it looks like the following: 2655 objects =2D1792622592 object bytes 3 pinned 19365888 pin bytes 233402368 gtt bytes 260308992 gtt total At the same time /proc//smaps has about 2500 similiar entries looking= =20 like 1007e000-1017e000 rw-s 20ac0e000 00:05 314 /dev/dri/card0 Size: 1024 kB Rss: 0 kB Pss: 0 kB Shared_Clean: 0 kB Shared_Dirty: 0 kB Private_Clean: 0 kB Private_Dirty: 0 kB Referenced: 0 kB Swap: 0 kB KernelPageSize: 4 kB MMUPageSize: 4 kB /proc/dri/0/gem_names contains name size handles refcount name 1 size 4194304 1 4194304 2 5 name 2 size 4194304 2 4194304 2 4 name 3 size 16777216 3 16777216 1 4 My setup is the latest kernel with latest commits from the drm-intel-next=20 branch. Same for xf86-video-intel and libdrm (both compiled from Git). XSer= ver=20 version is 1.8. Mesa as written above. What could I do further to track dow= n=20 the problem? =46rom what I saw in the source code, the program creates lots of textures = for individual glyphs of fonts. The extreme memory usage seems to start while t= he=20 several thousands of textures are being created. I'm just wondering that=20 everything works fine with other drivers. Best regards, Toby --nextPart1491484.Lvu9uK9pxZ Content-Type: application/pgp-signature; name=signature.asc Content-Description: This is a digitally signed message part. -----BEGIN PGP SIGNATURE----- Version: GnuPG v1.4.10 (GNU/Linux) iEYEABECAAYFAkvfWeYACgkQHsjdpw2OaadiYACgoMGiLdH1Omej2Fz+4cwP5hmc c4wAnAgASElY37wNPaO9BweC3fceWpLd =Snfw -----END PGP SIGNATURE----- --nextPart1491484.Lvu9uK9pxZ-- --===============1301613225== Content-Type: text/plain; charset="us-ascii" MIME-Version: 1.0 Content-Transfer-Encoding: 7bit Content-Disposition: inline _______________________________________________ Intel-gfx mailing list Intel-gfx@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/intel-gfx --===============1301613225==--