From info-volumizer-owner@els.sgi.com Sun Nov 2 23:19:19 2003 Received: (from majordom@localhost) by els.sgi.com (8.12.10/8.12.2/els-2.2) id hA37FQOm6172717; Sun, 2 Nov 2003 23:15:26 -0800 (PST) Received: from ext-nj2gw-3.online-age.net (ext-nj2gw-3.online-age.net [216.35.73.165]) by els.sgi.com (8.12.10/8.12.2/els-2.2) with ESMTP id hA350DCx6148365 for ; Sun, 2 Nov 2003 21:00:14 -0800 (PST) Received: from int-nj2gw-3.online-age.net (int-nj2gw-3 [3.159.236.67]) by ext-nj2gw-3.online-age.net (8.12.9/8.12.9/990426-RLH) with ESMTP id hA34uCw7005961 for ; Sun, 2 Nov 2003 23:56:13 -0500 (EST) Received: from cinmlef03.e2k.ad.ge.com (localhost [127.0.0.1]) by int-nj2gw-3.online-age.net (8.12.9/8.12.8/990426-RLH) with ESMTP id hA34uBXH003657 for ; Sun, 2 Nov 2003 23:56:12 -0500 (EST) Received: from banmlvem01.e2k.ad.ge.com ([3.159.224.30]) by cinmlef03.e2k.ad.ge.com with Microsoft SMTPSVC(5.0.2195.6673); Sun, 2 Nov 2003 23:56:10 -0500 content-class: urn:content-classes:message MIME-Version: 1.0 Content-Type: text/plain; charset="iso-8859-1" X-MimeOLE: Produced By Microsoft Exchange V6.0.6487.1 Subject: [info-volumizer] openGL volumizer 2.4 for linux Date: Mon, 3 Nov 2003 10:25:42 +0530 Message-ID: <62DD37292ED5464CBB142913FC65F8AB062985F8@BANMLVEM01.e2k.ad.ge.com> X-MS-Has-Attach: X-MS-TNEF-Correlator: Thread-Topic: openGL volumizer 2.4 for linux Thread-Index: AcOhxnWoMoXS2T3URuWTnHD1ZL3Kww== From: "S, Maheshanand (MED, Contractor)" To: X-OriginalArrivalTime: 03 Nov 2003 04:56:11.0020 (UTC) FILETIME=[CCED34C0:01C3A1C6] Content-Transfer-Encoding: 8bit X-MIME-Autoconverted: from quoted-printable to 8bit by els.sgi.com id hA350ECx6218656 Sender: owner-info-volumizer@els.sgi.com Precedence: bulk Reply-To: info-volumizer@els.sgi.com Hi >From where i can able to download openGL volumizer 2.4 .I found this link is providing downloads for IRIX series only.. http://www.sgi.com/products/evaluation/index.html#volumizer plz give me the suitable link for free downloading openGL volumizer 2.4 for Red hat Linux 7.2 thanx in advance rgds, mahesh. From info-volumizer-owner@els.sgi.com Sun Nov 2 23:19:19 2003 Received: (from majordom@localhost) by els.sgi.com (8.12.10/8.12.2/els-2.2) id hA37FQwi6099569; Sun, 2 Nov 2003 23:15:26 -0800 (PST) Received: from franklin.boulder.nist.gov (franklin-node1.boulder.nist.gov [132.163.128.81]) by els.sgi.com (8.12.10/8.12.2/els-2.2) with ESMTP id h9UKo9Cx6133913 for ; Thu, 30 Oct 2003 12:50:12 -0800 (PST) Received: from boulder.nist.gov (canvas.boulder.nist.gov [132.163.129.134]) by franklin.boulder.nist.gov (Switch-2.2.8/Switch-2.2.8) with ESMTP id h9UKkJP14754 for ; Thu, 30 Oct 2003 13:46:19 -0700 (MST) Message-ID: <3FA17892.50900@boulder.nist.gov> Date: Thu, 30 Oct 2003 13:46:10 -0700 From: "Adele P. Peskin" Organization: NIST User-Agent: Mozilla/5.0 (X11; U; SunOS sun4u; en-US; rv:1.0.1) Gecko/20020921 Netscape/7.0 X-Accept-Language: en-us, en MIME-Version: 1.0 To: info-volumizer@els.sgi.com Subject: [info-volumizer] volume loader question Content-Type: text/plain; charset=us-ascii; format=flowed Content-Transfer-Encoding: 7bit X-Scanned-By: milter-spamc/0.6 X-Spam-Flag: NO X-Spam-Status: NO, hits=-105.40 required=5.00 X-Spam-Level: Sender: owner-info-volumizer@els.sgi.com Precedence: bulk Reply-To: info-volumizer@els.sgi.com (This has also been posted to the Performer mailing list in case you have already seen this.) Hello, This is a question about the volume loader, code in the /usr/share/Performer/src/lib/libpfdb/libpfvol directory, used to load in *.vol files. When I load in a *.vol file using the built-in loader, all is fine - my data loads in without a problem. I am trying to create my own version of this loader, and as a first step, I just copied the code and compiled it on my system. The DSO from my compiled code (unchanged from the source version except to rename the file extension to *.vl3 in the appropriate places) will not load and I get the message: 2763 PF Debug: dlopen said: 2763:perfly: rld: Fatal Error: unresolvable symbol in /usr/local/diverse/savg/lib/OPT.N32.OPENGL/libpfvl3.so: __T_16vzTMRenderAction In the code, the vzTMRenderAction class is overloaded with a pfTMRenderAction structure that includes the PFSTRUCT_DECLARE macro to overload the new operator. Does anyone have any idea why the recompiled code will not load? Thanks, Adele Peskin From info-volumizer-owner@els.sgi.com Mon Nov 3 00:30:59 2003 Received: (from majordom@localhost) by els.sgi.com (8.12.10/8.12.2/els-2.2) id hA38TKT66181729; Mon, 3 Nov 2003 00:29:20 -0800 (PST) Received: from micaela.bio.dist.unige.it (micaela.bio.dist.unige.it [130.251.10.15]) by els.sgi.com (8.12.10/8.12.2/els-2.2) with ESMTP id hA388lCx6166432 for ; Mon, 3 Nov 2003 00:08:48 -0800 (PST) Received: from dafne (dafne.bio.dist.unige.it [130.251.10.12]) by micaela.bio.dist.unige.it (Postfix) with ESMTP id 46C9023867 for ; Mon, 3 Nov 2003 09:05:21 +0100 (CET) Content-Type: text/plain; charset="iso-8859-1" From: Ivan Porro Organization: Biolab - DIST To: info-volumizer@els.sgi.com Subject: Re: [info-volumizer] openGL volumizer 2.4 for linux Date: Mon, 3 Nov 2003 09:06:11 +0100 User-Agent: KMail/1.4.3 References: <62DD37292ED5464CBB142913FC65F8AB062985F8@BANMLVEM01.e2k.ad.ge.com> In-Reply-To: <62DD37292ED5464CBB142913FC65F8AB062985F8@BANMLVEM01.e2k.ad.ge.com> MIME-Version: 1.0 Message-Id: <200311030906.11156.pivan@bio.dist.unige.it> Content-Transfer-Encoding: 8bit X-MIME-Autoconverted: from quoted-printable to 8bit by els.sgi.com id hA388mCx6158190 Sender: owner-info-volumizer@els.sgi.com Precedence: bulk Reply-To: info-volumizer@els.sgi.com I think you can find in many previous posts and out there on Internet that there isn't any version of volumizer for linux.... hope SGI will "fix" this in future :-) regards, Ivan At 05:55, luned́ 3 novembre 2003, S, Maheshanand (MED, Contractor) wrote: > Hi > From where i can able to download openGL volumizer 2.4 .I found this link > is providing downloads for IRIX series only.. > > http://www.sgi.com/products/evaluation/index.html#volumizer > > plz give me the suitable link for free downloading openGL volumizer 2.4 for > Red hat Linux 7.2 > > thanx in advance > > rgds, > mahesh. -- _________________________________ www.bio.dist.unige.it/~pivan _________________________________ A bus station is where a bus stops. A train station is where a train stops. On your desk, you have a workstation... From info-volumizer-owner@els.sgi.com Mon Nov 3 01:05:43 2003 Received: (from majordom@localhost) by els.sgi.com (8.12.10/8.12.2/els-2.2) id hA394Fjv6210298; Mon, 3 Nov 2003 01:04:15 -0800 (PST) Received: from sgindia.newdelhi.sgi.com (newdelhi.sgi.com [134.14.90.2]) by els.sgi.com (8.12.10/8.12.2/els-2.2) with ESMTP id hA38v5Cx6221759 for ; Mon, 3 Nov 2003 00:57:08 -0800 (PST) Received: from praveenlaptop ([134.14.90.74]) by sgindia.newdelhi.sgi.com (980427.SGI.8.8.8/980728.SGI.AUTOCF) via SMTP id NAA51799 for ; Mon, 3 Nov 2003 13:10:41 +0530 (IST) Message-ID: <015b01c3a1e7$e8e20630$4a5a0e86@praveenlaptop> From: "Praveen Bhaniramka" To: References: <62DD37292ED5464CBB142913FC65F8AB062985F8@BANMLVEM01.e2k.ad.ge.com> <200311030906.11156.pivan@bio.dist.unige.it> Subject: Re: [info-volumizer] openGL volumizer 2.4 for linux Date: Mon, 3 Nov 2003 14:23:10 +0530 MIME-Version: 1.0 Content-Type: text/plain; charset="iso-8859-1" Content-Transfer-Encoding: 8bit X-Priority: 3 X-MSMail-Priority: Normal X-Mailer: Microsoft Outlook Express 6.00.2800.1106 X-MimeOLE: Produced By Microsoft MimeOLE V6.00.2800.1106 Sender: owner-info-volumizer@els.sgi.com Precedence: bulk Reply-To: info-volumizer@els.sgi.com Mahesh and Ivan, Yes... this is being 'fixed' with the upcoming Volumizer release - Volumizer 2.6! We had announced the release of Volumizer on Linux during the Volumizer BOF meeting at the IEEE Visualization conference in Seattle last week. More details will be send to info-volumizer soon. Cheers, Praveen ----- Original Message ----- From: "Ivan Porro" To: Sent: Monday, November 03, 2003 1:36 PM Subject: Re: [info-volumizer] openGL volumizer 2.4 for linux > I think you can find in many previous posts and out there on Internet that > there isn't any version of volumizer for linux.... hope SGI will "fix" this > in future :-) > > regards, > > Ivan > > > At 05:55, luned́ 3 novembre 2003, S, Maheshanand (MED, Contractor) wrote: > > Hi > > From where i can able to download openGL volumizer 2.4 .I found this link > > is providing downloads for IRIX series only.. > > > > http://www.sgi.com/products/evaluation/index.html#volumizer > > > > plz give me the suitable link for free downloading openGL volumizer 2.4 for > > Red hat Linux 7.2 > > > > thanx in advance > > > > rgds, > > mahesh. > > -- > _________________________________ > www.bio.dist.unige.it/~pivan > _________________________________ > A bus station is where a bus stops. > A train station is where a train stops. > On your desk, you have a workstation... > From info-volumizer-owner@els.sgi.com Mon Nov 3 22:58:29 2003 Received: (from majordom@localhost) by els.sgi.com (8.12.10/8.12.2/els-2.2) id hA46vmWN6270450; Mon, 3 Nov 2003 22:57:48 -0800 (PST) Received: from sgindia.newdelhi.sgi.com (newdelhi.sgi.com [134.14.90.2]) by els.sgi.com (8.12.10/8.12.2/els-2.2) with ESMTP id hA46phCx6052139 for ; Mon, 3 Nov 2003 22:51:45 -0800 (PST) Received: from praveenlaptop ([134.14.90.74]) by sgindia.newdelhi.sgi.com (980427.SGI.8.8.8/980728.SGI.AUTOCF) via SMTP id LAA30492; Tue, 4 Nov 2003 11:04:41 +0530 (IST) Message-ID: <00c201c3a29f$7c5438e0$4a5a0e86@praveenlaptop> From: "Praveen Bhaniramka" To: "Adele P. Peskin" , , "info-volumizer" References: <3FA157D0.1080701@boulder.nist.gov> Subject: [info-volumizer] Re: [info-performer] volume loader question Date: Tue, 4 Nov 2003 12:17:12 +0530 MIME-Version: 1.0 Content-Type: text/plain; charset="iso-8859-1" Content-Transfer-Encoding: 7bit X-Priority: 3 X-MSMail-Priority: Normal X-Mailer: Microsoft Outlook Express 6.00.2800.1106 X-MimeOLE: Produced By Microsoft MimeOLE V6.00.2800.1106 Sender: owner-info-volumizer@els.sgi.com Precedence: bulk Reply-To: info-volumizer@els.sgi.com Hi Adele, Is your loader DSO linking against the Volumizer library - libvz.so? I am not sure if the PFSTRUCT_DECLARE macro is responsible for the error that you are getting. Cheers, Praveen ----- Original Message ----- From: "Adele P. Peskin" To: Sent: Thursday, October 30, 2003 11:56 PM Subject: [info-performer] volume loader question > Hello, > > This is a question about the volume loader, code in > the /usr/share/Performer/src/lib/libpfdb/libpfvol > directory, used to load in *.vol files. > > When I load in a *.vol file using the built-in > loader, all is fine - my data loads in without > a problem. > > I am trying to create my own version of this > loader, and as a first step, I just copied > the code and compiled it on my system. The > DSO from my compiled code (unchanged from > the source version except to rename the file > extension to *.vl3 in the appropriate places) > will not load and I get the message: > > 2763 PF Debug: dlopen said: 2763:perfly: rld: Fatal > Error: unresolvable symbol in > /usr/local/diverse/savg/lib/OPT.N32.OPENGL/libpfvl3.so: > __T_16vzTMRenderAction > > In the code, the vzTMRenderAction class is overloaded > with a pfTMRenderAction structure that includes the > PFSTRUCT_DECLARE macro to overload the new operator. > > Does anyone have any idea why the recompiled code > will not load? > > Thanks, Adele Peskin > > ----------------------------------------------------------------------- > List Archives, Info, FAQ: http://www.sgi.com/software/performer/ > Open Development Project: http://oss.sgi.com/projects/performer/ > Submissions: info-performer@sgi.com > Admin. requests: info-performer-request@sgi.com > ----------------------------------------------------------------------- > From info-volumizer-owner@els.sgi.com Thu Nov 6 19:13:15 2003 Received: (from majordom@localhost) by els.sgi.com (8.12.10/8.12.2/els-2.2) id hA73C5YD6338364; Thu, 6 Nov 2003 19:12:05 -0800 (PST) Received: from tolkor.sgi.com (tolkor.sgi.com [198.149.18.6]) by els.sgi.com (8.12.10/8.12.2/els-2.2) with ESMTP id hA6GlVCx6323756 for ; Thu, 6 Nov 2003 08:47:31 -0800 (PST) Received: from flecktone.americas.sgi.com (flecktone.americas.sgi.com [192.48.203.135]) by tolkor.sgi.com (8.12.9/8.12.9/linux-outbound_gateway-1.1) with ESMTP id hA6H2uHc023158 for ; Thu, 6 Nov 2003 11:02:56 -0600 Received: from spindle.corp.sgi.com (spindle.corp.sgi.com [198.29.75.13]) by flecktone.americas.sgi.com (8.12.9/8.12.9/generic_config-1.2) with ESMTP id hA6GhOP513264754 for ; Thu, 6 Nov 2003 10:43:24 -0600 (CST) Received: from mtv-atc-009e--n.corp.sgi.com (mtv-atc-009e--n.corp.sgi.com [192.26.60.130]) by spindle.corp.sgi.com (8.12.9/8.12.9/generic_config-1.2) with ESMTP id hA6GhDij120189368 for ; Thu, 6 Nov 2003 08:43:13 -0800 (PST) Received: by mtv-atc-009e--n.corp.sgi.com with Internet Mail Service (5.5.2655.55) id ; Thu, 6 Nov 2003 08:43:13 -0800 Message-ID: <401AB3C78169D711B9F50004ACE84267013847A3@mtv-atc-009e--n.corp.sgi.com> From: Thomas Tannert To: "'info-volumizer@els.sgi.com'" Subject: [info-volumizer] IEEE Vis 2003 OpenGL Volumizer and Scalability Techniques BOF Rev iew Date: Thu, 6 Nov 2003 08:43:12 -0800 MIME-Version: 1.0 X-Mailer: Internet Mail Service (5.5.2655.55) Content-Type: text/plain; charset="iso-8859-1" Sender: owner-info-volumizer@els.sgi.com Precedence: bulk Reply-To: info-volumizer@els.sgi.com *** IEEE Vis 2003 OpenGL Volumizer and Scalability Techniques BOF Review *** Hello Volumizers, In case you missed the announcement at IEEE Vis last month in Seattle, have we got some exciting news for you! During the Volumizer & Scalability BOF, SGI previewed the work underway to make OpenGL Volumizer available for the Linux platform. The port is one of several additions we are making for the upcoming release of OpenGL Volumizer, version 2.6. Results are already "very" promising. The meeting was held on the evening of Tuesday October 29th and was well attended (45+ attendees). At the BOF was the Onyx4 UltimateVision advanced visualization system running Volumizer demos, along with a preview of Volumizer 2.6 running on Linux IA32. The speakers at the event and a short recap of their presentations follows: Praveen Bhaniramka covered several major topics on Volumizer. Praveen began with a recap on Volumizer and the various vertical markets it plays in along with the large data handling capabilities of SGI's advanced visualization systems. Praveen then covered topics around the Volumizer's integration capabilities with other API toolkits, like Open Inventor and OpenGL Performer and data loaders supported by Volumizer (DICOM, IFL, Plot3d). The next major topic was a preview of Volumizer running on Linux IA32. This was the first example of cross platform support for Volumizer. The first release will provide support for libraries compatible with GCC 3.x and above, and modern graphics cards like ATI Radeon 9000 series and Nvidia GeForce FX and above. Other topics included discussions around shading and computation on the GPU techniques that can be used within Volumizer. Allan Schaffer then presented on Scalability Techniques. He reviewed several ways to increase rendering performance as the number of CPUs and Graphics pipelines increase, and various kinds of decomposition: Application-based, Screen-based, Time-Based, Dataset-Based, Eye-Based, and Pixel-Based. Demos included an example of the integration between OpenGL Performer and OpenGL Volumizer along with a model running on a Linux IA32 machine during and after the BOF. There was also an announcement about the competition to win one of two new Onyx4 UltimateVision systems. The contest is currently on-going and runs through February 1, 2004. For more information on the contest and how to participate, please visit: http://www.sgi.com/visualization/onyx4/challenge We anticipate you'll have many questions about OpenGL Volumizer pricing, availability, licensing model, distribution model, etc. More information will be made available as we get closer to the release date and private feedback is welcome. We would also be pleased to discuss these topics and the details of the new features & capabilities coming in Volumizer 2.6 under a NDA. Please contact your SGI account representative to arrange a briefing; or contact me by e-mail at: ttannert@sgi.com Thank You- Thomas. -------------------------------------------- Thomas Tannert VSG Marketing - SGI 1600 Amphitheatre Pkwy., M/S 526 Mt View, CA 94043 E-mail: ttannert@sgi.com Phone: (650) 933 5237 From info-volumizer-owner@els.sgi.com Sun Nov 9 20:26:48 2003 Received: (from majordom@localhost) by els.sgi.com (8.12.10/8.12.2/els-2.2) id hAA4PjR06319588; Sun, 9 Nov 2003 20:25:45 -0800 (PST) Received: from aerodyne.technion.ac.il (aerodyne.technion.ac.il [132.68.147.2]) by els.sgi.com (8.12.10/8.12.2/els-2.2) with ESMTP id hA9M3ACx6349801 for ; Sun, 9 Nov 2003 14:03:11 -0800 (PST) Received: from [132.68.147.9] ([132.68.147.9]) by aerodyne.technion.ac.il (8.12.9/8.12.8) with ESMTP id hA9LxLA9079518 for ; Sun, 9 Nov 2003 23:59:21 +0200 Date: Sun, 9 Nov 2003 23:58:53 +0200 (IST) From: Yakov Mindelis X-X-Sender: jack@aeone.technion.ac.il To: info-volumizer@els.sgi.com Subject: Re: [info-volumizer] openGL volumizer 2.4 for linux In-Reply-To: <015b01c3a1e7$e8e20630$4a5a0e86@praveenlaptop> Message-ID: MIME-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII Sender: owner-info-volumizer@els.sgi.com Precedence: bulk Reply-To: info-volumizer@els.sgi.com Hi Praveen, > Mahesh and Ivan, > > Yes... this is being 'fixed' with the upcoming Volumizer release - Volumizer > 2.6! We had announced the release of Volumizer on Linux during the Volumizer > BOF meeting at the IEEE Visualization conference in Seattle last week. > > More details will be send to info-volumizer soon. begin(whine); I suppose this will be x86 only. Why not powerpc? Are there exclusively supported gfx card vendors(AFAIR, there were with performer)? end(whine); Jack. From info-volumizer-owner@els.sgi.com Sun Nov 9 20:26:48 2003 Received: (from majordom@localhost) by els.sgi.com (8.12.10/8.12.2/els-2.2) id hAA4PkSH6372962; Sun, 9 Nov 2003 20:25:46 -0800 (PST) Received: from aerodyne.technion.ac.il (aerodyne.technion.ac.il [132.68.147.2]) by els.sgi.com (8.12.10/8.12.2/els-2.2) with ESMTP id hA9LvsCx6291389 for ; Sun, 9 Nov 2003 13:57:55 -0800 (PST) Received: from [132.68.147.9] ([132.68.147.9]) by aerodyne.technion.ac.il (8.12.9/8.12.8) with ESMTP id hA9Ls3A9087008 for ; Sun, 9 Nov 2003 23:54:04 +0200 Date: Sun, 9 Nov 2003 23:53:35 +0200 (IST) From: Yakov Mindelis X-X-Sender: jack@aeone.technion.ac.il To: info-volumizer@els.sgi.com Subject: Re: [info-volumizer] Re: Alpha values in PTLUTShader In-Reply-To: <012501c39d9f$42d879f0$bf069aa3@praveenlaptop> Message-ID: MIME-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII Sender: owner-info-volumizer@els.sgi.com Precedence: bulk Reply-To: info-volumizer@els.sgi.com Hi Praveen, > The output opacity of a rendered vertex depends on the vertex_data value AND > the 'thickness' of the vertex when it is projected. You can look a bit more > into the projected tetrahedra algorithm by Shirley-Tuchman if you want to > get more details. IC, the result is totally unlike in prev version (I now that the APIs ar different :-( ) > If you are using a Linux system as a display, the graphics essentially runs > on the local machine. This means that you might be running the application > on an unsupported graphics system! Volumizer looks at the graphics system > and figures out the type and then does specific optimizations for the > system. If it does not recognize the graphics system, which seems to be the > case with you, it will print out the warning and use a set of default > parameters for the rendering. If things just work for you, then you might > not have anything to worry about :-) Seems to work here... Can't wait till the direct rendering driver is finished (using not common non-x86 hw here). I'll try tomorrow to display it on x86 with direct rendering, just for statistics. Jack. From info-volumizer-owner@els.sgi.com Sun Nov 9 22:00:19 2003 Received: (from majordom@localhost) by els.sgi.com (8.12.10/8.12.2/els-2.2) id hAA5xmfJ6336134; Sun, 9 Nov 2003 21:59:48 -0800 (PST) Received: from sgindia.newdelhi.sgi.com (newdelhi.sgi.com [134.14.90.2]) by els.sgi.com (8.12.10/8.12.2/els-2.2) with ESMTP id hAA5tFCx6312044 for ; Sun, 9 Nov 2003 21:55:17 -0800 (PST) Received: from praveenlaptop ([134.14.90.74]) by sgindia.newdelhi.sgi.com (980427.SGI.8.8.8/980728.SGI.AUTOCF) via SMTP id KAA85048 for ; Mon, 10 Nov 2003 10:08:11 +0530 (IST) Message-ID: <000f01c3a74e$9cdbe410$4a5a0e86@praveenlaptop> From: "Praveen Bhaniramka" To: References: Subject: Re: [info-volumizer] openGL volumizer 2.4 for linux Date: Mon, 10 Nov 2003 11:20:43 +0530 MIME-Version: 1.0 Content-Type: text/plain; charset="iso-8859-1" Content-Transfer-Encoding: 7bit X-Priority: 3 X-MSMail-Priority: Normal X-Mailer: Microsoft Outlook Express 6.00.2800.1106 X-MimeOLE: Produced By Microsoft MimeOLE V6.00.2800.1106 Sender: owner-info-volumizer@els.sgi.com Precedence: bulk Reply-To: info-volumizer@els.sgi.com Yakov, Yes, the 2.6 release will provide support for IA-based Linux systems only. We do not have any plans to provide support for Power PC with this release but please contact the product marketing manager (Thomas Tannert, ttannert@sgi.com) if you have such specific requests. The TMRenderAction requires support for 3D textures as well as support for dependent texture lookups / TLUTs. Since in the PC space there are huge number of vendors with multiple GL extensions for providing the same functionality, we decided to support only a few major ones. The 2.6 release would provide support for - - ATI Radeon 9000 series and above - ATI FireGL series - Nvidia GeForce FX and above If you are using the PTRenderAction only, then the API would work on a wider range of graphics platforms but we do not provide any official support for those platforms. Cheers, Praveen ----- Original Message ----- From: "Yakov Mindelis" To: Sent: Monday, November 10, 2003 3:28 AM Subject: Re: [info-volumizer] openGL volumizer 2.4 for linux > Hi Praveen, > > > > Mahesh and Ivan, > > > > Yes... this is being 'fixed' with the upcoming Volumizer release - Volumizer > > 2.6! We had announced the release of Volumizer on Linux during the Volumizer > > BOF meeting at the IEEE Visualization conference in Seattle last week. > > > > More details will be send to info-volumizer soon. > > begin(whine); > I suppose this will be x86 only. Why not powerpc? > Are there exclusively supported gfx card vendors(AFAIR, there were with > performer)? > end(whine); > > Jack. > > From info-volumizer-owner@els.sgi.com Mon Nov 10 20:18:24 2003 Received: (from majordom@localhost) by els.sgi.com (8.12.10/8.12.2/els-2.2) id hAB4GhhB6372164; Mon, 10 Nov 2003 20:16:43 -0800 (PST) Received: from aerodyne.technion.ac.il (aerodyne.technion.ac.il [132.68.147.2]) by els.sgi.com (8.12.10/8.12.2/els-2.2) with ESMTP id hAAJ1kCx6350064 for ; Mon, 10 Nov 2003 11:01:47 -0800 (PST) Received: from [132.68.147.9] ([132.68.147.9]) by aerodyne.technion.ac.il (8.12.9/8.12.8) with ESMTP id hAAIvrA9061104 for ; Mon, 10 Nov 2003 20:57:53 +0200 Date: Mon, 10 Nov 2003 20:56:44 +0200 (IST) From: Yakov Mindelis X-X-Sender: jack@aeone.technion.ac.il To: info-volumizer@els.sgi.com Subject: Re: [info-volumizer] openGL volumizer 2.4 for linux In-Reply-To: <000f01c3a74e$9cdbe410$4a5a0e86@praveenlaptop> Message-ID: MIME-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII Sender: owner-info-volumizer@els.sgi.com Precedence: bulk Reply-To: info-volumizer@els.sgi.com Hi Praveen, > Yes, the 2.6 release will provide support for IA-based Linux systems only. > We do not have any plans to provide support for Power PC with this release > but please contact the product marketing manager (Thomas Tannert, > ttannert@sgi.com) if you have such specific requests. Big thanx for the info > The TMRenderAction requires support for 3D textures as well as support for > dependent texture lookups / TLUTs. Since in the PC space there are huge > number of vendors with multiple GL extensions for providing the same > functionality, we decided to support only a few major ones. The 2.6 release > would provide support for - > - ATI Radeon 9000 series and above > - ATI FireGL series > - Nvidia GeForce FX and above > If you are using the PTRenderAction only, then the API would work on a wider > range of graphics platforms but we do not provide any official support for > those platforms. Few Qs and thoughts: Isn't it's easier to compile against Mesa which appears in all major distros? And then announce the list of ``entry-level'' cards which can cope with the load? That will simplify the porting to non-x86 market to. For example there is ppc market which includes things such as Macs/A1/Pegasos wich accept agp cards from pc market. Jack From info-volumizer-owner@els.sgi.com Wed Nov 12 21:12:47 2003 Received: (from majordom@localhost) by els.sgi.com (8.12.10/8.12.2/els-2.2) id hAD5AHGn6444590; Wed, 12 Nov 2003 21:10:17 -0800 (PST) Received: from franklin.boulder.nist.gov (franklin-node1.boulder.nist.gov [132.163.128.81]) by els.sgi.com (8.12.10/8.12.2/els-2.2) with ESMTP id hACLi0Cx6439800 for ; Wed, 12 Nov 2003 13:44:02 -0800 (PST) Received: from boulder.nist.gov (canvas.boulder.nist.gov [132.163.129.134]) by franklin.boulder.nist.gov (Switch-2.2.8/Switch-2.2.8) with ESMTP id hACLdaA13508; Wed, 12 Nov 2003 14:39:36 -0700 (MST) Message-ID: <3FB2A88E.9020108@boulder.nist.gov> Date: Wed, 12 Nov 2003 14:39:26 -0700 From: "Adele P. Peskin" Organization: NIST User-Agent: Mozilla/5.0 (X11; U; SunOS sun4u; en-US; rv:1.0.1) Gecko/20020921 Netscape/7.0 X-Accept-Language: en-us, en MIME-Version: 1.0 To: info-volumizer@els.sgi.com, kelso@nist.gov Subject: [info-volumizer] loader question continued Content-Type: text/plain; charset=us-ascii; format=flowed Content-Transfer-Encoding: 7bit X-Scanned-By: milter-spamc/0.6 X-Spam-Flag: NO X-Spam-Status: NO, hits=-105.40 required=5.00 X-Spam-Level: Sender: owner-info-volumizer@els.sgi.com Precedence: bulk Reply-To: info-volumizer@els.sgi.com This is a continuation of a loader question from last week - I am farther along with the problem, but stuck once again... The original question concerned recompiling the *.vol loader code that comes with Volumizer. When I load in a *.vol file using the built-in loader, all is fine - my data loads in without a problem. When I recompile the code, the resulting dso would not load. Praveen pointed out that the loader DSO might not be linking against the Volumizer library - libvz.so. This was the case, and linking in this library, the dso does load, but then crashes with a null vzShape pointer: The error message is: Volumizer Error: Invalid value vzTMRenderAction::manage(): shape is NULL Is there some other library that I am not loading??? Thank you, Adele Peskin From info-volumizer-owner@els.sgi.com Wed Nov 12 22:44:31 2003 Received: (from majordom@localhost) by els.sgi.com (8.12.10/8.12.2/els-2.2) id hAD6i3Lu6403786; Wed, 12 Nov 2003 22:44:03 -0800 (PST) Received: from sgindia.newdelhi.sgi.com (newdelhi.sgi.com [134.14.90.2]) by els.sgi.com (8.12.10/8.12.2/els-2.2) with ESMTP id hAD5xfCx6442784 for ; Wed, 12 Nov 2003 21:59:43 -0800 (PST) Received: from praveenlaptop ([134.14.90.74]) by sgindia.newdelhi.sgi.com (980427.SGI.8.8.8/980728.SGI.AUTOCF) via SMTP id KAA56658 for ; Thu, 13 Nov 2003 10:12:41 +0530 (IST) Message-ID: <018c01c3a9aa$b7b961a0$4a5a0e86@praveenlaptop> From: "Praveen Bhaniramka" To: References: Subject: Re: [info-volumizer] openGL volumizer 2.4 for linux Date: Thu, 13 Nov 2003 11:25:13 +0530 MIME-Version: 1.0 Content-Type: text/plain; charset="iso-8859-1" Content-Transfer-Encoding: 7bit X-Priority: 3 X-MSMail-Priority: Normal X-Mailer: Microsoft Outlook Express 6.00.2800.1106 X-MIMEOLE: Produced By Microsoft MimeOLE V6.00.2800.1106 Sender: owner-info-volumizer@els.sgi.com Precedence: bulk Reply-To: info-volumizer@els.sgi.com Hi again, Linking and running against Mesa (software GL) would technically work. Volume rendering using 3D textures, however, is very demanding from a performance perspective and so any software based implementation would probably be too slow to be useful. Additionally, quite a few of the features like transfer functions, gradient shading etc require certain OpenGL extensions to be present on the system which are querried at run-time by Volumizer. In the absence of those extensions, Volumizer would turn off the respective features. We do not prevent Volumizer from running on unsupported graphics systems (you can give it a try on other systems and let us know what you find!). What is supported simply means that those systems have been well tested and we will be able to fix bugs related to those systems. As for the Power PC question, we are currently planning to support Volumizer on IA-based Linux systems only. Even though a 'basic port' might appear to be trivial, providing a well optimized, tested, and packaged stable product is a lot more work. In any case, please feel free contact Thomas Tannert for such specific feature requests. Cheers, Praveen ----- Original Message ----- From: "Yakov Mindelis" To: Sent: Tuesday, November 11, 2003 12:26 AM Subject: Re: [info-volumizer] openGL volumizer 2.4 for linux > Hi Praveen, > > > Yes, the 2.6 release will provide support for IA-based Linux systems only. > > We do not have any plans to provide support for Power PC with this release > > but please contact the product marketing manager (Thomas Tannert, > > ttannert@sgi.com) if you have such specific requests. > > Big thanx for the info > > > The TMRenderAction requires support for 3D textures as well as support for > > dependent texture lookups / TLUTs. Since in the PC space there are huge > > number of vendors with multiple GL extensions for providing the same > > functionality, we decided to support only a few major ones. The 2.6 release > > would provide support for - > > - ATI Radeon 9000 series and above > > - ATI FireGL series > > - Nvidia GeForce FX and above > > > If you are using the PTRenderAction only, then the API would work on a wider > > range of graphics platforms but we do not provide any official support for > > those platforms. > > Few Qs and thoughts: > Isn't it's easier to compile against Mesa which appears in all major > distros? And then announce the list of ``entry-level'' cards which can > cope with the load? That will simplify the porting to non-x86 market to. For > example there is ppc market which includes things such as Macs/A1/Pegasos > wich accept agp cards from pc market. > > Jack > > From info-volumizer-owner@els.sgi.com Wed Nov 12 22:49:20 2003 Received: (from majordom@localhost) by els.sgi.com (8.12.10/8.12.2/els-2.2) id hAD6nA9G6430117; Wed, 12 Nov 2003 22:49:10 -0800 (PST) Received: from sgindia.newdelhi.sgi.com (newdelhi.sgi.com [134.14.90.2]) by els.sgi.com (8.12.10/8.12.2/els-2.2) with ESMTP id hAD6loCx6365910 for ; Wed, 12 Nov 2003 22:47:56 -0800 (PST) Received: from praveenlaptop ([134.14.90.74]) by sgindia.newdelhi.sgi.com (980427.SGI.8.8.8/980728.SGI.AUTOCF) via SMTP id LAA57363 for ; Thu, 13 Nov 2003 11:01:06 +0530 (IST) Message-ID: <01fa01c3a9b1$78223880$4a5a0e86@praveenlaptop> From: "Praveen Bhaniramka" To: References: <3FB2A88E.9020108@boulder.nist.gov> Subject: Re: [info-volumizer] loader question continued Date: Thu, 13 Nov 2003 12:13:38 +0530 MIME-Version: 1.0 Content-Type: text/plain; charset="iso-8859-1" Content-Transfer-Encoding: 7bit X-Priority: 3 X-MSMail-Priority: Normal X-Mailer: Microsoft Outlook Express 6.00.2800.1106 X-MIMEOLE: Produced By Microsoft MimeOLE V6.00.2800.1106 Sender: owner-info-volumizer@els.sgi.com Precedence: bulk Reply-To: info-volumizer@els.sgi.com Hi Adele, Your loader library is being loaded by Performer properly now, so I am guessing that you are linking against all the required libraries appropriately since probably you do not have any unresolved symbols in your DSO. However, since Volumizer is complaining about a NULL shape, I am guessing that the data files were not loaded appropriately. The loader by default should print some debug messages on the console which should help you debug your application. To increase the verbosity of the debug messages, simply set the VOLUMIZER_DEBUG_LEVEL environment variable to a higher number. A couple of minor comments - 1. The sources for the Volumizer/Performer loader is shipped as part of the Volumizer Development Environment along with the appropriate Makefiles. It might be a good idea for you to download that (sgi.com/software/volumizer) and use it for your development work. 2. The pfvol loader library was renamed to pfvz with the 2.4 release of Volumizer and the code was also modified substantially. It has a new XML file interface which will let you specify the Volumizer data in a cleaner high-level fashion so I would highly recommend using that. Cheers, Praveen ----- Original Message ----- From: "Adele P. Peskin" To: ; Sent: Thursday, November 13, 2003 3:09 AM Subject: [info-volumizer] loader question continued > This is a continuation of a loader question from last > week - I am farther along with the problem, but stuck > once again... > > The original question concerned recompiling the > *.vol loader code that comes with Volumizer. > When I load in a *.vol file using the built-in > loader, all is fine - my data loads in without > a problem. When I recompile the code, the > resulting dso would not load. > > Praveen pointed out that the loader DSO might not be > linking against the Volumizer library - libvz.so. > This was the case, and linking in this library, > the dso does load, but then crashes with a null > vzShape pointer: > > The error message is: > > Volumizer Error: Invalid value > vzTMRenderAction::manage(): shape is NULL > > Is there some other library that I am not > loading??? > > Thank you, > Adele Peskin > From info-volumizer-owner@els.sgi.com Thu Nov 13 10:17:08 2003 Received: (from majordom@localhost) by els.sgi.com (8.12.10/8.12.2/els-2.2) id hADIGTcG6413878; Thu, 13 Nov 2003 10:16:29 -0800 (PST) Received: from franklin.boulder.nist.gov (franklin-node1.boulder.nist.gov [132.163.128.81]) by els.sgi.com (8.12.10/8.12.2/els-2.2) with ESMTP id hADI1mCx6441693 for ; Thu, 13 Nov 2003 10:01:48 -0800 (PST) Received: from boulder.nist.gov (canvas.boulder.nist.gov [132.163.129.134]) by franklin.boulder.nist.gov (Switch-2.2.8/Switch-2.2.8) with ESMTP id hADHvMA06808 for ; Thu, 13 Nov 2003 10:57:22 -0700 (MST) Message-ID: <3FB3C5F7.9050305@boulder.nist.gov> Date: Thu, 13 Nov 2003 10:57:11 -0700 From: "Adele P. Peskin" Organization: NIST User-Agent: Mozilla/5.0 (X11; U; SunOS sun4u; en-US; rv:1.0.1) Gecko/20020921 Netscape/7.0 X-Accept-Language: en-us, en MIME-Version: 1.0 To: info-volumizer@els.sgi.com Subject: Re: [info-volumizer] loader question continued References: <3FB2A88E.9020108@boulder.nist.gov> <01fa01c3a9b1$78223880$4a5a0e86@praveenlaptop> Content-Type: text/plain; charset=us-ascii; format=flowed Content-Transfer-Encoding: 7bit X-Scanned-By: milter-spamc/0.6 X-Spam-Flag: NO X-Spam-Status: NO, hits=-106.30 required=5.00 X-Spam-Level: Sender: owner-info-volumizer@els.sgi.com Precedence: bulk Reply-To: info-volumizer@els.sgi.com We're running Performer 2.5. Is the pfvz loader compatible with that? Thanks, Adele Praveen Bhaniramka wrote: > Hi Adele, > > Your loader library is being loaded by Performer properly now, so I am > guessing that you are linking against all the required libraries > appropriately since probably you do not have any unresolved symbols in your > DSO. However, since Volumizer is complaining about a NULL shape, I am > guessing that the data files were not loaded appropriately. The loader by > default should print some debug messages on the console which should help > you debug your application. To increase the verbosity of the debug messages, > simply set the VOLUMIZER_DEBUG_LEVEL environment variable to a higher > number. > > A couple of minor comments - > 1. The sources for the Volumizer/Performer loader is shipped as part of the > Volumizer Development Environment along with the appropriate Makefiles. It > might be a good idea for you to download that (sgi.com/software/volumizer) > and use it for your development work. > 2. The pfvol loader library was renamed to pfvz with the 2.4 release of > Volumizer and the code was also modified substantially. It has a new XML > file interface which will let you specify the Volumizer data in a cleaner > high-level fashion so I would highly recommend using that. > > Cheers, > Praveen > > ----- Original Message ----- > From: "Adele P. Peskin" > To: ; > Sent: Thursday, November 13, 2003 3:09 AM > Subject: [info-volumizer] loader question continued > > > >>This is a continuation of a loader question from last >>week - I am farther along with the problem, but stuck >>once again... >> >>The original question concerned recompiling the >>*.vol loader code that comes with Volumizer. >>When I load in a *.vol file using the built-in >>loader, all is fine - my data loads in without >>a problem. When I recompile the code, the >>resulting dso would not load. >> >>Praveen pointed out that the loader DSO might not be >>linking against the Volumizer library - libvz.so. >>This was the case, and linking in this library, >>the dso does load, but then crashes with a null >>vzShape pointer: >> >>The error message is: >> >>Volumizer Error: Invalid value >>vzTMRenderAction::manage(): shape is NULL >> >>Is there some other library that I am not >>loading??? >> >>Thank you, >>Adele Peskin >> > > > > > From info-volumizer-owner@els.sgi.com Thu Nov 13 12:21:46 2003 Received: (from majordom@localhost) by els.sgi.com (8.12.10/8.12.2/els-2.2) id hADKKBv32447400; Thu, 13 Nov 2003 12:20:11 -0800 (PST) Received: from sgindia.newdelhi.sgi.com (newdelhi.sgi.com [134.14.90.2]) by els.sgi.com (8.12.10/8.12.2/els-2.2) with ESMTP id hADKC6Cx6453022 for ; Thu, 13 Nov 2003 12:12:07 -0800 (PST) Received: from praveenlaptop (mtv-vpn-sw-corp-0-86.corp.sgi.com [134.15.0.86]) by sgindia.newdelhi.sgi.com (980427.SGI.8.8.8/980728.SGI.AUTOCF) via SMTP id AAA65616 for ; Fri, 14 Nov 2003 00:24:57 +0530 (IST) Message-ID: <00e501c3aa21$cb2c0bd0$6401a8c0@praveenlaptop> From: "Praveen Bhaniramka" To: References: <3FB2A88E.9020108@boulder.nist.gov> <01fa01c3a9b1$78223880$4a5a0e86@praveenlaptop> <3FB3C5F7.9050305@boulder.nist.gov> Subject: Re: [info-volumizer] loader question continued Date: Fri, 14 Nov 2003 01:37:27 +0530 MIME-Version: 1.0 Content-Type: text/plain; charset="iso-8859-1" Content-Transfer-Encoding: 7bit X-Priority: 3 X-MSMail-Priority: Normal X-Mailer: Microsoft Outlook Express 6.00.2800.1106 X-MimeOLE: Produced By Microsoft MimeOLE V6.00.2800.1106 Sender: owner-info-volumizer@els.sgi.com Precedence: bulk Reply-To: info-volumizer@els.sgi.com Hi Adele, Yes.. the new loader should be compatible with Performer 2.5 since the pfGeoSetCB class that the loader uses was added with Performer 2.5. More recently, I have been testing mainly with Performer 3.0 though. Let me know if you have any issues getting it to work. - Praveen ----- Original Message ----- From: "Adele P. Peskin" To: Sent: Thursday, November 13, 2003 11:27 PM Subject: Re: [info-volumizer] loader question continued > We're running Performer 2.5. Is the pfvz loader > compatible with that? > Thanks, Adele > > Praveen Bhaniramka wrote: > > Hi Adele, > > > > Your loader library is being loaded by Performer properly now, so I am > > guessing that you are linking against all the required libraries > > appropriately since probably you do not have any unresolved symbols in your > > DSO. However, since Volumizer is complaining about a NULL shape, I am > > guessing that the data files were not loaded appropriately. The loader by > > default should print some debug messages on the console which should help > > you debug your application. To increase the verbosity of the debug messages, > > simply set the VOLUMIZER_DEBUG_LEVEL environment variable to a higher > > number. > > > > A couple of minor comments - > > 1. The sources for the Volumizer/Performer loader is shipped as part of the > > Volumizer Development Environment along with the appropriate Makefiles. It > > might be a good idea for you to download that (sgi.com/software/volumizer) > > and use it for your development work. > > 2. The pfvol loader library was renamed to pfvz with the 2.4 release of > > Volumizer and the code was also modified substantially. It has a new XML > > file interface which will let you specify the Volumizer data in a cleaner > > high-level fashion so I would highly recommend using that. > > > > Cheers, > > Praveen > > > > ----- Original Message ----- > > From: "Adele P. Peskin" > > To: ; > > Sent: Thursday, November 13, 2003 3:09 AM > > Subject: [info-volumizer] loader question continued > > > > > > > >>This is a continuation of a loader question from last > >>week - I am farther along with the problem, but stuck > >>once again... > >> > >>The original question concerned recompiling the > >>*.vol loader code that comes with Volumizer. > >>When I load in a *.vol file using the built-in > >>loader, all is fine - my data loads in without > >>a problem. When I recompile the code, the > >>resulting dso would not load. > >> > >>Praveen pointed out that the loader DSO might not be > >>linking against the Volumizer library - libvz.so. > >>This was the case, and linking in this library, > >>the dso does load, but then crashes with a null > >>vzShape pointer: > >> > >>The error message is: > >> > >>Volumizer Error: Invalid value > >>vzTMRenderAction::manage(): shape is NULL > >> > >>Is there some other library that I am not > >>loading??? > >> > >>Thank you, > >>Adele Peskin > >> > > > > > > > > > > > > From info-volumizer-owner@els.sgi.com Sun Nov 16 10:11:02 2003 Received: (from majordom@localhost) by els.sgi.com (8.12.10/8.12.2/els-2.2) id hAGI9pQR6474864; Sun, 16 Nov 2003 10:09:51 -0800 (PST) Received: from aerodyne.technion.ac.il (aerodyne.technion.ac.il [132.68.147.2]) by els.sgi.com (8.12.10/8.12.2/els-2.2) with ESMTP id hAFLlMCx6473343 for ; Sat, 15 Nov 2003 13:47:23 -0800 (PST) Received: from [132.68.147.9] ([132.68.147.9]) by aerodyne.technion.ac.il (8.12.9/8.12.8) with ESMTP id hAFLhGA9081652 for ; Sat, 15 Nov 2003 23:43:16 +0200 Date: Sat, 15 Nov 2003 23:39:20 +0200 (IST) From: Yakov Mindelis X-X-Sender: jack@aeone.technion.ac.il To: info-volumizer@els.sgi.com Subject: [info-volumizer] shapeset troubles Message-ID: MIME-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII Sender: owner-info-volumizer@els.sgi.com Precedence: bulk Reply-To: info-volumizer@els.sgi.com Hi all, I'm experiencing troubles implementing shapeSets here. the init of shapeSet is as follows: //remover shapes if(m_shapeSetTexture) { for(i = m_shapeSetTexture->getNumShapes()-1; i>0; i--) m_shapeSetTexture->remove(m_shapeTexture[i-1]); } //deletes existing, creates new array classes of: if(m_vtxLUTDataTexture.Size()) { //unref() all for(i = 0; i < m_LUTTexture.Size(); i++) while(m_LUTTexture[i]) m_LUTTexture[i]->unref(); for(i = 0; i < m_shapeTexture.Size(); i++) while(m_shapeTexture[i]) m_shapeTexture[i]->unref(); for(i = 0; i < m_appearanceTexture.Size(); i++) while(m_appearanceTexture[i]) m_appearanceTexture[i]->unref(); } m_vtxLUTDataTexture.renew (gridCells * 2 * 4);//floats m_LUTTexture.renew(gridCells);//(vzParameterLookupTable *) m_shapeTexture.renew(gridCells);//vzShapes m_appearanceTexture.renew(gridCells);//(vzAppearance *) for(i, j, k ...) { off = something(i, j, k); r = something(i, j, k); g = something(i, j, k); b = something(i, j, k); a = something(i, j, k); //puts 8 values in m_vtxLUTDataTexture data array in 8*off offset and // uses the first address as dataptr to create m_LUTTexture[off] // vzParameterLookupTable createLUT(m_vtxLUTDataTexture, r, g, b, a, off); //init appearance m_appearanceTexture[off] = new vzAppearance(m_shader); m_shader->unref(); m_appearanceTexture[off]->setParameter("vertex_data", m_vtxData); m_vtxData->unref(); m_appearanceTexture[off]->setParameter("lookup_table", m_LUTTexture[off]); m_LUTTexture[off]->unref(); //init shape m_shapeTexture[off] = new vzShape(m_hexaMeshTexture[off], m_appearanceTexture[off]); m_hexaMeshTexture[off]->unref(); m_appearanceTexture[off]->unref(); //add the shape to the set m_shapeSetTexture -> add(m_shapeTexture[off]); m_shapeTexture[off] -> unref(); } } The "unref all" stuff seems wrong to me. When I reenter the init for geometry init there's same reallocation for array of hexaMeshes (unref with, ref count shows huge value in debugger. Seems like ref count is unsigned, but when ref count reaches 0 the object should be destroyed. How do I know the reference count? What's the elegant(and correct) way to destroy the objects manually? TIA, Jack. From info-volumizer-owner@els.sgi.com Sun Nov 16 22:17:40 2003 Received: (from majordom@localhost) by els.sgi.com (8.12.10/8.12.2/els-2.2) id hAH6GtZf6452387; Sun, 16 Nov 2003 22:16:55 -0800 (PST) Received: from sgindia.newdelhi.sgi.com (newdelhi.sgi.com [134.14.90.2]) by els.sgi.com (8.12.10/8.12.2/els-2.2) with ESMTP id hAH6AdCx6492435 for ; Sun, 16 Nov 2003 22:10:40 -0800 (PST) Received: from praveenlaptop ([134.14.90.74]) by sgindia.newdelhi.sgi.com (980427.SGI.8.8.8/980728.SGI.AUTOCF) via SMTP id KAA42401 for ; Mon, 17 Nov 2003 10:23:34 +0530 (IST) Message-ID: <010a01c3acd0$e6d01f00$6401a8c0@praveenlaptop> From: "Praveen Bhaniramka" To: References: Subject: Re: [info-volumizer] shapeset troubles Date: Mon, 17 Nov 2003 11:36:11 +0530 MIME-Version: 1.0 Content-Type: text/plain; charset="iso-8859-1" Content-Transfer-Encoding: 7bit X-Priority: 3 X-MSMail-Priority: Normal X-Mailer: Microsoft Outlook Express 6.00.2800.1106 X-MimeOLE: Produced By Microsoft MimeOLE V6.00.2800.1106 Sender: owner-info-volumizer@els.sgi.com Precedence: bulk Reply-To: info-volumizer@els.sgi.com Hi Yakov, It is impossible for me to tell what you are trying to do in that code snippet without knowing what the various variables there are supposed to be :( In general, the way ref-counting is implemented in Volumizer is explained very well in the programming guide. Important points to remember : 1. Every object, when created, has a ref count of 1 by default 2. Every object is deleted _automatically_ when the ref count reaches 0 3. Every 'entity' which needs to keep around a copy of an object should call ref() on the object when it gets 'access' to it first and the unref() when it does not need it anymore 4. Every Volumizer object performs step 3 consistently througout the API. e.g. when you call vzApperance::setParameter(), the appearance calls ref() on the parameter. An easier way to debug reference counts in your case would be to use a small array of 1 or 2 objects and then set the env variable VOLUMIZER_DEBUG_LEVEL to 4. Each ref/unref call, whether called from within Volumizer or outside Volumizer, will print out a debug message indicating the object's current reference count. If the reference count is printed as a ridiculously large number, it usually means that you called unref() on an object which was already deleted, i.e. its reference count had already reached 0. The sample code tries to demonstrate the 'usual good practices' when dealing with reference counts. For example, vzBlock *geometry = new vzBlock(); // block's reference count = 1 vzAppearance *appearance = new vzAppearance(shader); // appearance's ref count = 1 vzShape *shape = new vzShape(geometry, appearance); // ref count of shape = 1, geometry and appearance = 2 geometry->unref(); // geometry = 1 appearance->unref(); // appearance = 1 At the end of the code, shape, geometry and appearance, all have a ref count of 1. Now, when you are done with the shape, you can call shape->unref(); // shape ref = 0, shape gets deleted The above causes the shape to be deleted, and shape now calls unref() on its components, namely the geometry and appearance. This reduces the shape counts of the two to 0 and they get deleted as well and everyone is happy! hth, Praveen ----- Original Message ----- From: "Yakov Mindelis" To: Sent: Sunday, November 16, 2003 3:09 AM Subject: [info-volumizer] shapeset troubles > Hi all, > > I'm experiencing troubles implementing shapeSets here. > > the init of shapeSet is as follows: > > //remover shapes > if(m_shapeSetTexture) > { > for(i = m_shapeSetTexture->getNumShapes()-1; i>0; i--) > m_shapeSetTexture->remove(m_shapeTexture[i-1]); > } > //deletes existing, creates new array classes of: > if(m_vtxLUTDataTexture.Size()) > { > //unref() all > for(i = 0; i < m_LUTTexture.Size(); i++) > while(m_LUTTexture[i]) > m_LUTTexture[i]->unref(); > for(i = 0; i < m_shapeTexture.Size(); i++) > while(m_shapeTexture[i]) > m_shapeTexture[i]->unref(); > for(i = 0; i < m_appearanceTexture.Size(); i++) > while(m_appearanceTexture[i]) > m_appearanceTexture[i]->unref(); > } > m_vtxLUTDataTexture.renew (gridCells * 2 * 4);//floats > m_LUTTexture.renew(gridCells);//(vzParameterLookupTable *) > m_shapeTexture.renew(gridCells);//vzShapes > m_appearanceTexture.renew(gridCells);//(vzAppearance *) > > for(i, j, k ...) > { > off = something(i, j, k); > r = something(i, j, k); > g = something(i, j, k); > b = something(i, j, k); > a = something(i, j, k); > > //puts 8 values in m_vtxLUTDataTexture data array in 8*off offset and > // uses the first address as dataptr to create m_LUTTexture[off] > // vzParameterLookupTable > createLUT(m_vtxLUTDataTexture, r, g, b, a, off); > //init appearance > m_appearanceTexture[off] = new vzAppearance(m_shader); > m_shader->unref(); > > m_appearanceTexture[off]->setParameter("vertex_data", > m_vtxData); > m_vtxData->unref(); > > m_appearanceTexture[off]->setParameter("lookup_table", > m_LUTTexture[off]); > m_LUTTexture[off]->unref(); > > //init shape > m_shapeTexture[off] = new vzShape(m_hexaMeshTexture[off], > m_appearanceTexture[off]); > m_hexaMeshTexture[off]->unref(); > m_appearanceTexture[off]->unref(); > //add the shape to the set > m_shapeSetTexture -> add(m_shapeTexture[off]); > m_shapeTexture[off] -> unref(); > } > } > > The "unref all" stuff seems wrong to me. When I reenter the init for > geometry init there's same reallocation for array of hexaMeshes (unref > with, ref count shows huge value in debugger. Seems like ref count is > unsigned, but when ref count reaches 0 the object should be destroyed. > How do I know the reference count? What's the elegant(and correct) way to > destroy the objects manually? > > TIA, > Jack. > From info-volumizer-owner@els.sgi.com Sun Nov 23 21:44:13 2003 Received: (from majordom@localhost) by els.sgi.com (8.12.10/8.12.2/els-2.2) id hAO5h1nM091393; Sun, 23 Nov 2003 21:43:01 -0800 (PST) Received: from aerodyne.technion.ac.il (aerodyne.technion.ac.il [132.68.147.2]) by els.sgi.com (8.12.10/8.12.2/els-2.2) with ESMTP id hALCOStp073052 for ; Fri, 21 Nov 2003 04:24:30 -0800 (PST) Received: from [132.68.147.9] ([132.68.147.9]) by aerodyne.technion.ac.il (8.12.9/8.12.8) with ESMTP id hALCKAA9071240 for ; Fri, 21 Nov 2003 14:20:10 +0200 Date: Fri, 21 Nov 2003 14:18:11 +0200 (IST) From: Yakov Mindelis X-X-Sender: jack@aeone.technion.ac.il To: info-volumizer@els.sgi.com Subject: Re: [info-volumizer] shapeset troubles In-Reply-To: <010a01c3acd0$e6d01f00$6401a8c0@praveenlaptop> Message-ID: MIME-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII Sender: owner-info-volumizer@els.sgi.com Precedence: bulk Reply-To: info-volumizer@els.sgi.com On Mon, 17 Nov 2003, Praveen Bhaniramka wrote: > It is impossible for me to tell what you are trying to do in that code > snippet without knowing what the various variables there are supposed to be > :( > > 4. Every Volumizer object performs step 3 consistently througout the API. > e.g. when you call vzApperance::setParameter(), the appearance calls ref() > on the parameter. Ok, I see, when I want all I need is ref() the geometry once manually, remove shapes form shapeset, delete shapes and they'll unref() appearances and gometries. Appearances will die, geometries will remain so I can reuse them. And it _works_. Big thanks! Now going to explore paging... Jack. From info-volumizer-owner@els.sgi.com Sun Nov 23 21:44:13 2003 Received: (from majordom@localhost) by els.sgi.com (8.12.10/8.12.2/els-2.2) id hAO5h0lH111307; Sun, 23 Nov 2003 21:43:00 -0800 (PST) Received: from aerodyne.technion.ac.il (aerodyne.technion.ac.il [132.68.147.2]) by els.sgi.com (8.12.10/8.12.2/els-2.2) with ESMTP id hALDEGtp086242 for ; Fri, 21 Nov 2003 05:14:17 -0800 (PST) Received: from [132.68.147.9] ([132.68.147.9]) by aerodyne.technion.ac.il (8.12.9/8.12.8) with ESMTP id hALD9wA9076204 for ; Fri, 21 Nov 2003 15:09:58 +0200 Date: Fri, 21 Nov 2003 15:09:32 +0200 (IST) From: Yakov Mindelis X-X-Sender: jack@aeone.technion.ac.il To: info-volumizer@els.sgi.com Subject: [info-volumizer] memory Message-ID: MIME-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII Sender: owner-info-volumizer@els.sgi.com Precedence: bulk Reply-To: info-volumizer@els.sgi.com Hi Praveen, There's following phenomena: 1. when I display graphics locally, the following happens: addition of 6000 shapes to the shapeset which is being drawn causes the memory size of the programm to increasey by approx. 230 Mb 2. When displaying remotely, the size inreased only by several Mb, while on displaying machine (Linux) I see increase in size of X, and the increase is much smaller than the one described in (1). What is the main difference in memory allocation meachanisms in these two cases? Thanks in advance, Jack. From info-volumizer-owner@els.sgi.com Mon Nov 24 23:03:34 2003 Received: (from majordom@localhost) by els.sgi.com (8.12.10/8.12.2/els-2.2) id hAP6xawT135324; Mon, 24 Nov 2003 22:59:36 -0800 (PST) Received: from sgindia.newdelhi.sgi.com (newdelhi.sgi.com [134.14.90.2]) by els.sgi.com (8.12.10/8.12.2/els-2.2) with ESMTP id hAP6Pctp134723 for ; Mon, 24 Nov 2003 22:25:40 -0800 (PST) Received: from praveenlaptop ([134.14.90.74]) by sgindia.newdelhi.sgi.com (980427.SGI.8.8.8/980728.SGI.AUTOCF) via SMTP id KAA39437 for ; Tue, 25 Nov 2003 10:38:28 +0530 (IST) Message-ID: <011001c3b3e5$7c95a350$4a5a0e86@praveenlaptop> From: "Praveen Bhaniramka" To: References: Subject: Re: [info-volumizer] memory Date: Wed, 26 Nov 2003 11:51:10 +0530 MIME-Version: 1.0 Content-Type: text/plain; charset="iso-8859-1" Content-Transfer-Encoding: 7bit X-Priority: 3 X-MSMail-Priority: Normal X-Mailer: Microsoft Outlook Express 6.00.2800.1106 X-MimeOLE: Produced By Microsoft MimeOLE V6.00.2800.1106 Sender: owner-info-volumizer@els.sgi.com Precedence: bulk Reply-To: info-volumizer@els.sgi.com Yakov, In the first case, 'everything' happens locally. Meaning that the application as well as the graphics are running locally. In the second case however, the application is running locally but the graphics is running remotely. Depending on what exactly your shapes contain, the OpenGL objects will take up some memory. The shapes themselves will be stored locally in both the cases. Cheers, Praveen ----- Original Message ----- From: "Yakov Mindelis" To: Sent: Friday, November 21, 2003 6:39 PM Subject: [info-volumizer] memory > Hi Praveen, > > There's following phenomena: > 1. when I display graphics locally, the following happens: > addition of 6000 shapes to the shapeset which is being drawn causes the > memory size of the programm to increasey by approx. 230 Mb > > 2. When displaying remotely, the size inreased only by several Mb, while > on displaying machine (Linux) I see increase in size of X, and the > increase is much smaller than the one described in (1). > > What is the main difference in memory allocation meachanisms in these two > cases? > > Thanks in advance, > Jack. > > > From info-volumizer-owner@els.sgi.com Tue Nov 25 21:36:13 2003 Received: (from majordom@localhost) by els.sgi.com (8.12.10/8.12.2/els-2.2) id hAQ5Z3XK158710; Tue, 25 Nov 2003 21:35:03 -0800 (PST) Received: from aerodyne.technion.ac.il (aerodyne.technion.ac.il [132.68.147.2]) by els.sgi.com (8.12.10/8.12.2/els-2.2) with ESMTP id hAPMQstp152518 for ; Tue, 25 Nov 2003 14:26:57 -0800 (PST) Received: from [132.68.147.9] ([132.68.147.9]) by aerodyne.technion.ac.il (8.12.9/8.12.8) with ESMTP id hAPMLvA9063856 for ; Wed, 26 Nov 2003 00:22:25 +0200 Date: Wed, 26 Nov 2003 00:21:31 +0200 (IST) From: Yakov Mindelis X-X-Sender: jack@aeone.technion.ac.il To: info-volumizer@els.sgi.com Subject: Re: [info-volumizer] memory In-Reply-To: <011001c3b3e5$7c95a350$4a5a0e86@praveenlaptop> Message-ID: MIME-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII Sender: owner-info-volumizer@els.sgi.com Precedence: bulk Reply-To: info-volumizer@els.sgi.com On Wed, 26 Nov 2003, Praveen Bhaniramka wrote: > Yakov, > > In the first case, 'everything' happens locally. Meaning that the > application as well as the graphics are running locally. In the second case > however, the application is running locally but the graphics is running > remotely. Depending on what exactly your shapes contain, the OpenGL objects > will take up some memory. The shapes themselves will be stored locally in > both the cases. So I can't get why ~100000 single-cell hexahedral shapes easily take ~900Mb physical ram and the the app must be killed to prevent insane swapping when displaying locally. When doing it remotely, the application hardly consumes 350Mb of physical ram, but displaying is dead slow (>30 minutes per frame!). It will force me to travel to another city to try onyx with 3Gb of ram :-( Jack. From info-volumizer-owner@els.sgi.com Wed Nov 26 21:59:16 2003 Received: (from majordom@localhost) by els.sgi.com (8.12.10/8.12.2/els-2.2) id hAR5pjZU004545; Wed, 26 Nov 2003 21:51:45 -0800 (PST) Received: from sgindia.newdelhi.sgi.com (newdelhi.sgi.com [134.14.90.2]) by els.sgi.com (8.12.10/8.12.2/els-2.2) with ESMTP id hAR5iFds004520 for ; Wed, 26 Nov 2003 21:44:23 -0800 (PST) Received: from praveenlaptop ([134.14.90.74]) by sgindia.newdelhi.sgi.com (980427.SGI.8.8.8/980728.SGI.AUTOCF) via SMTP id JAA27060 for ; Thu, 27 Nov 2003 09:56:57 +0530 (IST) Message-ID: <038e01c3b572$08fe7960$4a5a0e86@praveenlaptop> From: "Praveen Bhaniramka" To: References: Subject: Re: [info-volumizer] memory Date: Fri, 28 Nov 2003 11:09:08 +0530 MIME-Version: 1.0 Content-Type: text/plain; charset="iso-8859-1" Content-Transfer-Encoding: 7bit X-Priority: 3 X-MSMail-Priority: Normal X-Mailer: Microsoft Outlook Express 6.00.2800.1106 X-MIMEOLE: Produced By Microsoft MimeOLE V6.00.2800.1106 Sender: owner-info-volumizer@els.sgi.com Precedence: bulk Reply-To: info-volumizer@els.sgi.com Yakov, Hmm... it seems that the the PTRenderAction is not being very smart about its GL state management somewhere. That would explain the increase in graphics related memory requirements. Let me see what the issue might be. Thanks. Praveen ----- Original Message ----- From: "Yakov Mindelis" To: Sent: Wednesday, November 26, 2003 3:51 AM Subject: Re: [info-volumizer] memory > On Wed, 26 Nov 2003, Praveen Bhaniramka wrote: > > > Yakov, > > > > In the first case, 'everything' happens locally. Meaning that the > > application as well as the graphics are running locally. In the second case > > however, the application is running locally but the graphics is running > > remotely. Depending on what exactly your shapes contain, the OpenGL objects > > will take up some memory. The shapes themselves will be stored locally in > > both the cases. > > So I can't get why ~100000 single-cell hexahedral shapes easily take > ~900Mb physical ram and the the app must be killed to prevent insane > swapping when displaying locally. When doing it remotely, the application > hardly consumes 350Mb of physical ram, but displaying is dead slow (>30 > minutes per frame!). It will force me to travel to another city to try > onyx with 3Gb of ram :-( > > Jack. >