Re: [osg-users] Drawing complex labels of 3D points in screen space
Hi, On 21/10/10 18:26, Yurii Monakov wrote: Hi All After some googling I've discovered an AutoTransform node )) When setting it's auto-rotate mode to ROTATE_TO_SCREEN and setAutoScaleToScreen(true) it gives me exactly what I want. The only problem is that after looking at AutoTransform node I noticed that it does a lot of computation inside. The only thing I need is to push root's MVPW matrix to some part of the graph (Transform nodes, which are children of the screen-spaced node) to update their positions properly. Is there any other way to do this? not that I know of, but I think you've now discovered enough to create your own node, either based on AutoTransform (but for already screen-spaced geometry) or your own. jp Thank you! Yurii ___ osg-users mailing list osg-users@lists.openscenegraph.org http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org -- This message is subject to the CSIR's copyright terms and conditions, e-mail legal notice, and implemented Open Document Format (ODF) standard. The full disclaimer details can be found at http://www.csir.co.za/disclaimer.html. This message has been scanned for viruses and dangerous content by MailScanner, and is believed to be clean. MailScanner thanks Transtec Computers for their support. ___ osg-users mailing list osg-users@lists.openscenegraph.org http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org
Re: [osg-users] VPB problem with large database...
Hi, you can check in the tasks folder to see exactly the osgdem command(s) that would be executed by vpb. Try execute one of these manually and see if it does not report any useful errors. jp On 21/10/10 20:31, Tueller, Shayne R Civ USAF AFMC 519 SMXS/MXDEC wrote: All, I’m seeing some issues with VPB when attempting to build a very large database. My command is the following: vpbmaster –TERRAIN –PagedLOD –geocentric –whole-globe –t wholeearthtexture –t detailtexture –d dted –l 24 –o database.ive What I’m seeing is that vpbmaster creates the tasks and logs directories followed by the scheduling of tasks. During the scheduling of tasks, I see the threads “osgdem…” created on the machine and then die a few seconds later. vpbmaster then just waits for something to happen and eventually goes to sleep. I’m guessing that this is happening because the threads died for some reason and vpbmaster just sits there waiting for them to come back. I understand the VPB is able to handle the building of very large databases but it seems to be choking in our particular case. If I run the same command above on a smaller set of data, things run just fine. Our system has 16 processors with ample disk and memory space to handle what is needed. It also runs Fadora core 13. Has anyone else seen this issue? Is there a flag or setting that I need to use to build a large database? Any input would be appreciated… -Shayne ___ osg-users mailing list osg-users@lists.openscenegraph.org http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org -- This message is subject to the CSIR's copyright terms and conditions, e-mail legal notice, and implemented Open Document Format (ODF) standard. The full disclaimer details can be found at http://www.csir.co.za/disclaimer.html. This message has been scanned for viruses and dangerous content by MailScanner, and is believed to be clean. MailScanner thanks Transtec Computers for their support. ___ osg-users mailing list osg-users@lists.openscenegraph.org http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org
Re: [osg-users] Terrain geometry
Hi, osgTDS works not with VPB databases in the TERRAIN format, so its obsolete. I'm interested in such a terrain editing process too, maybe we can develop a solution together! ... Thank you! Cheers, Torben -- Read this topic online here: http://forum.openscenegraph.org/viewtopic.php?p=32963#32963 ___ osg-users mailing list osg-users@lists.openscenegraph.org http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org
[osg-users] color not the good one -_-
Helo ,
i need an image 200-0-255 for special use. I make the following code:
Code:
im_ = new Image;
im_.get()-allocateImage(WIDTH_CAM , HEIGHT_CAM, 1 , GL_RGB, GL_UNSIGNED_BYTE);
unsigned int r = 200;
unsigned int g = 0;
unsigned int b = 255;
std::cout Color for background : r g b
std::endl;
for(int i = 0 ; i HEIGHT_CAM*WIDTH_CAM ; i++)
{
memcpy(im_.get()-data()+i*3 , r, 1);
memcpy(im_.get()-data()+i*3+1, g, 1);
memcpy(im_.get()-data()+i*3+2, b, 1);
}
texture_.get()-setImage (im_.get());
texture_.get()-setFilter(osg::Texture2D::MIN_FILTER,osg::Texture2D::LINEAR_MIPMAP_LINEAR);
texture_.get()-setFilter(osg::Texture2D::MAG_FILTER,osg::Texture2D::LINEAR_MIPMAP_LINEAR);
stateset = geom.get()-getOrCreateStateSet();
stateset.get()-setTextureAttributeAndModes(0, texture_.get(),
osg::StateAttribute::ON);
stateset.get()-setMode(GL_BLEND,osg::StateAttribute::ON);
stateset.get()-setRenderingHint(osg::StateSet::OPAQUE_BIN);
geode.get()-getOrCreateStateSet()-setMode(GL_LIGHTING, false);
And result is 160-0-204
somebody has a clue?
thanks you !
(http://www.hordes.fr?ref=litllechicken)
--
Read this topic online here:
http://forum.openscenegraph.org/viewtopic.php?p=32964#32964
___
osg-users mailing list
osg-users@lists.openscenegraph.org
http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org
[osg-users] How to implement a fuel tank
Hi All: I want to use OSG to implement a irregular fuel tank, I can see the oil in this tank, oil can be changed by configuration parameters, you can rotate the tank, mobile, how should I achieve? ___ osg-users mailing list osg-users@lists.openscenegraph.org http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org
Re: [osg-users] OSG greensceen
Hi, you should have a background with texture witch is the film (copy image from camera each frame) it goes on 110 ms latence with out optimisation. You place other item from osg above this background and that will work. If all is done in osg you dont need your green screen. (http://www.hordes.fr?ref=litllechicken) -- Read this topic online here: http://forum.openscenegraph.org/viewtopic.php?p=32967#32967 ___ osg-users mailing list osg-users@lists.openscenegraph.org http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org
Re: [osg-users] color not the good one -_-
On 22/10/10 18:41 , Anthony Face wrote:
i need an image 200-0-255 for special use. I make the following code:
...
im_ = new Image;
im_.get()-allocateImage(WIDTH_CAM , HEIGHT_CAM, 1 , GL_RGB,
GL_UNSIGNED_BYTE);
unsigned int r = 200;
unsigned int g = 0;
unsigned int b = 255;
std::cout Color for background : r g b
std::endl;
for(int i = 0 ; i HEIGHT_CAM*WIDTH_CAM ; i++)
{
memcpy(im_.get()-data()+i*3 , r, 1);
memcpy(im_.get()-data()+i*3+1, g, 1);
memcpy(im_.get()-data()+i*3+2, b, 1);
}
...
stateset.get()-setTextureAttributeAndModes(0, texture_.get(),
osg::StateAttribute::ON);
stateset.get()-setMode(GL_BLEND,osg::StateAttribute::ON);
stateset.get()-setRenderingHint(osg::StateSet::OPAQUE_BIN);
geode.get()-getOrCreateStateSet()-setMode(GL_LIGHTING, false);
And result is 160-0-204
What do you mean by 'result'? Are you color picking the rendered geometry?
I'd make sure the texture environment mode is 'DECAL', otherwise it's modified
by the
color of the underlying geometry.
Also, why do you turn blending on but then put the object into the opaque bin?
It's
usually blending-on/transparent or blending-off/opaque. Just saying :-)
Cheers,
/ulrich
___
osg-users mailing list
osg-users@lists.openscenegraph.org
http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org
Re: [osg-users] How to implement a fuel tank
HI Juan, On Fri, Oct 22, 2010 at 8:42 AM, Juan forsakenwarl...@gmail.com wrote: I want to use OSG to implement a irregular fuel tank, I can see the oil in this tank, oil can be changed by configuration parameters, you can rotate the tank, mobile, how should I achieve? This is rather opened ended question and could span a multitude of topics. So.. lets start narrowing it down. Are you needing to do some kind of physical simulation of the oil? This really isn't an OSG topic but it's an important to be able to know what is appropriate to ask from the OSG community, you'll want a specialist that knows about fluid physics simulation. As for rotating a tank, are we just talking about moving it as a solid body, if so use a osg::Transform nodes such as MatrixTransform or PositionAttitudeTrasnform. See osganimation example. Robert. ___ osg-users mailing list osg-users@lists.openscenegraph.org http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org
Re: [osg-users] OSG on IPhone
Hi Ulrich, Am 22.10.10 04:32, schrieb Ulrich Hertlein: This should work, too: git clone git://github.com/stmh/osg.git git pull origin iphone This is giving me a heap of conflicts (on a clean checkout/clone) along the lines of From git://github.com/stmh/osg * branchiphone - FETCH_HEAD CONFLICT (rename/delete): Rename IPhone_Project/osgPlugins/imageio/ReaderWriterImageIO.cpp-src/osgPlugins/imageio/ReaderWriterImageIO_IPhone.cpp in b79883a69545f82a573ec8aaa350262b3dedf4b1 and deleted in HEAD CONFLICT (delete/modify): IPhone_Project/OSGIPhone.xcodeproj/hogbox.mode1v3 deleted in HEAD and modified in b79883a69545f82a573ec8aaa350262b3dedf4b1. Version b79883a69545f82a573ec8aaa350262b3dedf4b1 of IPhone_Project/OSGIPhone.xcodeproj/hogbox.mode1v3 left in tree. CONFLICT (delete/modify): IPhone_Project/OSGIPhone.xcodeproj/hogbox.pbxuser deleted in HEAD and modified in b79883a69545f82a573ec8aaa350262b3dedf4b1. Version b79883a69545f82a573ec8aaa350262b3dedf4b1 of IPhone_Project/OSGIPhone.xcodeproj/hogbox.pbxuser left in tree. ... The 'git clone, git branch, git checkout' that Florian described above works however... (I absolutely haven't got my head around git at all yet, so this might be a case of user error.) Or am I using the wrong repository? No the repo is fine. My experience with git is also limited and I did some mistakes when checking stuff into the master-branch in the past. (I did a git push origin master from my iphone-branch which pushed the iphone-related stuff into the master-branch. After I realized my fault, I couldn't revert the change, so I deleted the iPhone-specific stuff from this branch. I am still looking for a way to reset the master branch to a specific version on github.) Be sure to pull into the right branch e.g. git clone git://github.com/stmh/osg.git cd osg #switch to the iphone-branch git checkout iphone #pull new changes git pull origin iphone cheers, Stephan ___ osg-users mailing list osg-users@lists.openscenegraph.org http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org
Re: [osg-users] Better debug visualization in Visual Studio
Hi, have you tried it? Opinions, experience? Cheers, Dženan -- Read this topic online here: http://forum.openscenegraph.org/viewtopic.php?p=32971#32971 ___ osg-users mailing list osg-users@lists.openscenegraph.org http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org
Re: [osg-users] How to implement a fuel tank
HI Robert Thank you for your explanation, the focus of this implementation is the physics engine, and I intend to use physx to achieve physical simulation, using OSG rendering, should be able to get the effect I want, thanks again! ___ osg-users mailing list osg-users@lists.openscenegraph.org http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org
Re: [osg-users] How to implement a fuel tank
Hi Juan, On Fri, Oct 22, 2010 at 10:30 AM, Juan forsakenwarl...@gmail.com wrote: Thank you for your explanation, the focus of this implementation is the physics engine, and I intend to use physx to achieve physical simulation, using OSG rendering, should be able to get the effect I want, thanks again! There are users who tackle physics in the OSG community, but it's mainly centered on the rendering side. You'll certainly find more focussed assistance from communities around Physx or other physics engines so I'd recommend you ask you question there. Robert. ___ osg-users mailing list osg-users@lists.openscenegraph.org http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org
Re: [osg-users] Better debug visualization in Visual Studio
Hi, FYI, this is not compatible on VS2008 and previous versions. (I mean the _vcee_autoexp variable) For theses versions you have to edit the autoexp.dat file. Sorry I didn't tried already Vincent. Le 22/10/2010 10:35, Dženan Zukić a écrit : Hi, have you tried it? Opinions, experience? Cheers, Dženan -- Read this topic online here: http://forum.openscenegraph.org/viewtopic.php?p=32971#32971 ___ osg-users mailing list osg-users@lists.openscenegraph.org http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org ___ osg-users mailing list osg-users@lists.openscenegraph.org http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org
Re: [osg-users] Drawing complex labels of 3D points in screen space
Hi, Thank you for the answer. Now I'm searching for the right way to get to some level (for example, base level, e.g. root) of MVPW matrix stack from callback or accept() call. But all I see is top level, which is actual only for the current node. Yurii 2010/10/22 J.P. Delport jpdelp...@csir.co.za: Hi, On 21/10/10 18:26, Yurii Monakov wrote: Hi All After some googling I've discovered an AutoTransform node )) When setting it's auto-rotate mode to ROTATE_TO_SCREEN and setAutoScaleToScreen(true) it gives me exactly what I want. The only problem is that after looking at AutoTransform node I noticed that it does a lot of computation inside. The only thing I need is to push root's MVPW matrix to some part of the graph (Transform nodes, which are children of the screen-spaced node) to update their positions properly. Is there any other way to do this? not that I know of, but I think you've now discovered enough to create your own node, either based on AutoTransform (but for already screen-spaced geometry) or your own. jp Thank you! Yurii ___ osg-users mailing list osg-users@lists.openscenegraph.org http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org -- This message is subject to the CSIR's copyright terms and conditions, e-mail legal notice, and implemented Open Document Format (ODF) standard. The full disclaimer details can be found at http://www.csir.co.za/disclaimer.html. This message has been scanned for viruses and dangerous content by MailScanner, and is believed to be clean. MailScanner thanks Transtec Computers for their support. ___ osg-users mailing list osg-users@lists.openscenegraph.org http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org ___ osg-users mailing list osg-users@lists.openscenegraph.org http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org
Re: [osg-users] Drawing complex labels of 3D points in screen space
Hi, On 22/10/10 12:36, Yurii Monakov wrote: Hi, Thank you for the answer. Now I'm searching for the right way to get to some level (for example, base level, e.g. root) of MVPW matrix stack from callback or accept() call. But all I see is top level, which is actual only for the current node. I'm not 100% sure, but I think start by searching for use of node-getParentalNodePaths() or the use of CullStack. jp Yurii 2010/10/22 J.P. Delportjpdelp...@csir.co.za: Hi, On 21/10/10 18:26, Yurii Monakov wrote: Hi All After some googling I've discovered an AutoTransform node )) When setting it's auto-rotate mode to ROTATE_TO_SCREEN and setAutoScaleToScreen(true) it gives me exactly what I want. The only problem is that after looking at AutoTransform node I noticed that it does a lot of computation inside. The only thing I need is to push root's MVPW matrix to some part of the graph (Transform nodes, which are children of the screen-spaced node) to update their positions properly. Is there any other way to do this? not that I know of, but I think you've now discovered enough to create your own node, either based on AutoTransform (but for already screen-spaced geometry) or your own. jp Thank you! Yurii ___ osg-users mailing list osg-users@lists.openscenegraph.org http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org -- This message is subject to the CSIR's copyright terms and conditions, e-mail legal notice, and implemented Open Document Format (ODF) standard. The full disclaimer details can be found at http://www.csir.co.za/disclaimer.html. This message has been scanned for viruses and dangerous content by MailScanner, and is believed to be clean. MailScanner thanks Transtec Computers for their support. ___ osg-users mailing list osg-users@lists.openscenegraph.org http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org ___ osg-users mailing list osg-users@lists.openscenegraph.org http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org -- This message is subject to the CSIR's copyright terms and conditions, e-mail legal notice, and implemented Open Document Format (ODF) standard. The full disclaimer details can be found at http://www.csir.co.za/disclaimer.html. This message has been scanned for viruses and dangerous content by MailScanner, and is believed to be clean. MailScanner thanks Transtec Computers for their support. ___ osg-users mailing list osg-users@lists.openscenegraph.org http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org
Re: [osg-users] Animation with Cubic Bezier path
I've a pratical case, can anybody help me?? I've 4 keyframes: v[0] = osg::Vec3(0, 0, 0); v[1] = osg::Vec3(20, 0, 0); v[2] = osg::Vec3(40, 0, 0); v[3] = osg::Vec3(60, 0, 0); in times 0,2,4,6 How to move along this keyframes with a constant speed from 0,0 to 60,0??? thanks -- Read this topic online here: http://forum.openscenegraph.org/viewtopic.php?p=32978#32978 ___ osg-users mailing list osg-users@lists.openscenegraph.org http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org
Re: [osg-users] color not the good one -_-
Thanks for your answer... I just forgot to netral color on objets witch used this texture and it take something like 204-204-204 unsted of 0 0 0 now its works ... Thanks you (http://www.hordes.fr?ref=litllechicken) -- Read this topic online here: http://forum.openscenegraph.org/viewtopic.php?p=32979#32979 ___ osg-users mailing list osg-users@lists.openscenegraph.org http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org
Re: [osg-users] Animation with Cubic Bezier path
Hi Daniele, Have you looked at the osganimation example? Robert. On Fri, Oct 22, 2010 at 12:03 PM, daniele argiolas arj...@gmail.com wrote: I've a pratical case, can anybody help me?? I've 4 keyframes: v[0] = osg::Vec3(0, 0, 0); v[1] = osg::Vec3(20, 0, 0); v[2] = osg::Vec3(40, 0, 0); v[3] = osg::Vec3(60, 0, 0); in times 0,2,4,6 How to move along this keyframes with a constant speed from 0,0 to 60,0??? thanks -- Read this topic online here: http://forum.openscenegraph.org/viewtopic.php?p=32978#32978 ___ osg-users mailing list osg-users@lists.openscenegraph.org http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org ___ osg-users mailing list osg-users@lists.openscenegraph.org http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org
Re: [osg-users] Animation with Cubic Bezier path
yes, I've looked all examples. I need a bezier path animation with constant speed from first to last keyframe. Is it more clear? How? thanks robertosfield wrote: Hi Daniele, Have you looked at the osganimation example? Robert. On Fri, Oct 22, 2010 at 12:03 PM, daniele argiolas wrote: I've a pratical case, can anybody help me?? I've 4 keyframes: v[0] = osg::Vec3(0, 0, 0); v[1] = osg::Vec3(20, 0, 0); v[2] = osg::Vec3(40, 0, 0); v[3] = osg::Vec3(60, 0, 0); in times 0,2,4,6 How to move along this keyframes with a constant speed from 0,0 to 60,0??? thanks -- Read this topic online here: http://forum.openscenegraph.org/viewtopic.php?p=32978#32978 ___ osg-users mailing list http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org ___ osg-users mailing list http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org -- Post generated by Mail2Forum -- Read this topic online here: http://forum.openscenegraph.org/viewtopic.php?p=32981#32981 ___ osg-users mailing list osg-users@lists.openscenegraph.org http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org
[osg-users] Why is Vec not a template?
Hi, I just wondered why Vec3 and related Vec classes are all different classes and why they aren't realized as templates? best regards Mathias -- Institut für Informatik Universität Leipzig Johannisgasse 26, 04103 Leipzig Phone: +493419732283 ___ osg-users mailing list osg-users@lists.openscenegraph.org http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org
Re: [osg-users] integrating tessellation shaders into osg
Hi everyone, good News. I realized to add the tesselation shaders into OSG 2.9.9. It is still a beta version, so I have to test a little bit more before I will commit it to the community. I am gone over the week end but I will finish it during the next week. Hope to find some guys who will have a look over it and try some tests. Cheers, Holger -- Read this topic online here: http://forum.openscenegraph.org/viewtopic.php?p=32984#32984 ___ osg-users mailing list osg-users@lists.openscenegraph.org http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org
Re: [osg-users] Why is Vec not a template?
Hi Mathias, On Fri, Oct 22, 2010 at 1:01 PM, Mathias Goldau m...@informatik.uni-leipzig.de wrote: I just wondered why Vec3 and related Vec classes are all different classes and why they aren't realized as templates? Because templates aren't always the perfect tool for the job. Three main reasons : specialization of types required in various methods makes templates more awkward to write and maintain, secondly templates are slow to compile - the more low level templates you use the greater the slow down, and finally templates are pain in the butt to decifer when you get compile errors. Robert. ___ osg-users mailing list osg-users@lists.openscenegraph.org http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org
Re: [osg-users] Why is Vec not a template?
Am 22.10.2010 14:37, Robert Osfield wrote: Hi Mathias, On Fri, Oct 22, 2010 at 1:01 PM, Mathias Goldau m...@informatik.uni-leipzig.de wrote: I just wondered why Vec3 and related Vec classes are all different classes and why they aren't realized as templates? Because templates aren't always the perfect tool for the job. Three main reasons : specialization of types required in various methods makes templates more awkward to write and maintain, secondly templates are slow to compile - the more low level templates you use the greater the slow down, and finally templates are pain in the butt to decifer when you get compile errors. ok, I could imagine that at least the value_type for all Vec3 derivates would be a good choice for a template, since the operations for Vec3f and Vec3d, etc. are exactly the same (I guess). I really could imagine that a template size_t, typename value_type is indeed really awkward and hard to maintain. But so you will have much more code duplication. In point of compile time: Hmm the Vec[2-4] classes are not very big, there are many other templates in OSG which have much higher complexity. Mathias -- Institut für Informatik Universität Leipzig Johannisgasse 26, 04103 Leipzig Phone: +493419732283 ___ osg-users mailing list osg-users@lists.openscenegraph.org http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org
[osg-users] VertexBufferObects causing crash/bad geometry [2.8.2]
Hi,
This is a Windows 32/64app, fully single threaded(app and OSG in one thread).
Quadro FX1800, latest drivers.
I am rendering basic polygonal geometry - and I have been experimenting with
turning on use of VBO's. I want to emphasize that the app shows zero issues
until I turn VBO's on for geometry with a PrimitiveSet such as
osg::DrawElementsUInt(osg::PrimitiveSet::TRIANGLES)
.e.g
if(osgGrid_-areFastPathsUsed()){
osgGrid_-setUseVertexBufferObjects(true);
}else{
osgGrid_-setUseDisplayList(true);
}
When I turn on VBO's , the normally reliable app starts reproducibly showing
essentially random geometry for the objects (only) where UseVertexBufferObjects
is true. Non-VBO objects render fine. And the VBO objects render fine when
useVBO is off.
Crashes sometimes result. The symptoms are similar to multithreading issues
where one thread is accessing stale data from another thread.
Are the calls to the underlying rendering of VBO's asynchronous? It's as if
the VBO calls are getting stale data or something. Problems seem to happen
particularly if I close a window then open another one, again as if the
rendering was continuing with 'old' data in the old window.
All sounds a bit vague, I know, but these types of things are a nightmare to
debug.
Andrew
--
Read this topic online here:
http://forum.openscenegraph.org/viewtopic.php?p=32990#32990
___
osg-users mailing list
osg-users@lists.openscenegraph.org
http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org
Re: [osg-users] VPB problem with large database...
Jp, Thanks for the suggestion. I'll give it a try. Thinking that it might be a video card limitation (texture memory), I tried building the same content on a more capable video card (Nvidia 9800 GTX) and I'm still having the problem of having the osgdem threads dying out. I'm suspecting that VPB is having trouble with the large tif files and the number of them. I have 10,271 tif files in a single directory where each tif file is ~30MB in size. If I pare down the number of files to just a few, it builds fine. When I throw the full gamut of tif files at it, VPB chokes. Apparently there's some sort of limitation here to what it can handle. I'll continue to investigate and post my findings... -Shayne -Original Message- From: osg-users-boun...@lists.openscenegraph.org [mailto:osg-users-boun...@lists.openscenegraph.org] On Behalf Of J.P. Delport Sent: Friday, October 22, 2010 12:30 AM To: OpenSceneGraph Users Subject: Re: [osg-users] VPB problem with large database... Hi, you can check in the tasks folder to see exactly the osgdem command(s) that would be executed by vpb. Try execute one of these manually and see if it does not report any useful errors. jp On 21/10/10 20:31, Tueller, Shayne R Civ USAF AFMC 519 SMXS/MXDEC wrote: All, I'm seeing some issues with VPB when attempting to build a very large database. My command is the following: vpbmaster -TERRAIN -PagedLOD -geocentric -whole-globe -t wholeearthtexture -t detailtexture -d dted -l 24 -o database.ive What I'm seeing is that vpbmaster creates the tasks and logs directories followed by the scheduling of tasks. During the scheduling of tasks, I see the threads osgdem. created on the machine and then die a few seconds later. vpbmaster then just waits for something to happen and eventually goes to sleep. I'm guessing that this is happening because the threads died for some reason and vpbmaster just sits there waiting for them to come back. I understand the VPB is able to handle the building of very large databases but it seems to be choking in our particular case. If I run the same command above on a smaller set of data, things run just fine. Our system has 16 processors with ample disk and memory space to handle what is needed. It also runs Fadora core 13. Has anyone else seen this issue? Is there a flag or setting that I need to use to build a large database? Any input would be appreciated. -Shayne ___ osg-users mailing list osg-users@lists.openscenegraph.org http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org -- This message is subject to the CSIR's copyright terms and conditions, e-mail legal notice, and implemented Open Document Format (ODF) standard. The full disclaimer details can be found at http://www.csir.co.za/disclaimer.html. This message has been scanned for viruses and dangerous content by MailScanner, and is believed to be clean. MailScanner thanks Transtec Computers for their support. ___ osg-users mailing list osg-users@lists.openscenegraph.org http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org smime.p7s Description: S/MIME cryptographic signature ___ osg-users mailing list osg-users@lists.openscenegraph.org http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org
Re: [osg-users] Terrain geometry
On 10/21/2010 3:46 PM, Robert Gosztyla wrote: Hi, I'm back again with my terrain editing question :). I was serching forum for some solution, but i didn't find any. If i have found at least terraintile geometry and drawable associated with that how i could change this geometry? I've tried to use dirtyDisplayList() method, but it doesn't work as expected, i found also in some post that for terrain rendering this method won't work. There is also project called osgTDS, which modifies terrain geometry during loading from file and there is used method replaceDrawable() for geode. Is it the only solution for it? Or i can do is in some other way too? I rather don't want to modify height file, because probably i need some additional adjustment than only height on my terrain. I haven't tried modifying the heightfield on the fly while it is already displayed, but as far as I know, dirtying the Node should cause an update of any precomputed or cached data. I have used the modify-the-heightfield method during load and it works fine. perhaps Robert Osfield can comment on what's needed to update the displayed representation once it is already loaded. -- Chris 'Xenon' Hanson, omo sanza lettere. xe...@alphapixel.com http://www.alphapixel.com/ Digital Imaging. OpenGL. Scene Graphs. GIS. GPS. Training. Consulting. Contracting. There is no Truth. There is only Perception. To Perceive is to Exist. - Xen ___ osg-users mailing list osg-users@lists.openscenegraph.org http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org
Re: [osg-users] Upside down smoke trail
Hi Kevin,
Below is a piece of code that does exactly what you want in global WGS84
ellipsoid coord space. In your case just replace _xyz with the direction
vector you want the smoke to move.
osgParticle::FluidProgram* fp =
dynamic_castosgParticle::FluidProgram*
( _smokeTrailParticleEffect-getProgram() );
if( fp )
{
osg::Vec3 v = _xyz;
if( v.normalize() 0.0 )
fp-setAcceleration( v * 9.81 );
}
Cheers,
Wojtek Lewandowski
-Original Message-
From: Kevin Bentley
Sent: Thursday, October 21, 2010 10:35 PM
To: osg-users@lists.openscenegraph.org
Subject: [osg-users] Upside down smoke trail
Hi,
I want to have a particle effect that looks like falling dust/dirt. I was
thinking of using something like a smoketrail effect (with a different
texture), but I can't figure out how to make the particles go down instead
of up.
Amy I on the wrong track here? Any ideas how to best accomplish a falling
dust/dirt effect?
Thanks,
Kevin
--
Read this topic online here:
http://forum.openscenegraph.org/viewtopic.php?p=32953#32953
___
osg-users mailing list
osg-users@lists.openscenegraph.org
http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org
___
osg-users mailing list
osg-users@lists.openscenegraph.org
http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org
Re: [osg-users] Animation with Cubic Bezier path
daniele argiolas wrote: yes, I've looked all examples. I need a bezier path animation with constant speed from first to last keyframe. Is it more clear? You might need to implement it yourself if OSG's spline interpolation doesn't provide a constant velocity. Here's a link I found on the topic: http://community.spinxengine.com/content/19-constant-speed-cubic-spline-interpolation.html --J ___ osg-users mailing list osg-users@lists.openscenegraph.org http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org
Re: [osg-users] Why is Vec not a template?
Hi Mathias, On Fri, Oct 22, 2010 at 2:45 PM, Mathias Goldau m...@informatik.uni-leipzig.de wrote: I really could imagine that a template size_t, typename value_type is indeed really awkward and hard to maintain. But so you will have much more code duplication. There is practically no maintenance work required on the Vec classes, they just work, so having duplicated is of no cost to the community. In point of compile time: Hmm the Vec[2-4] classes are not very big, there are many other templates in OSG which have much higher complexity. Templates are used in the OSG pretty sparingly, and generally for very specific purposes where they add really extra value. For instance you ref_ptr and TriangleIndexFunctor are both examples where the alternatively of not using a template would be a nightmare of coding for end users. There are very few really complicated templates though, any type of coding that tries to be really clever is a recipe of a long term maintenance disaster so is something I prefer to avoid. Robert. ___ osg-users mailing list osg-users@lists.openscenegraph.org http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org
Re: [osg-users] VertexBufferObects causing crash/bad geometry [2.8.2]
Hi Andrew,
This sounds like a driver issue or perhaps corrupted data on the app
side. Could you try your app on another machine?
Robert.
On Fri, Oct 22, 2010 at 3:27 PM, Andrew Cunningham andr...@mac.com wrote:
Hi,
This is a Windows 32/64app, fully single threaded(app and OSG in one thread).
Quadro FX1800, latest drivers.
I am rendering basic polygonal geometry - and I have been experimenting with
turning on use of VBO's. I want to emphasize that the app shows zero issues
until I turn VBO's on for geometry with a PrimitiveSet such as
osg::DrawElementsUInt(osg::PrimitiveSet::TRIANGLES)
.e.g
if(osgGrid_-areFastPathsUsed()){
osgGrid_-setUseVertexBufferObjects(true);
}else{
osgGrid_-setUseDisplayList(true);
}
When I turn on VBO's , the normally reliable app starts reproducibly showing
essentially random geometry for the objects (only) where
UseVertexBufferObjects is true. Non-VBO objects render fine. And the VBO
objects render fine when useVBO is off.
Crashes sometimes result. The symptoms are similar to multithreading issues
where one thread is accessing stale data from another thread.
Are the calls to the underlying rendering of VBO's asynchronous? It's as if
the VBO calls are getting stale data or something. Problems seem to happen
particularly if I close a window then open another one, again as if the
rendering was continuing with 'old' data in the old window.
All sounds a bit vague, I know, but these types of things are a nightmare to
debug.
Andrew
--
Read this topic online here:
http://forum.openscenegraph.org/viewtopic.php?p=32990#32990
___
osg-users mailing list
osg-users@lists.openscenegraph.org
http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org
___
osg-users mailing list
osg-users@lists.openscenegraph.org
http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org
Re: [osg-users] [3rdparty] Sound In OSG
Are there standard ways to configure sound placement from outside the source code? Looking through the 1 page of documentation that comes with the source code as well as the examples, I see that I can place a sound object at an (x,y) position - so I can manually place it somewhere in my model. However, what I'd really like to do is configure Sound1 to position itself at Node1, so that when I make changes to my coordinate system, model, etc. I won't have to recalculate and re-code position information. I see that sound nodes are extensions of OSG's Node so I can manipulate them in the same way, but I would still need to change the source code if I ever moved or manipulated my 3D model. For example, I have a large model of a school and I'd like to attach a bell ring to the bell tower, cheers to the stadium, etc. It would be great if I could just write a config file or smarter source code that would always pair each sound with its part of the model in the scene. Has anyone solved this task before? Also, is there any detailed documentation for this system? The only documentation I found on the google project site was for building the system - and then I only saw 1 page of documentation in the source. Thanks. Matt -- Read this topic online here: http://forum.openscenegraph.org/viewtopic.php?p=32996#32996 ___ osg-users mailing list osg-users@lists.openscenegraph.org http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org
Re: [osg-users] VPB problem with large database...
Hi Shayne, I don't know what might be causing problems on your system, debugging this type of thing remotely is next to impossible. The best I can do is suggest trying the svn/trunk version of the OSG VPB. With VPB is now possible build using NVidia texture tools and with this avoid the need for graphics context and associated graphics card - instead you can just compile and run with Mesa libs. For more info have a look at the recent threads on osg-users and osg-submissions about NVidia Texture Tools integration. Robert. On Fri, Oct 22, 2010 at 3:48 PM, Tueller, Shayne R Civ USAF AFMC 519 SMXS/MXDEC shayne.tuel...@hill.af.mil wrote: Jp, Thanks for the suggestion. I'll give it a try. Thinking that it might be a video card limitation (texture memory), I tried building the same content on a more capable video card (Nvidia 9800 GTX) and I'm still having the problem of having the osgdem threads dying out. I'm suspecting that VPB is having trouble with the large tif files and the number of them. I have 10,271 tif files in a single directory where each tif file is ~30MB in size. If I pare down the number of files to just a few, it builds fine. When I throw the full gamut of tif files at it, VPB chokes. Apparently there's some sort of limitation here to what it can handle. I'll continue to investigate and post my findings... -Shayne -Original Message- From: osg-users-boun...@lists.openscenegraph.org [mailto:osg-users-boun...@lists.openscenegraph.org] On Behalf Of J.P. Delport Sent: Friday, October 22, 2010 12:30 AM To: OpenSceneGraph Users Subject: Re: [osg-users] VPB problem with large database... Hi, you can check in the tasks folder to see exactly the osgdem command(s) that would be executed by vpb. Try execute one of these manually and see if it does not report any useful errors. jp On 21/10/10 20:31, Tueller, Shayne R Civ USAF AFMC 519 SMXS/MXDEC wrote: All, I'm seeing some issues with VPB when attempting to build a very large database. My command is the following: vpbmaster -TERRAIN -PagedLOD -geocentric -whole-globe -t wholeearthtexture -t detailtexture -d dted -l 24 -o database.ive What I'm seeing is that vpbmaster creates the tasks and logs directories followed by the scheduling of tasks. During the scheduling of tasks, I see the threads osgdem. created on the machine and then die a few seconds later. vpbmaster then just waits for something to happen and eventually goes to sleep. I'm guessing that this is happening because the threads died for some reason and vpbmaster just sits there waiting for them to come back. I understand the VPB is able to handle the building of very large databases but it seems to be choking in our particular case. If I run the same command above on a smaller set of data, things run just fine. Our system has 16 processors with ample disk and memory space to handle what is needed. It also runs Fadora core 13. Has anyone else seen this issue? Is there a flag or setting that I need to use to build a large database? Any input would be appreciated. -Shayne ___ osg-users mailing list osg-users@lists.openscenegraph.org http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org -- This message is subject to the CSIR's copyright terms and conditions, e-mail legal notice, and implemented Open Document Format (ODF) standard. The full disclaimer details can be found at http://www.csir.co.za/disclaimer.html. This message has been scanned for viruses and dangerous content by MailScanner, and is believed to be clean. MailScanner thanks Transtec Computers for their support. ___ osg-users mailing list osg-users@lists.openscenegraph.org http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org ___ osg-users mailing list osg-users@lists.openscenegraph.org http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org ___ osg-users mailing list osg-users@lists.openscenegraph.org http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org
[osg-users] PagedLOD - too many nodes
Hi, I am using osg::PagedLOD nodes in my project for large models visualization. The model is splitted into several parts (like octree) and multiresolution tree is made - same as VPB uses: Code: Group PagedLOD child 0 - local tile geometry at low resolution child 1 - external file reference PagedLOD child 0 - local tile geometry at low resolution child 1 - external file reference I understand the system of unloading nodes by DatabasePager, when PagedLOD nodes count is bigger than OSG_MAX_PAGEDLOD (I hope that unloaded pagedlod nodes unload also corresponding data on GPU - when using VBO) The system works fine, but the problem is in good PagedLOD::setRangeMode and PagedLOD::setRange (between lower res. geode and external reference) data setting. Since my model is not terrain, definition of setRange value is crucial. I use PIXEL_SIZE_ON_SCREEN metrics and by low setRange number the high resolution files are requested, high number of Geometries is loaded and openGL out-of-memory occurs and no new nodes are loaded. The setRange number depends on graphical window size, GPU memory, so it should be calculated somehow - to get the best performance from hardware. Does anyone have tip how to dynamically adjust setRange values? Or is any better technique with different metrics? I have an idea to adjust setRange (or change LODScale) when an out-of-memory openGL error occurs. Then I would change the setRange value (it is same for all pagedLOD nodes), free the GPU memory and load again. What do you think about this idea? How do I free occupied GPU memory? Is is enough to reconnect the subtree from scene? How do I catch openGL error? Could culling settings help here to have smaller scene? Culling in my app is set to OSG default. Many thanks in advance, Tomas -- Read this topic online here: http://forum.openscenegraph.org/viewtopic.php?p=33000#33000 ___ osg-users mailing list osg-users@lists.openscenegraph.org http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org
Re: [osg-users] [3rdparty] Sound In OSG
On 10/22/2010 10:32 AM, Matt Caron wrote: Are there standard ways to configure sound placement from outside the source code? Looking through the 1 page of documentation that comes with the source code as well as the examples, I see that I can place a sound object at an (x,y) position - so I can manually place it somewhere in my model. However, what I'd really like to do is configure Sound1 to position itself at Node1, so that when I make changes to my coordinate system, model, etc. I won't have to recalculate and re-code position information. I see that sound nodes are extensions of OSG's Node so I can manipulate them in the same way, but I would still need to change the source code if I ever moved or manipulated my 3D model. For example, I have a large model of a school and I'd like to attach a bell ring to the bell tower, cheers to the stadium, etc. It would be great if I could just write a config file or smarter source code that would always pair each sound with its part of the model in the scene. Has anyone solved this task before? I guess I'd probably make a visitor that goes out and located the nodes you want by name, and attaches similarly-named sounds to them? That way wherever those nodes are in the model, they get their appropriate sounds. Also, is there any detailed documentation for this system? The only documentation I found on the google project site was for building the system - and then I only saw 1 page of documentation in the source. Every significant class has Doxygen-type comment documentation embedded in the source. That, coupled with the examples, is all there is. Similar to OSG itself. ;) Thanks. Matt -- Chris 'Xenon' Hanson, omo sanza lettere. xe...@alphapixel.com http://www.alphapixel.com/ Digital Imaging. OpenGL. Scene Graphs. GIS. GPS. Training. Consulting. Contracting. There is no Truth. There is only Perception. To Perceive is to Exist. - Xen ___ osg-users mailing list osg-users@lists.openscenegraph.org http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org
Re: [osg-users] PagedLOD - too many nodes
Hi Thomas, It's possible to adjust the children selection dynamically by adjusting the Camera::LODScale via viewer.getCamera()-setLODScale(myScale);. However, I think this is probably just a hack around what it essentially an issue of poor load balancing in the paged database you have built. In essence what you are trying to do when using LOD/PagedLOD's is to keep the load on the whole computer to a level that the system can handle it comfortably. Think about trying to aim for a constant number of nodes in memory and on-screen, and constant number of polygons and textures on screen, we can never attain this exactly but this is the goal of load balancing. Using of PagedLOD over LOD allows you to scale higher as you are no-longer limited by main memory, however, you can still easily build databases that are too big even for runtime use - if this is the case then you either have to simplify the children at each LOD range or make the LOD's ranges more conservative. With an Octree the potentially for an explosion of nodes and complexity of the scene is pretty high unless the Octree is sparse, this keep balancing the books you will have to be prepared to be quite conservative with your range settings. There is of course often other ways to solve problems. Perhaps some other type of rendering/scene graph approach can held lower the load, this needn't be in place of paging. What other techniques might appropriate I can't say as I know close to nothing about your application and the type of data you are trying to handle. Robert. Robert. On Fri, Oct 22, 2010 at 5:41 PM, Tomas Hnilica tomas.hnil...@webstep.net wrote: Hi, I am using osg::PagedLOD nodes in my project for large models visualization. The model is splitted into several parts (like octree) and multiresolution tree is made - same as VPB uses: Code: Group PagedLOD child 0 - local tile geometry at low resolution child 1 - external file reference PagedLOD child 0 - local tile geometry at low resolution child 1 - external file reference I understand the system of unloading nodes by DatabasePager, when PagedLOD nodes count is bigger than OSG_MAX_PAGEDLOD (I hope that unloaded pagedlod nodes unload also corresponding data on GPU - when using VBO) The system works fine, but the problem is in good PagedLOD::setRangeMode and PagedLOD::setRange (between lower res. geode and external reference) data setting. Since my model is not terrain, definition of setRange value is crucial. I use PIXEL_SIZE_ON_SCREEN metrics and by low setRange number the high resolution files are requested, high number of Geometries is loaded and openGL out-of-memory occurs and no new nodes are loaded. The setRange number depends on graphical window size, GPU memory, so it should be calculated somehow - to get the best performance from hardware. Does anyone have tip how to dynamically adjust setRange values? Or is any better technique with different metrics? I have an idea to adjust setRange (or change LODScale) when an out-of-memory openGL error occurs. Then I would change the setRange value (it is same for all pagedLOD nodes), free the GPU memory and load again. What do you think about this idea? How do I free occupied GPU memory? Is is enough to reconnect the subtree from scene? How do I catch openGL error? Could culling settings help here to have smaller scene? Culling in my app is set to OSG default. Many thanks in advance, Tomas -- Read this topic online here: http://forum.openscenegraph.org/viewtopic.php?p=33000#33000 ___ osg-users mailing list osg-users@lists.openscenegraph.org http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org ___ osg-users mailing list osg-users@lists.openscenegraph.org http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org
Re: [osg-users] VertexBufferObects causing crash/bad geometry [2.8.2]
Hi Robert, Tried two XP machines ... identical behavior as long as H/W acceleration is on. Both graphics cards are nVidia, though. If I disable the drivers, there are no problems , but that is not surprising as OSG defaults to a non-VBO code path. I am using the OSG API in a very vanilla way, and both debug and release builds show the issue. If I was exceeding vector bounds or otherwise abusing the OSG API that should be caught in the debug build or by OSG_NOTIFY_LEVEL Quite the mystery, I can toggle my use of VBOs at runtime so may look into this with some memory checking tools. Andrew -- Read this topic online here: http://forum.openscenegraph.org/viewtopic.php?p=33003#33003 ___ osg-users mailing list osg-users@lists.openscenegraph.org http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org
Re: [osg-users] Why is Vec not a template?
Gent's, I was wondering the same. A templated Vec class isn't all that hard to implement and maintain, it's pretty straight forward. What's a pain is a templated Matrix class, although totally doable (from experience). I would gladly help if someone wanted to start taking on the task. -B -Original Message- From: osg-users-boun...@lists.openscenegraph.org [mailto:osg-users-boun...@lists.openscenegraph.org] On Behalf Of Mathias Goldau Sent: Friday, October 22, 2010 6:46 AM To: osg-users@lists.openscenegraph.org Subject: Re: [osg-users] Why is Vec not a template? Am 22.10.2010 14:37, Robert Osfield wrote: Hi Mathias, On Fri, Oct 22, 2010 at 1:01 PM, Mathias Goldau m...@informatik.uni-leipzig.de wrote: I just wondered why Vec3 and related Vec classes are all different classes and why they aren't realized as templates? Because templates aren't always the perfect tool for the job. Three main reasons : specialization of types required in various methods makes templates more awkward to write and maintain, secondly templates are slow to compile - the more low level templates you use the greater the slow down, and finally templates are pain in the butt to decifer when you get compile errors. ok, I could imagine that at least the value_type for all Vec3 derivates would be a good choice for a template, since the operations for Vec3f and Vec3d, etc. are exactly the same (I guess). I really could imagine that a template size_t, typename value_type is indeed really awkward and hard to maintain. But so you will have much more code duplication. In point of compile time: Hmm the Vec[2-4] classes are not very big, there are many other templates in OSG which have much higher complexity. Mathias -- Institut für Informatik Universität Leipzig Johannisgasse 26, 04103 Leipzig Phone: +493419732283 ___ osg-users mailing list osg-users@lists.openscenegraph.org http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org ___ osg-users mailing list osg-users@lists.openscenegraph.org http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org
Re: [osg-users] VPB problem with large database...
Robert, Thanks for the reply. I can try the latest version of OSG/VPB to build the database but will the database still be compatible with an app that uses OSG 2.8.2 to fly through the database? I'm not really in a position to update our application software to the latest version of OSG at this time. If I can still get by with using OSG 2.8.2 in our app, then I'm more amenable to upgrading to the latest OSG/VPB for our build machine. Thanks for your input and help. -Shayne -Original Message- From: osg-users-boun...@lists.openscenegraph.org [mailto:osg-users-boun...@lists.openscenegraph.org] On Behalf Of Robert Osfield Sent: Friday, October 22, 2010 10:32 AM To: OpenSceneGraph Users Subject: Re: [osg-users] VPB problem with large database... Hi Shayne, I don't know what might be causing problems on your system, debugging this type of thing remotely is next to impossible. The best I can do is suggest trying the svn/trunk version of the OSG VPB. With VPB is now possible build using NVidia texture tools and with this avoid the need for graphics context and associated graphics card - instead you can just compile and run with Mesa libs. For more info have a look at the recent threads on osg-users and osg-submissions about NVidia Texture Tools integration. Robert. On Fri, Oct 22, 2010 at 3:48 PM, Tueller, Shayne R Civ USAF AFMC 519 SMXS/MXDEC shayne.tuel...@hill.af.mil wrote: Jp, Thanks for the suggestion. I'll give it a try. Thinking that it might be a video card limitation (texture memory), I tried building the same content on a more capable video card (Nvidia 9800 GTX) and I'm still having the problem of having the osgdem threads dying out. I'm suspecting that VPB is having trouble with the large tif files and the number of them. I have 10,271 tif files in a single directory where each tif file is ~30MB in size. If I pare down the number of files to just a few, it builds fine. When I throw the full gamut of tif files at it, VPB chokes. Apparently there's some sort of limitation here to what it can handle. I'll continue to investigate and post my findings... -Shayne -Original Message- From: osg-users-boun...@lists.openscenegraph.org [mailto:osg-users-boun...@lists.openscenegraph.org] On Behalf Of J.P. Delport Sent: Friday, October 22, 2010 12:30 AM To: OpenSceneGraph Users Subject: Re: [osg-users] VPB problem with large database... Hi, you can check in the tasks folder to see exactly the osgdem command(s) that would be executed by vpb. Try execute one of these manually and see if it does not report any useful errors. jp On 21/10/10 20:31, Tueller, Shayne R Civ USAF AFMC 519 SMXS/MXDEC wrote: All, I'm seeing some issues with VPB when attempting to build a very large database. My command is the following: vpbmaster -TERRAIN -PagedLOD -geocentric -whole-globe -t wholeearthtexture -t detailtexture -d dted -l 24 -o database.ive What I'm seeing is that vpbmaster creates the tasks and logs directories followed by the scheduling of tasks. During the scheduling of tasks, I see the threads osgdem. created on the machine and then die a few seconds later. vpbmaster then just waits for something to happen and eventually goes to sleep. I'm guessing that this is happening because the threads died for some reason and vpbmaster just sits there waiting for them to come back. I understand the VPB is able to handle the building of very large databases but it seems to be choking in our particular case. If I run the same command above on a smaller set of data, things run just fine. Our system has 16 processors with ample disk and memory space to handle what is needed. It also runs Fadora core 13. Has anyone else seen this issue? Is there a flag or setting that I need to use to build a large database? Any input would be appreciated. -Shayne ___ osg-users mailing list osg-users@lists.openscenegraph.org http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org -- This message is subject to the CSIR's copyright terms and conditions, e-mail legal notice, and implemented Open Document Format (ODF) standard. The full disclaimer details can be found at http://www.csir.co.za/disclaimer.html. This message has been scanned for viruses and dangerous content by MailScanner, and is believed to be clean. MailScanner thanks Transtec Computers for their support. ___ osg-users mailing list osg-users@lists.openscenegraph.org http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org ___ osg-users mailing list osg-users@lists.openscenegraph.org http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org ___ osg-users mailing list osg-users@lists.openscenegraph.org
[osg-users] Wrong CMake output directory using Visual - half submission
Hi Robert, hi all,
I spotted the binaries are not in /bin as usual with my new config (CMake
2.8.2, VS2010 x64), but in /bin/Debug and /bin/Release.
I think CMake 2.8.2 is the cause because the usual way to work around these
subdirs should not work anymore (SET_TARGET_PROPERTIES(Target PROPERTIES
PREFIX ../)). I found a way to make it work properly. However I don't post it
as a submussion because I didn't test it as it should (not tested under Linux),
and I guess we must write some versionning code (if cmake version is below
x.y.z...). Here is my code, to be put in the root CMakeLists.txt:
Instead of using
SET(CMAKE_ARCHIVE_OUTPUT_DIRECTORY ${OUTPUT_LIBDIR})
and such, use:
SET(CMAKE_ARCHIVE_OUTPUT_DIRECTORY_CONFIG ${OUTPUT_LIBDIR})
So the code should look like:
FOREACH(CONF ${CMAKE_CONFIGURATION_TYPES})
STRING(TOUPPER ${CONF} CONF)
SET(CMAKE_ARCHIVE_OUTPUT_DIRECTORY_${CONF} ${OUTPUT_LIBDIR})
SET(CMAKE_RUNTIME_OUTPUT_DIRECTORY_${CONF} ${OUTPUT_BINDIR})
IF(WIN32)
SET(CMAKE_LIBRARY_OUTPUT_DIRECTORY_${CONF} ${OUTPUT_BINDIR})
ELSE()
SET(CMAKE_LIBRARY_OUTPUT_DIRECTORY_${CONF} ${OUTPUT_LIBDIR})
ENDIF()
ENDFOREACH()
Hope someone will take a few minutes to write a nice CMake script (and figure
out the version number from which this can bu used, if this is really 100%
CMake related)
Cheers,
Sukender
PVLE - Lightweight cross-platform game engine - http://pvle.sourceforge.net/
___
osg-users mailing list
osg-users@lists.openscenegraph.org
http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org
Re: [osg-users] Wrong CMake output directory using Visual - half submission
Sukender,
It is indeed related to CMake = 2.8.1. I have had success setting all
configuration types including the generic one, which seems to work on Linux,
OSX and Windows. So:
SET(CMAKE_ARCHIVE_OUTPUT_DIRECTORY ${OUTPUT_LIBDIR})
SET(CMAKE_ARCHIVE_OUTPUT_DIRECTORY_RELEASE ${OUTPUT_LIBDIR})
SET(CMAKE_ARCHIVE_OUTPUT_DIRECTORY_DEBUG ${OUTPUT_LIBDIR})
That way it works with older and newer versions of CMake. Older versions don't
know/use/ignore the specific configuration types and newer CMake uses the most
specific setting.
The conditional below I use for other situations. It works for me since I am
simply checking against previous versions:
# CMake = 2.8.1 changed the output directory algorithm...
if (CMAKE_CONFIGURATION_TYPES AND NOT APPLE)
if (${CMAKE_MAJOR_VERSION} EQUAL 2 AND
(${CMAKE_MINOR_VERSION}.${CMAKE_PATCH_VERSION} LESS 8.1))
#Perform 'old' CMake method.
# ...
endif ()
endif ()
HTH
Chuck Seberino
On Oct 22, 2010, at 11:13 AM, Sukender wrote:
Hi Robert, hi all,
I spotted the binaries are not in /bin as usual with my new config (CMake
2.8.2, VS2010 x64), but in /bin/Debug and /bin/Release.
I think CMake 2.8.2 is the cause because the usual way to work around these
subdirs should not work anymore (SET_TARGET_PROPERTIES(Target PROPERTIES
PREFIX ../)). I found a way to make it work properly. However I don't post
it as a submussion because I didn't test it as it should (not tested under
Linux), and I guess we must write some versionning code (if cmake version is
below x.y.z...). Here is my code, to be put in the root CMakeLists.txt:
Instead of using
SET(CMAKE_ARCHIVE_OUTPUT_DIRECTORY ${OUTPUT_LIBDIR})
and such, use:
SET(CMAKE_ARCHIVE_OUTPUT_DIRECTORY_CONFIG ${OUTPUT_LIBDIR})
So the code should look like:
FOREACH(CONF ${CMAKE_CONFIGURATION_TYPES})
STRING(TOUPPER ${CONF} CONF)
SET(CMAKE_ARCHIVE_OUTPUT_DIRECTORY_${CONF} ${OUTPUT_LIBDIR})
SET(CMAKE_RUNTIME_OUTPUT_DIRECTORY_${CONF} ${OUTPUT_BINDIR})
IF(WIN32)
SET(CMAKE_LIBRARY_OUTPUT_DIRECTORY_${CONF} ${OUTPUT_BINDIR})
ELSE()
SET(CMAKE_LIBRARY_OUTPUT_DIRECTORY_${CONF} ${OUTPUT_LIBDIR})
ENDIF()
ENDFOREACH()
Hope someone will take a few minutes to write a nice CMake script (and figure
out the version number from which this can bu used, if this is really 100%
CMake related)
Cheers,
Sukender
PVLE - Lightweight cross-platform game engine - http://pvle.sourceforge.net/
___
osg-users mailing list
osg-users@lists.openscenegraph.org
http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org
___
osg-users mailing list
osg-users@lists.openscenegraph.org
http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org
Re: [osg-users] Why is Vec not a template?
Hi Brad, On Fri, Oct 22, 2010 at 6:17 PM, Brad Colbert bcolb...@rscusa.com wrote: A templated Vec class isn't all that hard to implement and maintain, it's pretty straight forward. There are awkward, and making it templated is totally pointless. What is there to gain? What's a pain is a templated Matrix class, although totally doable (from experience). I would gladly help if someone wanted to start taking on the task. I'm not about to merge code that offers no advancement in features, performance or maintainability. There are plenty of things the community can help out with, stuff that doesn't advance the OSG is a step backwards. Robert. ___ osg-users mailing list osg-users@lists.openscenegraph.org http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org
Re: [osg-users] Why is Vec not a template?
I'm not suggesting it happen, just that if anyone is interested I can give them moral support ;) For the sake of argument, however, the advantage of template vector and matrix classes is that where any operations can be evaluated during compile time, will be (assuming that the operations are coded using template metaprogramming techniques). This would result in less work for the CPU during run-time. One thing I would like to see is, where possible, taking advantage of SSE(2/3) instructions on CPU's that have them. I imagine that there could be some speed up, especially if operations could be batched (I'm sure a whole other can-o-worms). Anyway, for what it's worth. -B -Original Message- From: osg-users-boun...@lists.openscenegraph.org [mailto:osg-users-boun...@lists.openscenegraph.org] On Behalf Of Robert Osfield Sent: Friday, October 22, 2010 11:36 AM To: OpenSceneGraph Users Subject: Re: [osg-users] Why is Vec not a template? Hi Brad, On Fri, Oct 22, 2010 at 6:17 PM, Brad Colbert bcolb...@rscusa.com wrote: A templated Vec class isn't all that hard to implement and maintain, it's pretty straight forward. There are awkward, and making it templated is totally pointless. What is there to gain? What's a pain is a templated Matrix class, although totally doable (from experience). I would gladly help if someone wanted to start taking on the task. I'm not about to merge code that offers no advancement in features, performance or maintainability. There are plenty of things the community can help out with, stuff that doesn't advance the OSG is a step backwards. Robert. ___ osg-users mailing list osg-users@lists.openscenegraph.org http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org ___ osg-users mailing list osg-users@lists.openscenegraph.org http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org
Re: [osg-users] Box Selection Tool
I've come to realize that it wasn't the intersector that was being slow, it was selecting all of the nodes that the intersector found that was being slow. One thing that concerns me though is that the intersector is returning over 95 intersections, which is over double the amount of polygons in my geometry. Anybody know why this is happening? Thanks, Scott -- Read this topic online here: http://forum.openscenegraph.org/viewtopic.php?p=33010#33010 ___ osg-users mailing list osg-users@lists.openscenegraph.org http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org
Re: [osg-users] Why is Vec not a template?
Hi Brad, On Fri, Oct 22, 2010 at 8:13 PM, Brad Colbert bcolb...@rscusa.com wrote: One thing I would like to see is, where possible, taking advantage of SSE(2/3) instructions on CPU's that have them. I imagine that there could be some speed up, especially if operations could be batched (I'm sure a whole other can-o-worms). For apps that are limited by math performance this would be useful, but for most scene graph applications the matrix maths is never anything close to a bottleneck. We are mostly bandwidth limited, and what CPU overheads we see are largely down to cache misses as the scene graph churns through lots of data in memory. So stuff like SSE optimizations are really the for physics engines and alike rather than scene graphs. Robert. ___ osg-users mailing list osg-users@lists.openscenegraph.org http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org
[osg-users] OSGDEM output flattening
I'm passing along a question from our database team: We would like to flatten certain areas of the terrain polygons for a given DTED/DEM input file. Is there an option for VPB where we can specify a shapefile of the flattened area? Any help/info would be greatly appreciated. -B --- Brad Colbert Renaissance Sciences Corporation ___ osg-users mailing list osg-users@lists.openscenegraph.org http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org
Re: [osg-users] Why is Vec not a template?
-BEGIN PGP SIGNED MESSAGE- Hash: SHA1 Hello Robert, On 10/22/2010 10:03 PM, Robert Osfield wrote: Hi Brad, On Fri, Oct 22, 2010 at 8:13 PM, Brad Colbert bcolb...@rscusa.com wrote: One thing I would like to see is, where possible, taking advantage of SSE(2/3) instructions on CPU's that have them. I imagine that there could be some speed up, especially if operations could be batched (I'm sure a whole other can-o-worms). For apps that are limited by math performance this would be useful, but for most scene graph applications the matrix maths is never anything close to a bottleneck. We are mostly bandwidth limited, and what CPU overheads we see are largely down to cache misses as the scene graph churns through lots of data in memory. So stuff like SSE optimizations are really the for physics engines and alike rather than scene graphs. Not to mention that SSE and such are actually useful once you have more data to process than only two 4x4 matrices - the overhead of setting SSE up for the 16 values is probably higher than the benefit you could gain. Video decoding on the other hand is a completely different story. Now on a tangent a bit, but as we are discussing code design here: is there a particular reason why the osg::PositionAttitudeTransform is not a subclass of osg::MatrixTransform? It seems rather illogical and makes for messy problems when trying to write generic code dealing with scenegraphs that may contain either type of transform node. As they are not derived from one another, the code needs to check which one it is each time and handle them specially. Ugly, messy and bug prone. Regards, Jan -BEGIN PGP SIGNATURE- Version: GnuPG v1.4.10 (GNU/Linux) Comment: Using GnuPG with Mandriva - http://enigmail.mozdev.org/ iD8DBQFMwl17n11XseNj94gRAnqdAKChA/N4uWwNu/2Kirc4fGS6aj63ogCeKZck MmvBMQWrt8m9Mt0dHMMEVVY= =jCue -END PGP SIGNATURE- ___ osg-users mailing list osg-users@lists.openscenegraph.org http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org
