<?xml version="1.0" encoding="iso-8859-1" ?>
<rss version="2.0">
<channel>
	<title>New Posts</title>
	<description>All New Posts on the NVIDIA Developer Forum</description>
	<link>http://developer.nvidia.com/forums/index.php</link>
	<pubDate>Sun, 22 Nov 2009 23:15:26 -0800</pubDate>
	<ttl>15</ttl>
	<item>
		<title>Crash under PerfHUD while calling D3DXCompileShader</title>
		<link>http://developer.nvidia.com/forums/index.php?showtopic=3891</link>
		<description><![CDATA[While under PerfHUD calls our engine makes to D3DXCreateShader crash inside perfhud code with a NULL pointer dereference.<br /><br />We have tried this on multiple machines with the same result.<br /><br />Are you aware of any issues that might cause this?<br /><br />I can provide an example of shader code if required but it is fairly convoluted due to being generated by linking fragments of shader code together.]]></description>
		<pubDate>Sun, 22 Nov 2009 19:37:42 -0800</pubDate>
		<guid>http://developer.nvidia.com/forums/index.php?showtopic=3891</guid>
	</item>
	<item>
		<title>CAnt save as a dds in photoshop 7</title>
		<link>http://developer.nvidia.com/forums/index.php?showtopic=3890</link>
		<description><![CDATA[Hey all when i try and save as a dds file using the nvidia plugin in photoshop it comes up with a cant save filename as a dds because of a programme error can anyone help<br /><br />CHeers JZC]]></description>
		<pubDate>Fri, 20 Nov 2009 14:25:22 -0800</pubDate>
		<guid>http://developer.nvidia.com/forums/index.php?showtopic=3890</guid>
	</item>
	<item>
		<title>How download PhysX 2.8.3 on Windows ?</title>
		<link>http://developer.nvidia.com/forums/index.php?showtopic=3889</link>
		<description><![CDATA[How download PhysX 2.8.3 on Windows ? <img src="http://developer.nvidia.com/forums/style_emoticons/default/rolleyes.gif" style="vertical-align:middle" emoid=":rolleyes:" border="0" alt="rolleyes.gif" />]]></description>
		<pubDate>Fri, 20 Nov 2009 08:32:11 -0800</pubDate>
		<guid>http://developer.nvidia.com/forums/index.php?showtopic=3889</guid>
	</item>
	<item>
		<title>Adding Joints to Vehicles?</title>
		<link>http://developer.nvidia.com/forums/index.php?showtopic=3888</link>
		<description><![CDATA[Im using a Wrapper for the PhysX SDK with the Dark Basic Professional language, so Im not sure if this is an issue with my language of choice or the SDK, although from my tests it appears to be the SDK itself. <br /><br />I was wondering why its not possible to add joints between two vehicles, allowing me to create a truck vehicle and a trailer vehicle and attach them by a Sphere Joint having the trailer then follow behind the truck as it drives? To me it seems perfectly logical, a sphere joint to allow the vehicle to rotate on all axes', with it's thrust being applied by the truck vehicle. <br /><br />Is this impossible using the PhysX SDK?<br /><br />I have tried other options as well. Currently Im experimenting by creating a kinematic physics body ( a dummy object basically ) and positioning it behind my vehicle at all times, and then sphere jointing this dummy object to a large box rigid body to simulate the trailer. It works as you would expect it to, the box tumbles around behind the vehicle because it doesnt have wheels. I could add joints to some cylinders bellow the trailer body I suppose, to keep it rolling, but Im just wondering if there is another way to achieve what Im after.]]></description>
		<pubDate>Fri, 20 Nov 2009 08:22:13 -0800</pubDate>
		<guid>http://developer.nvidia.com/forums/index.php?showtopic=3888</guid>
	</item>
	<item>
		<title>NxConvexMeshDesc points and triangles attributes</title>
		<link>http://developer.nvidia.com/forums/index.php?showtopic=3887</link>
		<description><![CDATA[What is the exactly order in which the <b>points </b>( a pointer attribute of NxConvexMeshDesc class) is returned? It's differs from the order in which i have defined previously NxConvexMesh]]></description>
		<pubDate>Fri, 20 Nov 2009 08:10:36 -0800</pubDate>
		<guid>http://developer.nvidia.com/forums/index.php?showtopic=3887</guid>
	</item>
	<item>
		<title>Bug in 186.81 driver for notebooks?</title>
		<link>http://developer.nvidia.com/forums/index.php?showtopic=3886</link>
		<description><![CDATA[Apparently when I execute:<br /><br />m_pD3D10Device-&gt;ClearDepthStencilView( m_pCurrentDepthStencil, D3D10_CLEAR_STENCIL, 1.0f, i_nStencilValue );<br /><br />It also writes 1.0f to the depth buffer.<br /><br />I have an 8600M GT 512MB and using 186.81 drivers for notebooks.<br /><br />Can you confirm it's a driver bug or I'm just missing something?<br /><br />For those having 186.81, if you want to do a simple test, just add this command at the end of your "clear" method.<br />ClearDepthStencilView( m_pCurrentDepthStencil, D3D10_CLEAR_STENCIL, 0.0f, 0);<br /><br />If you're testing depth with less_equal, this should result in no output as the function erroneously clears the depth buffer.<br />If you're testing depth with greater_equal, use this:<br />ClearDepthStencilView( m_pCurrentDepthStencil, D3D10_CLEAR_STENCIL, 1.0f, 0);<br /><br />If the operation is working the rendering shouldn't be affected as the stencil buffer will just be set to 0 and the depth should be untouched.<br /><br />Thank you in advance!<br />]]></description>
		<pubDate>Fri, 20 Nov 2009 05:17:37 -0800</pubDate>
		<guid>http://developer.nvidia.com/forums/index.php?showtopic=3886</guid>
	</item>
	<item>
		<title>Bug in Physx plug-in for Maya 2008</title>
		<link>http://developer.nvidia.com/forums/index.php?showtopic=3885</link>
		<description><![CDATA[Hi <br />I am trying to build a collapsing building in Maya2008. And upon active/passive keyframing of rigid bodies, created using Physx, in the first cycle the simulation runs fine. But in the the subsequent cycles it doesnt start the animation of this body from the initial position but from the position the rigid body last held at the end of the previous simulation. <br /><br />On restarting maya, again the simulation runs fine in the first cycle but again subsequent cycles give the same problem.<br /><br />A request to the developers: Please fix the problem in the Plugin for Maya2008 and in the alpha, beta release of the plugin for maya2009...<br /><br />Awating a reponse from the developers if this could be fixed soon.<br />]]></description>
		<pubDate>Fri, 20 Nov 2009 03:59:15 -0800</pubDate>
		<guid>http://developer.nvidia.com/forums/index.php?showtopic=3885</guid>
	</item>
	<item>
		<title>PhysX 2.8.1 and PhysX 2.8.3 on Linux</title>
		<link>http://developer.nvidia.com/forums/index.php?showtopic=3884</link>
		<description><![CDATA[Hi, <br /><br />Where can I find PhysX 2.8.1 and/or PhysX 2.8.3 on Linux both built for 32 bit and 64 bit? <br /><br />I'm working debian Lenny using 64 bit. <br /><br />Cheers, <br /><br />Ole J.]]></description>
		<pubDate>Thu, 19 Nov 2009 03:12:45 -0800</pubDate>
		<guid>http://developer.nvidia.com/forums/index.php?showtopic=3884</guid>
	</item>
	<item>
		<title>Clarification on installation path</title>
		<link>http://developer.nvidia.com/forums/index.php?showtopic=3883</link>
		<description><![CDATA[Hello All,<br /><br />When I install the PhysX System Software, I donot get any option to select the folder where the files should go.<br />The installer creates C:&#092;Program Files&#092;Ageia Technologies&#092;.....<br /><br />When I install PhysX SDK, I get the option to change the path.<br /><br />I wish to install everything (SDK and the PhysX System Software) in D:&#092;Program Files&#092;.... so that in future if I need to reinstall the OS, I will not loose any of<br />my stuff.<br /><br />Can you please let me know, how I can get everything related to PhysX installed in D: ?<br /><br />Thanks.<br /><br />Regards,<br />Balaji]]></description>
		<pubDate>Thu, 19 Nov 2009 02:04:58 -0800</pubDate>
		<guid>http://developer.nvidia.com/forums/index.php?showtopic=3883</guid>
	</item>
	<item>
		<title>Does anyone have the PhysX  2.3. 2 plugin for Max8 ?</title>
		<link>http://developer.nvidia.com/forums/index.php?showtopic=3882</link>
		<description><![CDATA[I am using Gamebryo 2.2 , and it supports PhsX 2.3.2 only( what an old version!) <br />I tried hard to find the<u> PhysX  2.3. 2 plugin for Max8</u>, but got nothing. <br />Can anyone provide the plugin? <img src="http://developer.nvidia.com/forums/style_emoticons/default/sad.gif" style="vertical-align:middle" emoid=":(" border="0" alt="sad.gif" />]]></description>
		<pubDate>Wed, 18 Nov 2009 23:19:52 -0800</pubDate>
		<guid>http://developer.nvidia.com/forums/index.php?showtopic=3882</guid>
	</item>
	<item>
		<title>Visual Remote Debugger Hang</title>
		<link>http://developer.nvidia.com/forums/index.php?showtopic=3881</link>
		<description><![CDATA[Hi all,<br /><br />I face a weird problem. My Visual Remote Debugger (VRD) stop showing stimulation result. <img src="http://developer.nvidia.com/forums/style_emoticons/default/sad.gif" style="vertical-align:middle" emoid=":(" border="0" alt="sad.gif" /><br /><br />Previously my VRD is running correctly until just yesterday when it doesn't display stimulation result. The level still manage to load correctly.<br /><br />In fact, when VRD is running my stimulation hang.<br /><br />I revert to code base where I'm sure it worked before but it still hangs. -.-'<br /><br />I tried and search everywher to no avail.<br /><br />Anyone have faced and solved this problem? Thanks in advance. <img src="http://developer.nvidia.com/forums/style_emoticons/default/smile.gif" style="vertical-align:middle" emoid=":)" border="0" alt="smile.gif" />]]></description>
		<pubDate>Wed, 18 Nov 2009 20:47:46 -0800</pubDate>
		<guid>http://developer.nvidia.com/forums/index.php?showtopic=3881</guid>
	</item>
	<item>
		<title>Request for Blender3d plugin</title>
		<link>http://developer.nvidia.com/forums/index.php?showtopic=3880</link>
		<description>I would like to request a Blender3d plugin be released. Currently there is only support for Softimage, 3dsMax, and Maya, which while the industry standards, are very expensive. In the current economic situation, most Indy, open source, and mod developers can not afford these kinds of tools. It would be greatly appreciated for Blender3d tools that work on both Linux and Windows (and mac if a mac version of PhysX is released) which would allow for more financially limited developers to use PhysX in their apps.</description>
		<pubDate>Wed, 18 Nov 2009 19:55:56 -0800</pubDate>
		<guid>http://developer.nvidia.com/forums/index.php?showtopic=3880</guid>
	</item>
	<item>
		<title><![CDATA[NX_NOTIFY_ON_ROLL - Has this been implemented yet as I can't get it to work?]]></title>
		<link>http://developer.nvidia.com/forums/index.php?showtopic=3879</link>
		<description><![CDATA[Hi guys,<br /><br />Can anyone tell me the status of NX_NOTIFY_ON_ROLL, as I can't seem to get it to flag with my contacts.<br /><br />Thank you.]]></description>
		<pubDate>Wed, 18 Nov 2009 13:51:11 -0800</pubDate>
		<guid>http://developer.nvidia.com/forums/index.php?showtopic=3879</guid>
	</item>
	<item>
		<title>cg tutorial examples</title>
		<link>http://developer.nvidia.com/forums/index.php?showtopic=3878</link>
		<description>i ve download and installed the cg tutorial examples and of course the cg compiler . ( see my post) i take a look on the examples but some of them do not open . the application crash. i ve red the doc included. i have a Nvidia Geforce 8400m gt, is this card supported for the examples? thanks</description>
		<pubDate>Wed, 18 Nov 2009 12:52:41 -0800</pubDate>
		<guid>http://developer.nvidia.com/forums/index.php?showtopic=3878</guid>
	</item>
	<item>
		<title>Problem with attachVertexToShape</title>
		<link>http://developer.nvidia.com/forums/index.php?showtopic=3877</link>
		<description><![CDATA[Hello,<br /><br />i tried to use the attachVertexToShape method for a softbody. I was able to attach serveral vertices to a kinematic shape. Everythinks works like expectd. But there seems to be a problem with some vertices. Using the same code, but attaching e.g vertex i+1 (that is existing) instead of vertex i (which is doing fine) leads to chaotic simulation. Some vertices fly away from the softbody and pull it with them. The softbody will not come to rest. <br /><br />What could be the problem?<br /><br />Thanks,<br />Ilja]]></description>
		<pubDate>Wed, 18 Nov 2009 08:58:08 -0800</pubDate>
		<guid>http://developer.nvidia.com/forums/index.php?showtopic=3877</guid>
	</item>
	<item>
		<title>Motor velocity target implemented as a constraint</title>
		<link>http://developer.nvidia.com/forums/index.php?showtopic=3876</link>
		<description><![CDATA[In PhysX if you attach a joint to an actor and set a velocity target in the motor it will work as you would expect.  The problem comes in when external forces resist the motor.  The motor will only ever exert the amount force it would take to bring <i>the mass of the actor it's attached to</i> up to speed.  It does not scale up the force to account for other actors attached, collisions or gravity.  This means that it will likely never get to the desired speed even though it is not exerting all its alloted force. <br />This implementation is only useful in very simple and controlled simulations.  It amounts to little more than a simple impulse applied scaled by the mass of the attached actor.<br /> <br />Implementing the velocity target as a constraint in the solver (ODE does this btw) the joint is equally effective in the simple case, but it can essentually look ahead one timestep and see how fast the actor is <i>going to be moving</i> after the step and apply more/less force as needed.  This allows a solver-constraint based motor to effectivly and accuratly counter-act forces like gravity or collisions with other bodies.  The amount of force applied is based on the actual resulting velocity of the actor, not just a simple impulse to the body scaled for its mass. <br /><br />using joint motors to get believable motion from a sophisticated articulated character with the current motor options is just not possible.  Spring/damper only works in controlled situations with tons of tuning, velocity targets are implemented to simplisticly.  PhysX needs a way to add constraints to the solver like this to make it more practical for use outside the basic debris, barrels and dead character ragdolls.<br /><br />Does this make sense?  Is there a way to add constraints to the solver myself to workaround the lack of this feature, or some other solution?]]></description>
		<pubDate>Wed, 18 Nov 2009 07:09:58 -0800</pubDate>
		<guid>http://developer.nvidia.com/forums/index.php?showtopic=3876</guid>
	</item>
	<item>
		<title>problems with examples</title>
		<link>http://developer.nvidia.com/forums/index.php?showtopic=3875</link>
		<description><![CDATA[hi! i've tried to use the basic example for d3d9. i compile it. no errors, all right. when i try to debug i receive this error"<!--c1--><div class='codetop'>CODE</div><div class='codemain'><!--ec1-->'01_vertex_program.exe'&#58; caricato 'C&#58;&#092;Program Files&#092;NVIDIA Corporation&#092;Cg&#092;examples&#092;Direct3D9&#092;basic&#92;&#48;1_vertex_program&#92;&#48;1_vertex_program.exe'<br />'01_vertex_program.exe'&#58; caricato 'C&#58;&#092;Windows&#092;System32&#092;ntdll.dll', Esportazioni caricate.<br />'01_vertex_program.exe'&#58; caricato 'C&#58;&#092;Windows&#092;System32&#092;kernel32.dll', Esportazioni caricate.<br />'01_vertex_program.exe'&#58; caricato 'C&#58;&#092;Program Files&#092;Autodesk&#092;Maya2008&#092;bin&#092;cg.dll', Esportazioni caricate.<br />'01_vertex_program.exe'&#58; caricato 'C&#58;&#092;Program Files&#092;NVIDIA Corporation&#092;Cg&#092;bin&#092;cgD3D9.dll', Esportazioni caricate.<br />Debugger&#58;&#58; durante il caricamento del processo è stata generata un'eccezione critica non gestita.<br />Il programma '&#91;2952&#93; 01_vertex_program.exe&#58; Nativo' è terminato con il codice -1073741511 &#40;0xc0000139&#41;.<!--c2--></div><!--ec2--><br />what's the problem? <img src="http://developer.nvidia.com/forums/style_emoticons/default/dry.gif" style="vertical-align:middle" emoid="&lt;_&lt;" border="0" alt="dry.gif" /> <br />thanks  <img src="http://developer.nvidia.com/forums/style_emoticons/default/smile.gif" style="vertical-align:middle" emoid=":)" border="0" alt="smile.gif" />]]></description>
		<pubDate>Tue, 17 Nov 2009 13:03:04 -0800</pubDate>
		<guid>http://developer.nvidia.com/forums/index.php?showtopic=3875</guid>
	</item>
	<item>
		<title>Scaling Question</title>
		<link>http://developer.nvidia.com/forums/index.php?showtopic=3874</link>
		<description><![CDATA[Hello,<br /><br />we use our Graphic Engine with the scaling 100:1. For Example something that is 1m = 100cm in the real world we use 100 units in the engine.<br /><br />The physic system must now be in the same scale which means speeds are a bit higher (factor 100). Also other values have to be adjusted like the various SDK Parameters and other parameters at the entity.<br /><br />My Question is now if someone had ever Issues or experience with another scaling then 1:1 and PhysX.<br />Another more specific question is about the SDK Parameter NX_DYN_FRICT_SCALING and NX_STA_FRICT_SCALING. Do I have to set these to 100?<br /><br />greetings]]></description>
		<pubDate>Tue, 17 Nov 2009 07:48:28 -0800</pubDate>
		<guid>http://developer.nvidia.com/forums/index.php?showtopic=3874</guid>
	</item>
	<item>
		<title>Physx Plugin x64 - a small wish</title>
		<link>http://developer.nvidia.com/forums/index.php?showtopic=3873</link>
		<description><![CDATA[Hello<br /><br />playing around with the new beta plugin for max 64...works good for now but I have a little wish:<br /><br />Multiple Selection in the Meshlist (physxmodRB modifier).<br /><br />It would be really awesome if there would be such a feature...selecting many Items in the Meshlist and changing the properties of the selected mass.<br />Maybe this feature will be added with the modifier TAB of the missing Physx modify Panel in the Physx Toolbar...but I just wanted to ask because its a very important missing feature.<br /><br />Thanks.<br /><br />PS: I could not find the shared example scenes...are there any ones&#96;?]]></description>
		<pubDate>Tue, 17 Nov 2009 06:57:38 -0800</pubDate>
		<guid>http://developer.nvidia.com/forums/index.php?showtopic=3873</guid>
	</item>
	<item>
		<title>Why does changing the dt for simulation not have any effect in Lesson101?</title>
		<link>http://developer.nvidia.com/forums/index.php?showtopic=3872</link>
		<description><![CDATA[Hi folks,<br /><br />Just wondering why if I change gDeltaTime in the simulate function to any other number, it makes no difference?<br /><br />gScene-&gt;simulate(gDeltaTime); // changed to n<br /><br />Thanks]]></description>
		<pubDate>Mon, 16 Nov 2009 06:26:31 -0800</pubDate>
		<guid>http://developer.nvidia.com/forums/index.php?showtopic=3872</guid>
	</item>
	<item>
		<title>Maya Plugin Issues: SCV causes crash/warning</title>
		<link>http://developer.nvidia.com/forums/index.php?showtopic=3871</link>
		<description><![CDATA[I'm trying to create a ragdoll in Maya 2008 using the PhysX plugin.  I'm using the orc_painted_skin.ma that comes with the plugin and the following the user guide.<br />When I try to create the SCV, Maya crashes.  I tried seeing what would happen if I deleted a few joints from the skeleton (specifically the joints in both hands) and instead of crashing, Maya gave a warning that says:<br />// Warning: SkinCollisionVolumeCmd::createHullMesh() - hullNum (28) exceeds dimensions of joints array (28).<br /><br />I tried to create a ragdoll using my own skinned model and I got the same warning.  Has anyone encountered this error before? What could be causing it?<br />Also, is there another method I could use to make a ragdoll if I can't get this to work?]]></description>
		<pubDate>Sun, 15 Nov 2009 20:13:07 -0800</pubDate>
		<guid>http://developer.nvidia.com/forums/index.php?showtopic=3871</guid>
	</item>
	<item>
		<title>Apply force to rear wheels</title>
		<link>http://developer.nvidia.com/forums/index.php?showtopic=3870</link>
		<description>Whats the best way to apply a force to the wheels?</description>
		<pubDate>Sat, 14 Nov 2009 16:38:48 -0800</pubDate>
		<guid>http://developer.nvidia.com/forums/index.php?showtopic=3870</guid>
	</item>
	<item>
		<title>drawing wire shapes for wrong actor</title>
		<link>http://developer.nvidia.com/forums/index.php?showtopic=3869</link>
		<description><![CDATA[I am using this code from the training programs shipped with PhysX:<br /><br />NxShape*const* shapes;<br />shapes = actor-&gt;getShapes();<br />NxU32 nShapes = actor-&gt;getNbShapes();<br />if (actor == selectedActor) <br />{<br />while (nShapes--)<br />{<br />DrawWireShape(shapes[nShapes], NxVec3(1,1,1), useShapeUserData);<br />}<br />}<br /><br />This i supposed to draw wire shapes around the selected actor. In fact it draws wires at the location of the selected actor, but with shapes corresponding to the last actor created on the scene. For example, if I create a box and set it as the selected actor, a box wire shape will be rendered around the box. But if I create a box and a capsule and select the box, a *capsule* wire shape will be rendered around the *box*. As far as I can tell, the function DrawWireShapes simply calls gl primitives to draw lines from the points and triangles of shapes[nShapes] so there is no mystery there. Any ideas why this is happening?<br />Thanks,<br />-nuun]]></description>
		<pubDate>Fri, 13 Nov 2009 14:40:28 -0800</pubDate>
		<guid>http://developer.nvidia.com/forums/index.php?showtopic=3869</guid>
	</item>
	<item>
		<title>PhysX for 3ds Max beta plug-in available</title>
		<link>http://developer.nvidia.com/forums/index.php?showtopic=3868</link>
		<description><![CDATA[The extranet we want to use to distribute the beta PhysX Plug-in for 3ds Max is taking longer than expected to finalize. However, I'd love to get beta testing underway sooner rather than later. To that end, if you would like to be a beta tester for the plug-in (supporting Max 2008, 2009, 2010 on 32- and 64-bit OS), please send me an email. I'll get you the plug-in if you will in turn give me feedback.<br /><br />Send an email with subject "PhysX Plug-in for 3ds Max Beta" to "gkistner@nvidia.com", and include your real name, email address, and company that you'll be using the plug-in for. Optionally, I'd also love to hear what you intend to use the plug-in for: content development for PhysX-based games; developing for other real-time simulations; improved simulation for standalone rendering; or just repeatedly smashing wrecking balls into brick silos and watching them fall.  <img src="http://developer.nvidia.com/forums/style_emoticons/default/smile.gif" style="vertical-align:middle" emoid=":)" border="0" alt="smile.gif" /><br /><br />The best way to make the plug-ins better is to test them and send us feedback, so hop to it and flood my inbox.]]></description>
		<pubDate>Fri, 13 Nov 2009 08:43:52 -0800</pubDate>
		<guid>http://developer.nvidia.com/forums/index.php?showtopic=3868</guid>
	</item>
	<item>
		<title>no wireframe</title>
		<link>http://developer.nvidia.com/forums/index.php?showtopic=3867</link>
		<description><![CDATA[In the NVIDIA Physx tutorials , from the beginning, among the things which i suppose to learn, are A HUNDREDS of things over which i just cannot step over....<br />One of these are the follow:<br />the SELECTED objects are in wireframe mode....<br />How i can maintain the selection of the actors BUT DISABLE the wireframe mode OF SELECTED OBJECT :....]]></description>
		<pubDate>Fri, 13 Nov 2009 07:56:13 -0800</pubDate>
		<guid>http://developer.nvidia.com/forums/index.php?showtopic=3867</guid>
	</item>
	<item>
		<title>PhysX Issue</title>
		<link>http://developer.nvidia.com/forums/index.php?showtopic=3866</link>
		<description><![CDATA[Hi, I am trying to create a bounding box for a gutter in a bowling lane but it creates a regular box when i do create hull. How can I create it for this kind of shape.<br /><br />Thank you,<br />Lucas]]></description>
		<pubDate>Thu, 12 Nov 2009 20:21:33 -0800</pubDate>
		<guid>http://developer.nvidia.com/forums/index.php?showtopic=3866</guid>
	</item>
	<item>
		<title>crash with NxTriangleMesh</title>
		<link>http://developer.nvidia.com/forums/index.php?showtopic=3864</link>
		<description><![CDATA[I created some NxtriangleMeshes in my scene. These meshes work well when i set them to static. But if I set them to dynamic, i have a crash in PhysXCore.dll. The crash doesn't happen in the first frame , but after 1 or 2 second (it seems that the crash happens when we get a TriangleMesh vs TriangleMesh collision)<br /><br />Someone have a idea about it ?<br /><br />Thx<br /><br />Error message : <br /><br />Unhandled exception at 0x02956fa3 in MyApplication.exe: 0xC0000005: Access violation writing location 0xde7832e0.<br /><br />IDE : Microsoft Visual Studio 2008]]></description>
		<pubDate>Thu, 12 Nov 2009 11:22:27 -0800</pubDate>
		<guid>http://developer.nvidia.com/forums/index.php?showtopic=3864</guid>
	</item>
	<item>
		<title><![CDATA[Triggers & Kinematics]]></title>
		<link>http://developer.nvidia.com/forums/index.php?showtopic=3863</link>
		<description><![CDATA[I have my world made up by Kinematic shapes (Capsules), each Tick() I will manually handle collision detection & response using linearSweeps & setGlobalPosition on the kinematics... This works OK, until I add a trigger to the scene (static trigger), once this trigger is added the whole system<br />starts to behave strange in a number of ways:<br /><br />1. When my kinematics are solving detection & response within a trigger volume the behavior will not be correct, my character can get stuck midair, my character cannot even move from its original position sometimes. If I remove the trigger however, everything is back to the normal and working state, so I'm 100% sure the cause has something to do with triggers.<br /><br />2. The trigger report does not fire fore my kinematics at all, BUT if I add a dynamic entity to the scene the trigger will invoke the callback whenever it enters the volume.<br /><br />When reading the docs it appears to me that simply creating the kinematic is enough to make it interact with triggers, maybe this is not the case?<br />Also, it seems like any sweep test / overlap test will ignore triggers, thus it should not mess up the collision detection / response phase I'm currently using.<br /><br />I dont know where to start looking for the solution so if anyone knows something that could shed some light on my problem Id be very grateful.<br /><br />Thanks.<br /><br />If it helps, here is how I create my kinematic actors and my trigger actor:<br /><br />//<br />// KINEMATIC:<br />//<br />NxBodyDesc bodyDesc;<br />bodyDesc.mass = 1.0f;<br />bodyDesc.flags = NX_BF_KINEMATIC;<br />EE_ASSERT(bodyDesc.isValid());<br /><br />NxCapsuleShapeDesc shapeDesc;<br />shapeDesc.group = 0;<br />shapeDesc.radius = radius;<br />shapeDesc.height = height - (radius * 2.0f);<br /><br />efd::Point3 offset = efd::Point3::ZERO;<br /><br />efd::Matrix3 rot;<br />rot.FromEulerAnglesXYZ(NI_HALF_PI, 0, 0);<br />offset = efd::Point3(0, 0, height * 0.5f);<br /><br />NxMat34 nxPose;<br />efdPhysX::PhysXTypes::MatrixPointToNxMat34(rot, offset, nxPose);<br />shapeDesc.localPose = nxPose;<br />EE_ASSERT(shapeDesc.isValid());<br />		<br />NxActorDesc actorDesc;<br />actorDesc.body = &bodyDesc;<br />actorDesc.shapes.push_back(&shapeDesc);<br />EE_ASSERT(actorDesc.isValid());<br /><br />efdPhysX::PhysXSDKManager::GetManager()-&gt;WaitSDKLock();<br />NxActor* pNxActor = pNxScene-&gt;createActor(actorDesc);<br />EE_ASSERT(pNxActor);<br />pNxActor-&gt;userData = this;<br />efdPhysX::PhysXSDKManager::GetManager()-&gt;ReleaseSDKLock();<br /><br />//<br />// TRIGGER:<br />//<br />NxBoxShapeDesc shapeDesc;<br />shapeDesc.group = 0;<br />shapeDesc.dimensions.x = m_dimensions.x * 0.5f;<br />shapeDesc.dimensions.y = m_dimensions.y * 0.5f;<br />shapeDesc.dimensions.z = m_dimensions.z * 0.5f;<br />shapeDesc.userData = this;<br />		<br />shapeDesc.shapeFlags |= NX_TRIGGER_ENABLE;<br /><br />EE_ASSERT(shapeDesc.isValid());<br /><br />NxVec3 p;<br />efdPhysX::PhysXTypes::Point3ToNxVec3(GetPosition(), p);<br /><br />p.z += shapeDesc.dimensions.z;<br />		<br />NxActorDesc actorDesc;<br />actorDesc.globalPose.t = p;<br />actorDesc.shapes.push_back(&shapeDesc);<br />EE_ASSERT(actorDesc.isValid());<br /><br />efdPhysX::PhysXSDKManager::GetManager()-&gt;WaitSDKLock();<br />NxActor* pNxActor = pNxScene-&gt;createActor(actorDesc);<br />EE_ASSERT(pNxActor);<br />pNxActor-&gt;userData = this;<br />efdPhysX::PhysXSDKManager::GetManager()-&gt;ReleaseSDKLock();]]></description>
		<pubDate>Thu, 12 Nov 2009 07:17:03 -0800</pubDate>
		<guid>http://developer.nvidia.com/forums/index.php?showtopic=3863</guid>
	</item>
	<item>
		<title>FX Composer as Collada viewer (urgent request)</title>
		<link>http://developer.nvidia.com/forums/index.php?showtopic=3862</link>
		<description><![CDATA[Hi all,<br /><br />I'm new to the tool and most of the shading world but I need to visualize multiple Collada files and I cannot seem to find a proper tool for this.<br />Trying FX C it is actually the first tool that enables me to load dae + textures and I'm very excited about this.<br /><br />Now I would like to brighten the scene.  As I have a ton of Collada files to check - each files represents a part of a 3D world with buildings, roads and land - I would like to change some default values to FX C that increase/change the lighting.<br /><br />My target platform only uses OpenGL 1.1 (mobile devices mainly) so I don't require the fancy lighting, I just want to ensure my full scene is lit so I can check overall texture quality and look&feel.<br /><br />Can anyone assist me in this, it is rather urgent as I need to start checking this week.  Is there a script available that allows me to add a light which is then automatically assigned to all materials?<br /><br />thanks,<br />3DC]]></description>
		<pubDate>Thu, 12 Nov 2009 03:09:01 -0800</pubDate>
		<guid>http://developer.nvidia.com/forums/index.php?showtopic=3862</guid>
	</item>
	<item>
		<title>PhysX Issue</title>
		<link>http://developer.nvidia.com/forums/index.php?showtopic=3861</link>
		<description><![CDATA[Hi, I installed the PhysX SDK to be able to create a bowling game. When I try to open the Lessons it tells me that it has encountered a problem and i tneeds to be closed. Also, when I use the plug in for 3DS Max, it gives me a warning that says that the SDK could not be created and to make sure I installed the correct version. I checked everything, I re-intalled everything a lot of times. I've created an Install Path in the Environment variables as well. Any help?<br /><br />thank you,<br /><br />Lucas Sanjuan]]></description>
		<pubDate>Wed, 11 Nov 2009 15:37:35 -0800</pubDate>
		<guid>http://developer.nvidia.com/forums/index.php?showtopic=3861</guid>
	</item>
	<item>
		<title>NxHeightfield creation</title>
		<link>http://developer.nvidia.com/forums/index.php?showtopic=3860</link>
		<description><![CDATA[Greetings. Here I've got such a problem: i'm trying to create a Heightfield based on heightmap image. Everything works almoust fine with NxTriangleMesh (i simply got 30 sec loading program and 256*512 instead 512*512 terrain). But when i try to use NxHeightField it compiles as well, but crashes when i run it in couple with RemoteDebugger... <br /><br /> I've tried a lot of variants, but nothing (and no one) helped me  <img src="http://developer.nvidia.com/forums/style_emoticons/default/sad.gif" style="vertical-align:middle" emoid=":(" border="0" alt="sad.gif" /> <br /><br /> Here are my source codes:<br /><br />*common part of code. works just fine*<br /><br /><div class='codetop'>CODE</div><div class='codemain' style='height:200px;white-space:pre;overflow:auto'><br />Ogre::ConfigFile config;<br />	String value;<br /><br />	config.load(filename);<br /><br />	Vector3 dimensions = Vector3::UNIT_SCALE;<br />    <br />	value = config.getSetting( "PageWorldX" );<br />    <br />	if ( ! value.empty() )<br />		dimensions.x = atof( value.c_str() );<br />    <br />	value = config.getSetting( "MaxHeight" );<br />            <br />	if ( ! value.empty() )<br />		dimensions.y = atof( value.c_str() );<br />    <br />	value = config.getSetting( "PageWorldZ" );<br />    <br />	if ( ! value.empty() )<br />		dimensions.z = atof( value.c_str() );<br />    <br />	float pageSize, tileSize;<br />    <br />	value = config.getSetting( "PageSize" );<br />    <br />	if ( ! value.empty() )<br />		pageSize = atoi( value.c_str() );<br />	else<br />		LogManager::getSingleton().logMessage("Terrain says: Missing option &#092;"PageSize&#092;" in terrain config!");<br /><br />	value = config.getSetting( "TileSize" );<br /><br />	if ( ! value.empty() )<br />		tileSize = atoi( value.c_str() );<br />	else<br />		LogManager::getSingleton().logMessage("Terrain says: Missing option &#092;"TileSize&#092;" in terrain config!");<br />    <br />	dimensions.x /= pageSize - 1;<br />	dimensions.z /= pageSize - 1;<br />   <br />	String imageFilename = config.getSetting( "Heightmap.image" );<br />    <br />	if (imageFilename.empty())<br />		LogManager::getSingleton().logMessage("Terrain says: No image for heightField!"); else<br />			LogManager::getSingleton().logMessage("Terrain says: Loading image: " + imageFilename);<br />	<br />    // current variant of loading heightfield<br />	Image image;<br />    <br />	image.load(imageFilename, ResourceGroupManager::DEFAULT_RESOURCE_GROUP_NAME);<br />    <br />	size_t sizeX = image.getWidth();<br />	size_t sizeZ = image.getHeight();<br /><br />	size_t numHeightfieldVerts = sizeX * sizeZ;<br />	size_t numHeightfieldFaces = ( sizeX + 1 ) * ( sizeZ + 1 );<br /><br />	Real invScale;<br />    <br />	PixelFormat pf = image.getFormat();<br />    <br />	bool is16bit = (pf == Ogre::PF_L16);<br />    <br />	if( pf != Ogre::PF_L8 && pf != Ogre::PF_L16 )<br />	{<br />		LogManager::getSingleton().logMessage("Terrain says: HeightMap image is not a grayscale one!");<br />	}<br /><br />	LogManager::getSingleton().logMessage("Terrain says: My size is " + StringConverter::toString(sizeX) + "x" + StringConverter::toString(sizeZ));<br />    <br />	if ( is16bit )<br />	{<br />		invScale = 1.0f / 65535.0f;<br />	}<br />	else<br />	{<br />		invScale = 1.0f / 255.0f;<br />	}<br />   <br />    uchar* pSrc = image.getData();<br /><br />	LogManager::getSingleton().logMessage("Terrain says: My heightmap size is " + StringConverter::toString(sizeof(pSrc)));<br /></div><br /><br />*half-working Trimesh*<br /><div class='codetop'>CODE</div><div class='codemain' style='height:200px;white-space:pre;overflow:auto'>NxVec3* gTerrainVerts = new NxVec3[numHeightfieldVerts];<br />	NxU32 y, numVerts = 0;<br /><br />	for(y = 0; y &lt; sizeX; y++)<br />	{<br />		for(NxU32 x = 0; x &lt; sizeX; x++)<br />		{<br />			//gTerrainVerts[x + (y * sizeZ)] = NxVec3(x, pSrc[x + (y * sizeX)] * invScale, y);<br />			gTerrainVerts[numVerts++] = NxVec3(x, pSrc[numVerts] / (image.getBPP() * 2), y);<br />		}<br />	}<br /><br />	LogManager::getSingleton().logMessage("Terrain says: My Y-scale is " + StringConverter::toString(invScale));<br /><br />	// Initialize terrain faces<br />	NxU32* gTerrainFaces = new NxU32[numHeightfieldFaces * 6];<br />	NxU32 k = 0;<br /><br />	for(NxU32 j = 0; j &lt; sizeX - 1; j++)<br />	{<br />		for(NxU32 i = 0; i &lt; sizeX - 1; i++)<br />		{<br />			// Create first triangle<br />			gTerrainFaces[k++] = i   + j * sizeX;<br />			gTerrainFaces[k++] = i   + (j + 1) * sizeX;<br />			gTerrainFaces[k++] = i + 1 + (j + 1) * sizeX;<br /><br />			// Create second triangle<br />			gTerrainFaces[k++] = i   + j * sizeX;<br />			gTerrainFaces[k++] = i + 1 + (j + 1) * sizeX;<br />			gTerrainFaces[k++] = i + 1 + j * sizeX;<br />		}<br />	}<br /><br />	// Build physical model<br />	NxTriangleMeshDesc terrainDesc;<br />	//terrainDesc.numVertices					= numHeightfieldVerts;<br />	terrainDesc.numVertices					= numVerts - 1;<br />	terrainDesc.numTriangles				= numHeightfieldFaces;<br />	terrainDesc.pointStrideBytes			= sizeof(NxVec3);<br />	terrainDesc.triangleStrideBytes			= 3 * sizeof(NxU32);<br />	terrainDesc.points						= gTerrainVerts;<br />	terrainDesc.triangles					= gTerrainFaces;							<br />	terrainDesc.flags						= 0;<br /><br />	terrainDesc.heightFieldVerticalAxis		= NX_Y;<br />	terrainDesc.heightFieldVerticalExtent	= -1000.0f;<br /><br />	NxTriangleMeshShapeDesc terrainShapeDesc;<br /><br />	NxInitCooking();<br /><br />	MemoryWriteBuffer buf;<br />	bool status = NxCookTriangleMesh(terrainDesc, buf);<br />	NxTriangleMesh* triangleMesh = gPhysicsSDK-&gt;createTriangleMesh(MemoryReadBuffer(buf.data));<br /><br />	terrainShapeDesc.meshData  = triangleMesh;<br /><br />	NxActorDesc actorDesc;<br />	actorDesc.shapes.pushBack(&terrainShapeDesc);<br />	actorDesc.body		   = NULL;<br />	NxActor* newActor = gScene-&gt;createActor(actorDesc);<br /><br />	delete []gTerrainVerts;<br />	delete []gTerrainFaces;<br /></div><br /><br />*and crashing Heightfield*<br /><div class='codetop'>CODE</div><div class='codemain' style='height:200px;white-space:pre;overflow:auto'>NxHeightFieldDesc heightFieldDesc;<br />    heightFieldDesc.nbColumns           = sizeX;<br />    heightFieldDesc.nbRows              = sizeZ;<br />    heightFieldDesc.thickness = -1000;<br />    heightFieldDesc.convexEdgeThreshold = 0;<br />    heightFieldDesc.samples             = new NxU32[sizeX * sizeZ];<br />    heightFieldDesc.sampleStride        = sizeof(NxU32);<br /><br />    NxU8* currentByte = (NxU8*) heightFieldDesc.samples;<br /><br />    for (NxU32 row = 0; row &lt; sizeX; row++)<br />    {<br />        for (NxU32 column = 0; column &lt; sizeZ; column++)<br />        {<br />            NxHeightFieldSample* currentSample = (NxHeightFieldSample*) currentByte;<br /><br />            //currentSample-&gt;height         = 0;<br />			currentSample-&gt;height         = pSrc[(row * sizeZ) + column] / 16;<br /><br />            currentSample-&gt;materialIndex0 = 1;<br />            currentSample-&gt;materialIndex1 = 1;<br />            currentSample-&gt;tessFlag = 0;<br /><br />            currentByte += heightFieldDesc.sampleStride;<br />        }<br />    }<br /><br />    NxHeightField* heightField = gPhysicsSDK()-&gt;createHeightField(heightFieldDesc);<br /><br />    delete[] heightFieldDesc.samples;<br /><br />    NxHeightFieldShapeDesc heightFieldShapeDesc;<br />    heightFieldShapeDesc.heightField     = heightField;<br />    heightFieldShapeDesc.heightScale     = 1;<br />    heightFieldShapeDesc.rowScale        = dimensions.x / sizeX; // tileSize<br />    heightFieldShapeDesc.columnScale     = dimensions.z / sizeZ; // tileSize<br />    heightFieldShapeDesc.materialIndexHighBits = 0;<br />    heightFieldShapeDesc.skinWidth = 0.001f;<br /></div><br /><br /> Rest of the code creates NxActor, based on shape i got in previous code pieces.<br /><br /> Edit: when i run RemoteDebugger in couple with second part of code (HeightField), only RemoteDebugger crashes ("Application has raised an exception and will be closed..."). My application continues to work...  <img src="http://developer.nvidia.com/forums/style_emoticons/default/blink.gif" style="vertical-align:middle" emoid=":blink:" border="0" alt="blink.gif" /> <br /><br /> What should i do to fix this error(-s)?]]></description>
		<pubDate>Wed, 11 Nov 2009 12:32:44 -0800</pubDate>
		<guid>http://developer.nvidia.com/forums/index.php?showtopic=3860</guid>
	</item>
	<item>
		<title>I have 2 actors but they dont collide with each other</title>
		<link>http://developer.nvidia.com/forums/index.php?showtopic=3859</link>
		<description><![CDATA[Hey all,<br />Currently i started with physx and made a test file.<br />i have 2 actors, 1 plane and one box...<br />when the application begins the box is suppose to fall on the plane. <br />i know that the box doesn't collide with the plain cause every-step my console outputs the box actor Y-axis and when it reaches 0 (aka it collides)it doesn't stay at 0 or -1 but it keeps going ...<br /><br />My actors creation code :<br /><!--c1--><div class='codetop'>CODE</div><div class='codemain'><!--ec1-->// Create body of the box<br />&nbsp;&nbsp;&nbsp;&nbsp;NxBodyDesc bodyDesc;<br />&nbsp;&nbsp;&nbsp;&nbsp;bodyDesc.angularDamping&nbsp;&nbsp;&nbsp;&nbsp;= 0.5f;<br /><br />&nbsp;&nbsp;&nbsp;&nbsp;//make the box shape<br />&nbsp;&nbsp;&nbsp;&nbsp;NxBoxShapeDesc boxShapeDesc;<br />&nbsp;&nbsp;&nbsp;&nbsp;boxShapeDesc.mass=3;<br /><br />&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;NxActorDesc actorDesc;<br />&nbsp;&nbsp;&nbsp;&nbsp;actorDesc.shapes.pushBack&#40;&boxShapeDesc&#41;;<br />&nbsp;&nbsp;&nbsp;&nbsp;actorDesc.body&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;= &bodyDesc;<br />&nbsp;&nbsp;&nbsp;&nbsp;actorDesc.density&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;= 10.0f;<br />&nbsp;&nbsp;&nbsp;&nbsp;NxActor* boxActor=gScene-&#62;createActor&#40;actorDesc&#41;;<br /><br />NxPlaneShapeDesc planeDesc;<br />// plane equation &#58; ax + by + cz + d = 0<br />planeDesc.normal = NxVec3 &#40; 0, 0, 1 &#41;;<br />planeDesc.d = 0.0f;<br />// actor descriptor with collection of shapes<br />NxActorDesc pactorDesc;<br />pactorDesc.shapes.pushBack&#40; &planeDesc &#41;;<br />// NxScene creates actor and returns a pointer.<br />NxActor* pActor = gScene-&#62;createActor&#40; pactorDesc &#41;;<br />pActor-&#62;userData = NULL;<!--c2--></div><!--ec2--><br /><br />and code in my while loop for the simulation to take place :<br /><!--c1--><div class='codetop'>CODE</div><div class='codemain'><!--ec1-->cubepos=boxActor-&#62;getGlobalPosition&#40;&#41;;<br />&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;cout&#60;&#60;cubepos.y &#60;&#60;endl;<br />&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;gScene-&#62;setTiming&#40;myTimestep / 4.0f, 4, NX_TIMESTEP_FIXED&#41;;<br /><br />&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;gScene-&#62;simulate&#40;myTimestep&#41;; //perform 4 substeps to advance simulation by 1/60th of a second.<br />&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;gScene-&#62;flushStream&#40;&#41;;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;//...perform useful work here using previous frame's state data&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;<br />&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;gScene-&#62;fetchResults&#40;NX_RIGID_BODY_FINISHED, true&#41;;<!--c2--></div><!--ec2--><br /><br /><br />According to my code everything is ok...what am i missing that my cube doesn't collide with the plane?<br />Thanks in advanced.<br /><br /><br /><br />-Margin]]></description>
		<pubDate>Tue, 10 Nov 2009 18:21:05 -0800</pubDate>
		<guid>http://developer.nvidia.com/forums/index.php?showtopic=3859</guid>
	</item>
	<item>
		<title><![CDATA[[Error Report] driver 9.09.0814]]></title>
		<link>http://developer.nvidia.com/forums/index.php?showtopic=3858</link>
		<description><![CDATA[I think there is a trouble on driver note  at <a href="http://www.nvidia.com/object/physx_9.09.0814.html" target="_blank">http://www.nvidia.com/object/physx_9.09.0814.html</a><br />after some test ... i m pretty sure that it dont  includes runtimes for the PhysX SDK 2.8.3, instead 9.09.030 works fine.<br />i invite nvidia to check it.<br />Carlo<br />]]></description>
		<pubDate>Tue, 10 Nov 2009 15:23:28 -0800</pubDate>
		<guid>http://developer.nvidia.com/forums/index.php?showtopic=3858</guid>
	</item>
	<item>
		<title>Surface drawing problem - GTX275 - Dx9</title>
		<link>http://developer.nvidia.com/forums/index.php?showtopic=3857</link>
		<description><![CDATA[Dear Developer nVidia community,<br /> first of all, i very apologize for my bad english, but I do not know what to do. Please give me a moment of your time and share some point of your views.<br /><br />I am a developer of freeware games and it struck me that the drivers have problems with the processing surface in 2D Direct-X. We use the program: GAME MAKER (you can find on www.yoyogames.com)<br />The only problem is, that the surface cant draw images over self. It means, that this problem freezing image in game. <br />While using a surface processing - there is also problem with image transparency (example below).<br /><br />I wrote to Gainward support and they advised me: use the old driver, but that nothing solved.<br /><br /><br />I have NOW:<br /><br />ASUS P5KPL-C/1600 motherboard<br />Intel Core 2 Duo E8400 3 Ghz<br />RAM 4 GB DDR2 - 800 MHz<br />Gainward Bliss GTX275<br />Windows 7 CZ 64bit Professional<br /><br />and display driver: 191.07 from nVidia site.<br /><br />I also tried install old drivers, but everytime there is there same problem.<br /><br /> <br />My friend have same problem. It is also developer of games.<br /><br />Asus P6T SE - Intel X58<br />Intel Core i7-920<br />Corsair DIMM 6144MB DDR III 1333MHz<br />Gainward Bliss GTX275<br />Windows 7 CZ 64bit Ultimate<br /><br />And he have 190.62 version of Gainward site driver. But he have same problem not only in my game, but in every games, which are using surface drawing processing (All game maker games).<br /><br /> <br />I'd like to recommend you, for better understanding, look at our game (also is in English) here: <a href="http://www.fiolasoft.cz/stahuj/flip-flap/" target="_blank">http://www.fiolasoft.cz/stahuj/flip-flap/</a> with Gainward GTX275, and you will see, bad transparency while using transitions effect to next level.<br /><br />Example of error here:<br /><a href='http://developer.nvidia.com/forums/index.php?act=attach&type=post&id=288'>http://developer.nvidia.com/forums/index.php?act=attach&type=post&id=288</a><br /><br />Other graphics cards, as my older nVidia GeForce 8600 GT hasnt this problem. We tested this game for many graphics cards. Mostly with the chip from nVidia and now, when i finally spent some money for the proper graphics card, ... so not working properly <img src="http://developer.nvidia.com/forums/style_emoticons/default/sad.gif" style="vertical-align:middle" emoid=":(" border="0" alt="sad.gif" />. <br /><br />Do you know what may therefore be a problem?<br /><br />I really dont know, how i can help you to imagine my problem. I tried reinstall my system. I tried use drivers on Windows Vista 32bit, also on Windows 7 32bit, everytime its same problem... And relly, this isnt only my problem - many peoples have this.<br /><br />My main questions are ...  Can something nVidia do? And if so, where and how I can report this problem? We prepared one game at this time ... and on GTX275 is unplayable. Everywhere else works, on GTX275 not. Unfortunately I dont know, if the problem concerns only the GTX275.<br /><br />Thank you for any respond! And i really apologize for my english ...<br /><br />Yours sincerely,<br /> Filip Kraucher]]></description>
		<pubDate>Tue, 10 Nov 2009 14:32:53 -0800</pubDate>
		<guid>http://developer.nvidia.com/forums/index.php?showtopic=3857</guid>
	</item>
	<item>
		<title>Texturing without API specific calls</title>
		<link>http://developer.nvidia.com/forums/index.php?showtopic=3856</link>
		<description><![CDATA[Hey all, <br /><br />Currently, I'm trying to write an engine. Things are looking pretty good so far. But there's one problem: I'd like to use CG for all my shading. My problem is that I don't really like using API specific calls. Right now I handle this problem by using a set of interfaces in a statically linked library and a separate OpenGL and DirectX implementation in a shared library (dynamically linked). My goal with this question is to move all implementation ("pure" CG if you will) to the static library. This is so I can more easily derive from the base effect class that I use.<br /><br />Cheers,<br />Emiel]]></description>
		<pubDate>Tue, 10 Nov 2009 13:56:56 -0800</pubDate>
		<guid>http://developer.nvidia.com/forums/index.php?showtopic=3856</guid>
	</item>
	<item>
		<title>Deformation of a nail possible?</title>
		<link>http://developer.nvidia.com/forums/index.php?showtopic=3855</link>
		<description><![CDATA[Hi,<br />i posted a related topic in the soft body section but i think it's more related to cloth now. So basically i want to know if it is possible to simulate the plastical deformation of a nail resulting from some force applied let's say by a hammer (the head of the nail doesn't have to be deformable, that can be done as a second step for my simulation).<br /><br />Because if it's not possible i will have to write the simulation on my own. It would be nice if someone can give me a hint for that too (i found something here <a href="http://developer.download.nvidia.com/SDK/10.5/direct3d/Source/DeformableBody/doc/DeformableBody.pdf" target="_blank">http://developer.download.nvidia.com/SDK/1...ormableBody.pdf</a> but i do not know how this is related to physx).<br /><br />Additionaly is it possible to simulate plastical deformation of a plane? Let's say i want to put the nail into a flat desk. In order to achieve this the desk has to be deformable too, right? So is this possible with Physx?<br /><br />To be clear, the deformation has not to be 100% perfect. I am happy if it works at all. <br /><br />Thanks,<br />Thomas]]></description>
		<pubDate>Tue, 10 Nov 2009 13:28:47 -0800</pubDate>
		<guid>http://developer.nvidia.com/forums/index.php?showtopic=3855</guid>
	</item>
	<item>
		<title>Perfhud and Direct3D9 Ex</title>
		<link>http://developer.nvidia.com/forums/index.php?showtopic=3854</link>
		<description><![CDATA[hello!<br />Are there any plans for Perfhud to support D3D9Ex in future?<br />I read earlier on this forum that perfhud was supposed to support this in<br />version 6.6 but it don't. So, are there any plans for this?<br />Thanks for answer<br /><br />Greetings<br />Zachar]]></description>
		<pubDate>Tue, 10 Nov 2009 00:46:53 -0800</pubDate>
		<guid>http://developer.nvidia.com/forums/index.php?showtopic=3854</guid>
	</item>
	<item>
		<title>Linux Support Discontinued?</title>
		<link>http://developer.nvidia.com/forums/index.php?showtopic=3853</link>
		<description><![CDATA[Today, the public SDK download disappeared, along with the mentioning of the Linux SDK. Is this it?<br />Missing 64bit and hardware acceleration support were indicators for cancellation?]]></description>
		<pubDate>Mon, 09 Nov 2009 22:43:32 -0800</pubDate>
		<guid>http://developer.nvidia.com/forums/index.php?showtopic=3853</guid>
	</item>
	<item>
		<title>raycastClosestShape returns NULL on Xbox</title>
		<link>http://developer.nvidia.com/forums/index.php?showtopic=3852</link>
		<description><![CDATA[We use <b>raycastClosestShape()</b> for our AI characters to find out where the ground/surface is and ensure characters feet are always snapped to the surface.  The following code (which uses the 2.8.1 SDK) works just fine under Win32:<br /><br /><!--c1--><div class='codetop'>CODE</div><div class='codemain'><!--ec1-->efd&#58;&#58;Point3 shape;<br />m_pOwningEntity-&#62;GetPropertyValue&#40;&#34;KyDimensions&#34;, shape&#41;;<br />NxRay ray&#40;NxVec3&#40;currentPosition.x, currentPosition.y, currentPosition.z+shape.z&#41;,&nbsp;&nbsp;// origin<br />&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;NxVec3&#40;0.0, 0.0, -1.0&#41;&#41;;&nbsp;&nbsp;// direction<br />NxRaycastHit hit;<br />NxShape* closestShape = pkScene-&#62;raycastClosestShape&#40;ray, NX_STATIC_SHAPES, hit&#41;;<br /><br />// If close to the surface, snap Z position to the ground<br />NxShape *closestShape = pkScene-&#62;raycastClosestShape&#40;ray, NX_STATIC_SHAPES, hit, 0xffffffff, maxDist, 0xffffffff, NULL, NULL&#41;;<br />if &#40;closestShape&#41;<br />{<br />&nbsp;&nbsp;&nbsp;&nbsp;newPosition.z = hit.worldImpact.z + 0.002f;<br />}<!--c2--></div><!--ec2--><br /><br />When compiled for Xbox 360, however, this same exact code always returns NULL and nothing for the hit parameter.  Playing around with the function parameters I found that the problem seemed to be with the <b>maxDist</b> parameter.  When we stick with the default for <b>maxDist</b>, it always works; but if I specify anything else (other than infinity), the function fails.  What's even funnier is that the ray always finds an intersection within 1 unit (1 meter) of the its starting location.  So it really should be working, but it's just not.<br /><br />Is this a bug?  A known feature?  Or what?  I just don't see what I'm doing wrong here.<br /><br />For now, we modified our code to produce the same functional result without using the <b>maxDist</b> parameter as follows:<br /><br /><!--c1--><div class='codetop'>CODE</div><div class='codemain'><!--ec1-->// Cast ray towards ground and find closest static object it intersects.<br />//<br />// NOTE&#58; On the XBox 360, passing maxDist does not work &#40;NULL result always&#41;.<br />//&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; Workaround is to raycast to infinity and check hit info after return. -- DCNefd&#58;&#58;Point3 shape;<br />efd&#58;&#58;Point3 shape;<br />m_pOwningEntity-&#62;GetPropertyValue&#40;&#34;KyDimensions&#34;, shape&#41;;<br />NxRay ray&#40;NxVec3&#40;currentPosition.x, currentPosition.y, currentPosition.z+shape.z&#41;,&nbsp;&nbsp;// origin<br />&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;NxVec3&#40;0.0, 0.0, -1.0&#41;&#41;;&nbsp;&nbsp;// direction<br />NxRaycastHit hit;<br />NxShape* closestShape = pkScene-&#62;raycastClosestShape&#40;ray, NX_STATIC_SHAPES, hit&#41;;<br /><br />// If close to the surface, snap Z position to the ground<br />NxReal hitDist = fabs&#40;ray.orig.z - hit.worldImpact.z&#41;;<br />NxReal maxDist = &#40;NxReal&#41;&#40;shape.z + 3.5f&#41;;&nbsp;&nbsp;// surface must be within 3.5 meters from character<br />if &#40;closestShape&nbsp;&nbsp;&&&nbsp;&nbsp;hitDist &#60; maxDist&#41;<br />{<br />&nbsp;&nbsp;&nbsp;&nbsp;newPosition.z = hit.worldImpact.z + 0.002f;<br />}<!--c2--></div><!--ec2--><br /><br />]]></description>
		<pubDate>Mon, 09 Nov 2009 14:46:20 -0800</pubDate>
		<guid>http://developer.nvidia.com/forums/index.php?showtopic=3852</guid>
	</item>
	<item>
		<title>How to rendea path with a Brush texture ?</title>
		<link>http://developer.nvidia.com/forums/index.php?showtopic=3851</link>
		<description><![CDATA[Hi,<br /><br />i would like  to draw a collection of segment with a texture as brush, like in Photoshop for exemple.<br /><br />The collection of segment is defined with a collection of points, stored in a texture (here we use 64 point of a 256*256 texture, passed as parameter to the shader).<br /><br />I'm currently able to draw a collection of segment with a basic cylindric brush, which is a very simplified situation.<br /><br /><br />My question is : are you aware of any paper / ressource describing a fast technic to calculate the value of a pixel between 2 control points respectively to the brush ?<br /><br />It seems to me to be an usual problematic but I was unable to find any documentation on the subject.<br /><br />Thx]]></description>
		<pubDate>Mon, 09 Nov 2009 12:45:30 -0800</pubDate>
		<guid>http://developer.nvidia.com/forums/index.php?showtopic=3851</guid>
	</item>
	<item>
		<title>nvlddmkm display  driver error</title>
		<link>http://developer.nvidia.com/forums/index.php?showtopic=3850</link>
		<description><![CDATA[please help nvidia,<br /><br />     OS: windows vista &#153; home premium, 32-bit (service pack 2)<br />     directX version: 10.0<br /><br />Graphics Card information-----------------------------------------<br />     <br />    GeForce 9200M GS<br />    Details: Driver version 186.81<br /><br />Problem:<br /><br />(same as the others)<br /><br />i recently installed driver mentioned above, and i always incountered problem in regards with the display driver name "nvlddmkm", it appears when i used to played games and making a 3d drawings in related with my jobs, i really need to fix it urgently. my computers problem and reports sent me a detailed problem solution that this NVLDDMKM display driver must be uninstalled, but i wonder where can find this driver when all that i have in my program and features is only nvidia drivers, may be there are some solutions that NVIDIA can recommend rather than uninstalling it to my computer. please help people......<br /><br />arnelrdumawal@yahoo.com.ph]]></description>
		<pubDate>Mon, 09 Nov 2009 11:01:08 -0800</pubDate>
		<guid>http://developer.nvidia.com/forums/index.php?showtopic=3850</guid>
	</item>
	<item>
		<title>Rest angle of cloth</title>
		<link>http://developer.nvidia.com/forums/index.php?showtopic=3849</link>
		<description><![CDATA[Hi,<br /><br />I'm currently loading using an imported triangle mesh for some cloth simulation. When cooking the mesh, PhysX apparently sets the rest angle for the calculation of the bending forces to the angle between the triangles in their current position. Is there anyway to change the rest angle for the simulation to something? I would like to run a simulation on an arbitrary mesh and specify a rest angle of 0 (i.e. that the rest state of all triangles is planar). As I can't really modify the original mesh to be flat (as it doesn't have a disc-like topology) I don't really see any way of faking it by modifying the initial mesh either.<br /><br />Thanks!]]></description>
		<pubDate>Mon, 09 Nov 2009 05:02:46 -0800</pubDate>
		<guid>http://developer.nvidia.com/forums/index.php?showtopic=3849</guid>
	</item>
	<item>
		<title>2.8.3 for mere Mortals</title>
		<link>http://developer.nvidia.com/forums/index.php?showtopic=3848</link>
		<description><![CDATA[Hi,<br /><br />I know that this has been asked before, but I have not seen an answer so far.<br /><br />So, when is 2.8.3 going to be released to us mere non-developers?<br /><br />Cheers,<br /><br />Ander]]></description>
		<pubDate>Sun, 08 Nov 2009 21:21:56 -0800</pubDate>
		<guid>http://developer.nvidia.com/forums/index.php?showtopic=3848</guid>
	</item>
	<item>
		<title>Windows Virtualization with Parallels</title>
		<link>http://developer.nvidia.com/forums/index.php?showtopic=3847</link>
		<description><![CDATA[Does any one know if the NVIdia SDK tools work under virtualization with Parallels 5 specifically?  I am running a MAC PRO with Snow Leopard<br />with Parallels 5 and Windows 7.  Pretty flawless overall, but Parallels has  tight control over the graphics drive and their generic one usually<br />acts like a firewall.  I am specifically interested in developing for OpenGL ES 2.0 (under emulation of course). My card is a Geforce 8800.  The one sold by Apple and pre-configured.<br /><br />Any guidance is appreciated.<br /><br />Thanks, Patrice]]></description>
		<pubDate>Sun, 08 Nov 2009 20:52:55 -0800</pubDate>
		<guid>http://developer.nvidia.com/forums/index.php?showtopic=3847</guid>
	</item>
	<item>
		<title>Does NVIdia Emulator in Parallels 5.0 on a Mac</title>
		<link>http://developer.nvidia.com/forums/index.php?showtopic=3846</link>
		<description><![CDATA[Hi - <br /><br />Does anyone know if NVEmulate works under Parallels 5 (or any version for that matter) on a MAC PRO + Geforce running<br />Parallels+Windows 7 for windows emulation?<br /><br />Thanks in advance]]></description>
		<pubDate>Sun, 08 Nov 2009 20:47:23 -0800</pubDate>
		<guid>http://developer.nvidia.com/forums/index.php?showtopic=3846</guid>
	</item>
	<item>
		<title>Whether there is an OpenCL emulation?</title>
		<link>http://developer.nvidia.com/forums/index.php?showtopic=3844</link>
		<description><![CDATA[Whether prompt, please, at use OpenCL emulation, how at Cuda is possible? And that at me a video card without support Cuda.<br />]]></description>
		<pubDate>Sun, 08 Nov 2009 09:15:56 -0800</pubDate>
		<guid>http://developer.nvidia.com/forums/index.php?showtopic=3844</guid>
	</item>
	<item>
		<title>Actors wont move</title>
		<link>http://developer.nvidia.com/forums/index.php?showtopic=3843</link>
		<description><![CDATA[I'm using PhysX 2.8.1 API under Linux. My scene initializes and I get all the objects I would expect. I have a loop that retrieves actor positions at any point, but none of my actors seem to move. With just a plane and a box floating above it, nothing seems to move. I have set  gravity in the sceneDesc and the actors starts with a velocity. Here is the code I use every frame to update and get coords:<br /><br />		int nbActors = pScene-&gt;getNbActors();<br />		NxActor** actors = pScene-&gt;getActors();<br />		while(nbActors--)<br />		{<br />			NxActor* actor = *actors++;<br />			if(!actor-&gt;userData) continue;<br /><br />			Entity *e = (Entity *)actor-&gt;userData;<br />			if(e && !e-&gt;props) e = e-&gt;parent;<br />			if(e && !e-&gt;props) e = e-&gt;parent;<br />			if(!e || !e-&gt;props) continue;<br /><br />			nProperties::OriginProperty *prop =<br />				(nProperties::OriginProperty *)e-&gt;props-&gt;GetProp(PropertyManager::ORIGIN);<br />			if(prop) {<br />				cmMatrix m;<br />				//actor-&gt;getGlobalPose().getColumnMajor44(m.mbuf);<br /><br />				NxVec3 nxo = actor-&gt;getGlobalPose().t;<br />				prop-&gt;orig = Vec3(VEC3_3f(nxo));<br />				t2printf("Orig update %s: %f %f %f&#092;n", e-&gt;id, VEC3_3f(prop-&gt;orig));<br />			}<br />		}<br /><br />		pScene-&gt;simulate(oTimer.renderTimer.totaltime / 1000.0f);<br />		pScene-&gt;flushStream();<br />		pScene-&gt;fetchResults(NX_RIGID_BODY_FINISHED, true);<br /><br />i get the position reported back properly in the actor loop, but it never changes...]]></description>
		<pubDate>Sat, 07 Nov 2009 19:49:48 -0800</pubDate>
		<guid>http://developer.nvidia.com/forums/index.php?showtopic=3843</guid>
	</item>
	<item>
		<title>how to statically link nvcompress for Linux?</title>
		<link>http://developer.nvidia.com/forums/index.php?showtopic=3841</link>
		<description><![CDATA[Hi<br /><br />I built the nvcompress command line tool for Linux with...<br /><br />$ wget <a href="http://nvidia-texture-tools.googlecode.com/files/nvidia-texture-tools-2.0.7-1.tar.gz" target="_blank">http://nvidia-texture-tools.googlecode.com...-2.0.7-1.tar.gz</a><br />$ tar zxvf nvidia-texture-tools-2.0.7-1.tar.gz<br />$ cd nvidia-texture-tools<br />$ ./configure --release<br />$ make<br /><br />It creates (among other things) the tool ./build/src/nvtt/nvcompress<br />with dynamic libraries:<br /><br />$ ldd buid/src/nvtt/nvcompres<br />        linux-vdso.so.1 =&gt;  (0x00007fff335fe000)<br />        libnvtt.so =&gt; /home/dope/sb/nvidia-texture-tools/build/src/nvtt/libnvtt.so (0x00007fe82b040000)<br />        libnvimage.so =&gt; /home/dope/sb/nvidia-texture-tools/build/src/nvimage/libnvimage.so (0x00007fe82ae16000)<br />        libnvmath.so =&gt; /home/dope/sb/nvidia-texture-tools/build/src/nvmath/libnvmath.so (0x00007fe82ac0f000)<br />        libnvcore.so =&gt; /home/dope/sb/nvidia-texture-tools/build/src/nvcore/libnvcore.so (0x00007fe82aa06000)<br />        libdl.so.2 =&gt; /lib64/libdl.so.2 (0x00007fe82a802000)<br />        libpng12.so.0 =&gt; /usr/lib64/libpng12.so.0 (0x00007fe82a5da000)<br />        libz.so.1 =&gt; /lib64/libz.so.1 (0x00007fe82a3c4000)<br />        libjpeg.so.62 =&gt; /usr/lib64/libjpeg.so.62 (0x00007fe82a1a0000)<br />        libtiff.so.3 =&gt; /usr/lib64/libtiff.so.3 (0x00007fe829f44000)<br />        libstdc++.so.6 =&gt; /usr/lib64/libstdc++.so.6 (0x00007fe829c38000)<br />        libm.so.6 =&gt; /lib64/libm.so.6 (0x00007fe8299e2000)<br />        libgcc_s.so.1 =&gt; /lib64/libgcc_s.so.1 (0x00007fe8297cb000)<br />        libc.so.6 =&gt; /lib64/libc.so.6 (0x00007fe829472000)<br />        /lib64/ld-linux-x86-64.so.2 (0x00007fe82b25c000)<br /><br />*Question*: how can I build nvcompress with static libraries rather<br />than dynamic libs?<br /><br />I would prefer to have nvcompress statically linked to make it easier<br />to copy nvcompress on several hosts without to worry about missing<br />dynamic libs for libnvimage.so, libnvcore.so, libnvmath.so, libnvtt.so.<br /><br />I was looking through the makefiles but they are not straightforward.<br /><br />Thanks<br />]]></description>
		<pubDate>Sat, 07 Nov 2009 00:31:36 -0800</pubDate>
		<guid>http://developer.nvidia.com/forums/index.php?showtopic=3841</guid>
	</item>
	<item>
		<title>strange vehicles behaviour</title>
		<link>http://developer.nvidia.com/forums/index.php?showtopic=3840</link>
		<description><![CDATA[My car flip when turning at higher speeds... my code is similiar to createCarWithDesc from Car Example (car.cpp). Do anyone know which parameter is responsible for such behaviour? I tried to change a lot of them, especially wheels friction and vehicleDesc.mass but I wasnt able to make my car feel realistic, like in for example THIS thread:<br /><a href="http://developer.nvidia.com/forums/index.php?showtopic=2774" target="_blank">http://developer.nvidia.com/forums/index.php?showtopic=2774</a><br /><br />any help is welcomed...]]></description>
		<pubDate>Sat, 07 Nov 2009 00:26:24 -0800</pubDate>
		<guid>http://developer.nvidia.com/forums/index.php?showtopic=3840</guid>
	</item>
	<item>
		<title>Vehicle Speed</title>
		<link>http://developer.nvidia.com/forums/index.php?showtopic=3839</link>
		<description><![CDATA[Hi There,<br /><br />             I am developing a car simulation, in that, i am modifies the samples in the physx, but i cant increase the car's speed, i tried to change the various settings, but still i cant increase the speed,<br /><br />Any hint please?<br /><br />Thanks,<br />Sara.]]></description>
		<pubDate>Sat, 07 Nov 2009 00:22:06 -0800</pubDate>
		<guid>http://developer.nvidia.com/forums/index.php?showtopic=3839</guid>
	</item>
	<item>
		<title>WheelShape code</title>
		<link>http://developer.nvidia.com/forums/index.php?showtopic=3838</link>
		<description><![CDATA[Taken from the Raycastcar sample, what does this line mean? I'm assuming thats setting the vehicle up to use the WheelShape?<br /><br />wheelDesc[i].wheelFlags |= NX_WF_USE_WHEELSHAPE;<br /><br />And then when the wheel is created later:<br /><br />if(wheelDesc-&gt;wheelFlags & NX_WF_USE_WHEELSHAPE)<br /><br />This is a true statement correct?<br />Sorry I am not an expert C++ programmer and don't understand the usage of the operators here.]]></description>
		<pubDate>Fri, 06 Nov 2009 18:18:53 -0800</pubDate>
		<guid>http://developer.nvidia.com/forums/index.php?showtopic=3838</guid>
	</item>
	<item>
		<title><![CDATA[Global position from actor's frame position]]></title>
		<link>http://developer.nvidia.com/forums/index.php?showtopic=3837</link>
		<description><![CDATA[Hi,<br />i have the local positions of the shapes of a NxActor. But this positions aren't in relation to the universal space. How i can to translate from actor's frame position to global position. Look at:<br /><br />	NxShape *mesh =  this-&gt;actor-&gt;getShapes()[0];<br /><br />	NxMat34 pose = mesh-&gt;getGlobalPose();<br />	void* ptr = mesh-&gt;userData;<br />	NxConvexMeshDesc meshDesc = *((NxConvexMeshDesc*)(ptr));<br /><br />	NxU32 nbVerts = meshDesc.numVertices;	<br />	NxVec3* points = (NxVec3 *)meshDesc.points;<br /><br />The *points variable is a vector of NxVec3 representing the positions of the vertices of the actor in relation to the actor's frame. How i can obtain the global position of these points?]]></description>
		<pubDate>Fri, 06 Nov 2009 12:52:50 -0800</pubDate>
		<guid>http://developer.nvidia.com/forums/index.php?showtopic=3837</guid>
	</item>
	<item>
		<title>Actor memory usage</title>
		<link>http://developer.nvidia.com/forums/index.php?showtopic=3836</link>
		<description><![CDATA[Hi,<br />Can someone help me calculate how much memory is required by the SDK for one dynamic actor with one shape, say a convex-mesh with N triangular faces? Thanks!<br />-nuun]]></description>
		<pubDate>Fri, 06 Nov 2009 12:28:08 -0800</pubDate>
		<guid>http://developer.nvidia.com/forums/index.php?showtopic=3836</guid>
	</item>
	<item>
		<title>NVidia GeForce FX 5600</title>
		<link>http://developer.nvidia.com/forums/index.php?showtopic=3835</link>
		<description><![CDATA[Windows 7 Ultimate (7600)<br /><br />nVIDIA ForceWare 96.85<br /><br /><br /><div class='codetop'>CODE</div><div class='codemain' style='height:200px;white-space:pre;overflow:auto'><br />- System <br /><br />  - Provider <br /><br />   [ Name]  Microsoft-Windows-ApplicationExperienceInfrastructure <br />   [ Guid]  {5EC13D8E-4B3F-422E-A7E7-3121A1D90C7A} <br /> <br />   EventID 1 <br /> <br />   Version 0 <br /> <br />   Level 3 <br /> <br />   Task 0 <br /> <br />   Opcode 2 <br /> <br />   Keywords 0x8000000000000000 <br /> <br />  - TimeCreated <br /><br />   [ SystemTime]  2009-11-06T18:46:49.151367100Z <br /> <br />   EventRecordID 3581 <br /> <br />   Correlation <br /> <br />  - Execution <br /><br />   [ ProcessID]  1608 <br />   [ ThreadID]  2180 <br /> <br />   Channel Application <br /> <br />   Computer  <br /> <br />  - Security <br /><br />   [ UserID]  S-1-5-21-1057919540-1362870897-3904713084-1000 <br /> <br /><br />- EventData <br /><br />  DBType 15 <br />  AppNameCount 62 <br />  AppName nVidia Display drivers/nView Desktop Manager (NVIDIA Drivers) <br />  VendorNameCount 7 <br />  VendorName NVidia <br />  SummaryCount 171 <br />  Summary &#1055;&#1088;&#1086;&#1075;&#1088;&#1072;&#1084;&#1084;&#1072; nVidia Display drivers/nView Desktop Manager (NVIDIA Drivers) &#1085;&#1077;&#1089;&#1086;&#1074;&#1084;&#1077;&#1089;&#1090;&#1080;&#1084;&#1072; &#1089; &#1076;&#1072;&#1085;&#1085;&#1086;&#1081; &#1074;&#1077;&#1088;&#1089;&#1080;&#1077;&#1081; Windows. &#1044;&#1083;&#1103; &#1087;&#1086;&#1083;&#1091;&#1095;&#1077;&#1085;&#1080;&#1103; &#1076;&#1086;&#1087;&#1086;&#1083;&#1085;&#1080;&#1090;&#1077;&#1083;&#1100;&#1085;&#1099;&#1093; &#1089;&#1074;&#1077;&#1076;&#1077;&#1085;&#1080;&#1081; &#1086;&#1073;&#1088;&#1072;&#1097;&#1072;&#1081;&#1090;&#1077;&#1089;&#1100; &#1082; NVidia. <br /> <br /><br /></div>]]></description>
		<pubDate>Fri, 06 Nov 2009 09:48:50 -0800</pubDate>
		<guid>http://developer.nvidia.com/forums/index.php?showtopic=3835</guid>
	</item>
	<item>
		<title>brightness contrast</title>
		<link>http://developer.nvidia.com/forums/index.php?showtopic=3834</link>
		<description><![CDATA[hi<br /><br />can someone give me a hint on a correct contrast function for my pixelshader?<br /><br />thx]]></description>
		<pubDate>Fri, 06 Nov 2009 03:39:54 -0800</pubDate>
		<guid>http://developer.nvidia.com/forums/index.php?showtopic=3834</guid>
	</item>
	<item>
		<title><![CDATA[[bug] - createActor returns 0]]></title>
		<link>http://developer.nvidia.com/forums/index.php?showtopic=3833</link>
		<description><![CDATA[Its something related to bodyDesc.mass = 1 line because when its commented out everything works... the problem occurs when im trying to set the mass of bodyDesc. Then "createActor" just returns 0. Anyone know what cause it?<br /><!--c1--><div class='codetop'>CODE</div><div class='codemain'><!--ec1-->&nbsp;&nbsp;&nbsp;&nbsp;// Create body<br />&nbsp;&nbsp;&nbsp;&nbsp;NxBodyDesc bodyDesc;<br />&nbsp;&nbsp;&nbsp;&nbsp;bodyDesc.mass = 1; // i have to remove it otherwise we get acces violation a-&#62;doSomething because createActor returns 0<br />&nbsp;&nbsp;&nbsp;&nbsp;bodyDesc.angularDamping&nbsp;&nbsp;&nbsp;&nbsp;= 0.5f;<br />&nbsp;&nbsp;&nbsp;&nbsp;if&#40;initialVelocity&#41;<br />&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;bodyDesc.linearVelocity = *initialVelocity;<br /><br />&nbsp;&nbsp;&nbsp;&nbsp;NxBoxShapeDesc boxDesc;<br />&nbsp;&nbsp;&nbsp;&nbsp;boxDesc.group = GROUP_COLLIDABLE_PUSHABLE;<br />&nbsp;&nbsp;&nbsp;&nbsp;boxDesc.dimensions = NxVec3&#40;&#40;float&#41;size, &#40;float&#41;size, &#40;float&#41;size&#41;;<br />&nbsp;&nbsp;&nbsp;&nbsp;NxActorDesc actorDesc;<br />&nbsp;&nbsp;&nbsp;&nbsp;actorDesc.shapes.pushBack&#40;&boxDesc&#41;;<br />&nbsp;&nbsp;&nbsp;&nbsp;actorDesc.body&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;= &bodyDesc;<br />&nbsp;&nbsp;&nbsp;&nbsp;actorDesc.density&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;= 10.0f;<br />&nbsp;&nbsp;&nbsp;&nbsp;actorDesc.globalPose.t&nbsp;&nbsp;= pos;<br />&nbsp;&nbsp;&nbsp;&nbsp;NxActor *a = gScene-&#62;createActor&#40;actorDesc&#41;;<br />&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;a-&#62;doSomething /// acces violation, a == 0 !<!--c2--></div><!--ec2-->]]></description>
		<pubDate>Fri, 06 Nov 2009 03:27:14 -0800</pubDate>
		<guid>http://developer.nvidia.com/forums/index.php?showtopic=3833</guid>
	</item>
	<item>
		<title>Sampling a texture at a given point</title>
		<link>http://developer.nvidia.com/forums/index.php?showtopic=3832</link>
		<description><![CDATA[Hello everyone,<br /><br />I have a simple particle system implemented on the GPU (one that I borrowed from the CG book) and I am trying to make some simple changes to it.<br /><br />Now, my pixel shader takes a texture as an input and what I am trying to do is figure out the colour of the texture at the current location. My plan was to change the brightness of the particles based on the<br />underlying colour.<br /><br />So, my vertex shader looks as follows:<br /><br /><div class='codetop'>CODE</div><div class='codemain' style='height:200px;white-space:pre;overflow:auto'><br />struct fragment<br />{<br />       float4  position    : POSITION;<br />       float4  colour      : COLOR0;<br />       float2  texcoord0   : TEXCOORD0;<br />       float   pointSize   : PSIZE;<br />};<br /><br /><br />fragment main(float3 pInitial   : POSITION,<br />             float3 vInitial   : TEXCOORD0,<br />             float tInitial    : TEXCOORD1,<br />             uniform float     globalTime,<br />             uniform float     particleSize,<br />             uniform float4x4  modelViewProj)<br />{<br />   fragment OUT;<br />   float t = globalTime - tInitial;<br />   float4 final = float4(pInitial, 1) + (float4(vInitial, 0) + float4(0, (-4.9f), 0, 0) * t * t;<br />   OUT.position = mul(modelViewProj, final);<br />   OUT.texcoord0 = OUT.position; // store the position for the pixel shader<br />   OUT.pointSize = particleSize;<br />   return OUT;<br />}<br /><br /></div><br /><br />So, I store the position (x, y) value in the texcoord0 parameter and my plan is to sample the colour at this position and use it as the particle colour.<br /><br />My pixel shader looks as follows:<br /><br /><div class='codetop'>CODE</div><div class='codemain' style='height:200px;white-space:pre;overflow:auto'><br />struct fragment<br />{<br />       float4 position  : POSITION;<br />       float4 colour0    : COLOR0;<br />       float2 texcoord0 : TEXCOORD0;<br />};<br /><br />struct pixel<br />{<br />       float4 colour : COLOR;<br />};<br /><br /><br /><br />pixel main( fragment IN,<br />                       uniform sampler2D inputTexture,<br />                       uniform float brightness,<br />                       uniform float contrast,<br />                       uniform float gamma)<br />{<br />   pixel OUT;<br />   OUT.colour = tex2D(inputTexture, IN.texcoord0);<br />   return OUT;<br />}<br /></div><br /><br />What I was expecting was the particle to have the sampled colour but what I actually see is that the texture gets mapped to the particle. I would actually like it to take a simple colour value rather than doing texture mapping. I have played around with the code in many ways but am unable to get the desired functionality.<br /><br />I would really appreciate any help as this is driving me up the wall!<br /><br />Thanks,<br /><br />xarg]]></description>
		<pubDate>Thu, 05 Nov 2009 23:55:30 -0800</pubDate>
		<guid>http://developer.nvidia.com/forums/index.php?showtopic=3832</guid>
	</item>
	<item>
		<title><![CDATA[&#26377;&#27809;&#26377;&#20013;&#22269;&#30340;&#26379;&#21451;&#65311;]]></title>
		<link>http://developer.nvidia.com/forums/index.php?showtopic=3831</link>
		<description><![CDATA[&#26412;&#20154;&#19968;&#30452;&#20174;&#20107;notebook&#31508;&#35760;&#26412;&#30828;&#20214;&#26041;&#38754;&#30340;&#24037;&#20316;&#65292;&#20174;06&#24180;NVIDIA&#36136;&#37327;&#38376;&#23601;&#19968;&#30452;&#20851;&#27880;&#26174;&#21345;&#30828;&#20214;&#20102;&#65292;&#20294;&#36825;&#37324;&#20132;&#27969;&#30828;&#20214;&#30340;&#26379;&#21451;&#30495;&#23569;&#65292;&#24212;&#29992;&#26041;&#38754;&#30340;&#22810;&#65281;&#36825;&#37324;&#26377;&#20013;&#22269;&#30340;&#26379;&#21451;&#21527;?&#25105;&#22312;&#19978;&#28023;]]></description>
		<pubDate>Thu, 05 Nov 2009 22:09:40 -0800</pubDate>
		<guid>http://developer.nvidia.com/forums/index.php?showtopic=3831</guid>
	</item>
	<item>
		<title>Need help in revolving actors about a specified anchor point</title>
		<link>http://developer.nvidia.com/forums/index.php?showtopic=3830</link>
		<description><![CDATA[Hello All,<br /><br />In the PhysX scene that Iam working on, I have two chain conveyors that transport blocks. These conveyors are made of various NxActors such as wheels, frame, chain elements and revolute joints. This works fine.<br /><br />I now want another conveyor at right angles to the first one. Since I donot want to do the math again for positioning all the above NxActors again, is there a way to revolve the conveyor about a specified anchor point ? This way, I will not have to do the math again for getting another identical conveyor at right angles to the first one.<br /><br />This is what I have tried so far. <br /><br />--------------<br />frame-&gt;setGlobalOrientationQuat(AnglesToQuat(NxVec3(0,90,0)));<br />wheel0-&gt;setGlobalOrientationQuat(AnglesToQuat(NxVec3(0,90,0)));<br />wheel1-&gt;setGlobalOrientationQuat(AnglesToQuat(NxVec3(0,90,0)));<br />for(int cnt = 0; cnt &lt; numOfChainsLinks; ++cnt)<br />     chainLink[cnt]-&gt;setGlobalOrientationQuat(AnglesToQuat(NxVec3(0,90,0)));<br />--------------<br /><br />But this does not work since each actor get rotated about their own center and not about a single global anchor point.<br />Please let me know if there is lack of clarity in my question and I will try to elaborate it with some more code.<br /><br />Thanks for your time.   <img src="http://developer.nvidia.com/forums/style_emoticons/default/smile.gif" style="vertical-align:middle" emoid=":)" border="0" alt="smile.gif" /> <br /><br />Cheers,<br />Balaji]]></description>
		<pubDate>Thu, 05 Nov 2009 21:55:19 -0800</pubDate>
		<guid>http://developer.nvidia.com/forums/index.php?showtopic=3830</guid>
	</item>
	<item>
		<title>Actor Vertices</title>
		<link>http://developer.nvidia.com/forums/index.php?showtopic=3829</link>
		<description>There is some way to determine the vertices of the meshs of a actor?</description>
		<pubDate>Thu, 05 Nov 2009 14:48:20 -0800</pubDate>
		<guid>http://developer.nvidia.com/forums/index.php?showtopic=3829</guid>
	</item>
	<item>
		<title>Friction in the wrong direction?</title>
		<link>http://developer.nvidia.com/forums/index.php?showtopic=3828</link>
		<description><![CDATA[I’m looking at a simple model of blocks sliding down a ramp. This is the frictionramp.max example from the samples directory of the 3ds Max plugin.<br /><br />I would expect the blocks to slide straight down. Instead they curve to the side. If I set the friction (static and dynamic) to 0 they do slide straight down.<br /><br />If I delete all but 1 block, it still curves, so it’s not the blocks influencing eachother.<br /><br />If I export the model to .xml it has the same problem, so it doesn’t seem to have anything to do with the plugin.<br /><br />This forum message <a href="http://developer.nvidia.com/forums/index.php?showtopic=3379" target="_blank">http://developer.nvidia.com/forums/index.php?showtopic=3379</a> also seems to report a friction in the wrong direction, but doesn’t have any replies yet. <br /><br />Can somebody please explain what’s going on here?<br /><br />Martijn<br />]]></description>
		<pubDate>Thu, 05 Nov 2009 11:46:15 -0800</pubDate>
		<guid>http://developer.nvidia.com/forums/index.php?showtopic=3828</guid>
	</item>
	<item>
		<title>Enabling Quadro professional stereo?</title>
		<link>http://developer.nvidia.com/forums/index.php?showtopic=3827</link>
		<description><![CDATA[Hi<br /><br />I'm developing an application using the Quadro professional stereoscopic features via SceniX (former NVSG SDK). OS is Windows Vista 64 bit.<br />As far as I know, SceniX requires the Quadro stereo option to be activated in the NVidia control panel. This option is located at the 3d settings, it's a simple drop-down list (on/off) plus a control for switching left/right eye. But I didn't find a way to control this option by my application.<br />My customer requires a simple setup (1-click, if possible), so at least I have to assure this option is enabled after setup and stays enabled after any update of Windows and Quadro drivers and so on.<br /><br />I already took a look at (official) NVAPI, but it seems to me to provide only methods to control 3d vision, nothing for the Quadro professional stereo.<br />The NVControlPanel API seems to provide no stereo features at all.<br /><br /><br />Is there a way to enable the Quadro stereo feature by my application?<br />(an API or the like)<br /><br /><br />Thank you for your answers]]></description>
		<pubDate>Thu, 05 Nov 2009 11:36:09 -0800</pubDate>
		<guid>http://developer.nvidia.com/forums/index.php?showtopic=3827</guid>
	</item>
	<item>
		<title>NO AGEIA PhysX processor installd :s plzzzzzzzz help</title>
		<link>http://developer.nvidia.com/forums/index.php?showtopic=3826</link>
		<description><![CDATA[i hav opend Physx properties n over there it says tht no AGEIA PhysX processor installed.....there is nthng writtn on physX proccesr n i dnt c any demos n last thin when i go 2 settings there is GeForce PhysX n no acceleration when i click on GeForce nthn happens....plzzzzz help me..... <img src="http://developer.nvidia.com/forums/style_emoticons/default/unsure.gif" style="vertical-align:middle" emoid=":unsure:" border="0" alt="unsure.gif" />]]></description>
		<pubDate>Thu, 05 Nov 2009 10:43:46 -0800</pubDate>
		<guid>http://developer.nvidia.com/forums/index.php?showtopic=3826</guid>
	</item>
	<item>
		<title>SampleCharacterController - PhysX SDK</title>
		<link>http://developer.nvidia.com/forums/index.php?showtopic=3825</link>
		<description><![CDATA[I started implementin character controllers to my app, whole code is based on SampleCharacterController example... everything should work correct, BUT there's one problem with collision character-plane<br /><br />heres my spawn plane code:<br /><br /><br /><!--c1--><div class='codetop'>CODE</div><div class='codemain'><!--ec1-->&nbsp;&nbsp;&nbsp;&nbsp;NxPlaneShapeDesc planeDesc;&nbsp;&nbsp;&nbsp;&nbsp;<br />&nbsp;&nbsp;&nbsp;&nbsp;planeDesc.group = GROUP_COLLIDABLE_NON_PUSHABLE;<br />&nbsp;&nbsp;&nbsp;&nbsp;NxActorDesc actorDesc;<br />&nbsp;&nbsp;&nbsp;&nbsp;actorDesc.shapes.pushBack&#40;&planeDesc&#41;;<br />&nbsp;&nbsp;&nbsp;&nbsp;ac = gScene-&#62;createActor&#40;actorDesc&#41;;<!--c2--></div><!--ec2--><br />as you see, im settin the proper GROUP_COLLIDABLE_NON_PUSHABLE on my plane BUT for some reasons my character doesnt even collide with it... my character just fall down, down.... I dont know whats wrong. Everything should be correct because my character collide with boxes properly, and boxes collide with plane...<br /><br />anyone know what can be wrong? Thanks for any help guys.....]]></description>
		<pubDate>Thu, 05 Nov 2009 03:02:05 -0800</pubDate>
		<guid>http://developer.nvidia.com/forums/index.php?showtopic=3825</guid>
	</item>
	<item>
		<title><![CDATA[glDrawRangeElements & IndexBuffer]]></title>
		<link>http://developer.nvidia.com/forums/index.php?showtopic=3824</link>
		<description><![CDATA[hey folks,<br />I've got a little question. I'm using VBO and and IBO (GL_ELEMENT_ARRAY_BUFFER) and bind them.<br />I found the glDrawRangeElements call quite nice since it allows me to store a single large index buffer per mesh und use differnet "views" to set up the material.<br />Everything in my code is running fine, my question is performancewise.<br />my call looks more or less like this<br />glDrawRangeElements( GL_TRIANGLES,x1,x2,(GLsizei)pIdxBufferView-&gt;GetCount(),GL_UNSIGNED_INT,(GLvoid*)pIdxBufferView-&gt;GetStartIndex()*sizeof(uint32_t) );<br /><br />I can see that x1 and x2 are valuable when not using VBO & IBO.<br />Is there a performance hit when using x1=0 and x2 = GL_MAX_ELEMENTS_VERTICES (when using VBO&IBO)???<br /><br />thanks a lot in advance<br />zqueezy]]></description>
		<pubDate>Thu, 05 Nov 2009 00:42:53 -0800</pubDate>
		<guid>http://developer.nvidia.com/forums/index.php?showtopic=3824</guid>
	</item>
	<item>
		<title>Determining Vertices</title>
		<link>http://developer.nvidia.com/forums/index.php?showtopic=3823</link>
		<description>How i can to determine the vertices of an NxConvexMeshDesc based actor (such as a pyramid, or a prism) in physX?</description>
		<pubDate>Wed, 04 Nov 2009 20:32:05 -0800</pubDate>
		<guid>http://developer.nvidia.com/forums/index.php?showtopic=3823</guid>
	</item>
	<item>
		<title>dynamic simulation</title>
		<link>http://developer.nvidia.com/forums/index.php?showtopic=3821</link>
		<description><![CDATA[Hi everyone, I just registered today in hopes of having some questions answered.<br /><br />I'll be popping in over the next few weeks but just want to get a quick heads-up on the PhysX plugin. I am looking for a piece of software that can simulate a pretty specific situation but am not sure where to start!<br />I'll briefly describe the situation here and once I get a chance to organize some details I will post more information.<br /><br />Picture a large rectangular tank to be filled with some material. The floor of the tank is essentially a feed mechanism to move the material toward the rear of the tank. At the rear of the tank there is another mechanism which projects the material outward.<br />My question is this: can the plugin roughly simulate the area covered by the material as it is projected from the machine? It doesn't need to be perfectly accurate, but I'm looking for a rough idea of where the material could end up given different conditions.<br />Another bonus would be to simulate how the material acts while it is in the tank, and how a piece of debris may affect the mechanisms.<br /><br />Sorry this is so vague but I don't have much time right now...I'll be in tomorrow to provide specifics, links, and hopefully models if anyone wants to tackle this. But for the moment I'd be happy if someone told me I was looking in the right place even!<br /><br />Thanks!]]></description>
		<pubDate>Wed, 04 Nov 2009 14:00:09 -0800</pubDate>
		<guid>http://developer.nvidia.com/forums/index.php?showtopic=3821</guid>
	</item>
	<item>
		<title>PhysX Visual Debugger - formerly VRD</title>
		<link>http://developer.nvidia.com/forums/index.php?showtopic=3819</link>
		<description>Today we released PVD version 1.1.8 which supports PhysX SDK Version 2.8.3 and all previous SDK versions.  The PhysX PS3 SDK does not support this version of PVD but we are working on a fix which should be released shortly.</description>
		<pubDate>Wed, 04 Nov 2009 13:26:47 -0800</pubDate>
		<guid>http://developer.nvidia.com/forums/index.php?showtopic=3819</guid>
	</item>
	<item>
		<title>attach camera</title>
		<link>http://developer.nvidia.com/forums/index.php?showtopic=3817</link>
		<description><![CDATA[How could i attach the camera by a created object ( a box created over a ground plane)?<br />I want to move the camera only rotating around his z axis, and the rest of orientation being only maked by the movement of the object on which is attached...<br />I am a beginner in physX , and this question i consider a verry  important one ( the parent object ) ...<br />But nowere to find an answer.]]></description>
		<pubDate>Wed, 04 Nov 2009 06:26:54 -0800</pubDate>
		<guid>http://developer.nvidia.com/forums/index.php?showtopic=3817</guid>
	</item>
	<item>
		<title>Racing game Tutor</title>
		<link>http://developer.nvidia.com/forums/index.php?showtopic=3816</link>
		<description><![CDATA[Is there a concise tutorial on making a racing game using physx. I searched on the internet but only found bits and bytes of information.<br />The physx documentation is pretty good but are there resources other that it.]]></description>
		<pubDate>Wed, 04 Nov 2009 02:28:03 -0800</pubDate>
		<guid>http://developer.nvidia.com/forums/index.php?showtopic=3816</guid>
	</item>
	<item>
		<title><![CDATA[loading static world mesh into physx, [problem]]]></title>
		<link>http://developer.nvidia.com/forums/index.php?showtopic=3812</link>
		<description>solved.</description>
		<pubDate>Tue, 03 Nov 2009 22:00:12 -0800</pubDate>
		<guid>http://developer.nvidia.com/forums/index.php?showtopic=3812</guid>
	</item>
	<item>
		<title>November and still no new plugins for max or maya?</title>
		<link>http://developer.nvidia.com/forums/index.php?showtopic=3811</link>
		<description><![CDATA[Hey Nvidia devs did you forget something?<br /> I'm seeing all of these post that said that it was gonna be going beta in a few weeks. That was back in June. Then I saw one that you guys were looking at mid Oct to release. now it is Nov. and still not a word. What happend? Have there been cutbacks there to? Did all the devs get sacked?<br /><br />Come on guys, You guys are really throwing a wrench in some our plans here we ahve no idea what is going on and when you are making all of our lives very difficult!.]]></description>
		<pubDate>Tue, 03 Nov 2009 14:19:48 -0800</pubDate>
		<guid>http://developer.nvidia.com/forums/index.php?showtopic=3811</guid>
	</item>
	<item>
		<title>Determining Vertices</title>
		<link>http://developer.nvidia.com/forums/index.php?showtopic=3810</link>
		<description>How i can to determine the vertices of an NxConvexMeshDesc based actor (such as a pyramid, or a prism) in physX?</description>
		<pubDate>Tue, 03 Nov 2009 13:45:44 -0800</pubDate>
		<guid>http://developer.nvidia.com/forums/index.php?showtopic=3810</guid>
	</item>
	<item>
		<title>Switching Textures, Parameters, etc</title>
		<link>http://developer.nvidia.com/forums/index.php?showtopic=3809</link>
		<description>Is Cg smart enough to ignore repeated cgD3D9SetTexture() which pass the same pointer to a texture?</description>
		<pubDate>Tue, 03 Nov 2009 13:11:30 -0800</pubDate>
		<guid>http://developer.nvidia.com/forums/index.php?showtopic=3809</guid>
	</item>
	<item>
		<title>SolidWireFrame very small bug</title>
		<link>http://developer.nvidia.com/forums/index.php?showtopic=3807</link>
		<description><![CDATA[the SolidWireFrame sample has a small bug in SolidWireFrame.cpp:<br /><br />void destroyGeometries()<br />{<br />    for (unsigned int meshNb = 0; meshNb &lt; g_Geometries.size(); ++meshNb)<br />    {<br />       g_Geometries[meshNb]-&gt;Destroy();<br /><br />        CDXUTMesh10* pmesh = g_Geometries[meshNb]; // added to fix bug<br />        delete pmesh; // added to fix bug<br /><br />        g_Geometries[meshNb] = NULL;<br /><br />    } <br />    g_Geometries.clear(); <br />}]]></description>
		<pubDate>Tue, 03 Nov 2009 08:59:27 -0800</pubDate>
		<guid>http://developer.nvidia.com/forums/index.php?showtopic=3807</guid>
	</item>
	<item>
		<title>DXt1 Cuda compression error</title>
		<link>http://developer.nvidia.com/forums/index.php?showtopic=3806</link>
		<description><![CDATA[Hi<br /><br />I have checked out nvtt from the trunk 5 days ago (I wanted the latest float image functionalities) and it seems the hardware cuda dxt1 compression error is not working properly. It works fine when using software compression though. To make sure it was not my cuda driver that was the cause of this corruption, I tested against the branch version 2.0.6 (by recompiling nvtt and nvcompress) and the cuda hardware compression went fine.<br /><br />The bug can be reproduced any time. Any idea why?<br /><br />Thanks<br />]]></description>
		<pubDate>Tue, 03 Nov 2009 08:59:02 -0800</pubDate>
		<guid>http://developer.nvidia.com/forums/index.php?showtopic=3806</guid>
	</item>
	<item>
		<title>Plane cooking (Wrong result)</title>
		<link>http://developer.nvidia.com/forums/index.php?showtopic=3805</link>
		<description><![CDATA[Hello,<br /><br />I'm new to PhysX since 3 days.<br />I managed to create a scene with 2 Actors (1 dynamic, 1 static)<br />The dynamic object is a simple box cooked from a mesh wich has 8 vertices and dimensions of -0.5 - 0.5.<br />the static Actor is a Plane, cooked from a mesh with 256 vertices also dimensions of -0.5 - 0.5.<br /><br />The problem is, when the box falls on the Plane Edge, it won't fall.<br />The visual debugger shows that the plane is cooked to a box with wrong dimensions, it has a width and depth of -1.0 - 1.0 and height of -0.5 - 0.5<br /><br />I searched on google and in this forum, but i can't find such a problem nor a solution.<br />I think the problem is that there are so much vertices, but i'm sure it is possible somehow, elsewise it is impossible to add a high poly character or other high poly dynamic objects.<br /><br />PS: i cooked with the flag NX_CF_COMPUTE_CONVEX]]></description>
		<pubDate>Tue, 03 Nov 2009 03:47:23 -0800</pubDate>
		<guid>http://developer.nvidia.com/forums/index.php?showtopic=3805</guid>
	</item>
	<item>
		<title>Where do I have to modify joints?</title>
		<link>http://developer.nvidia.com/forums/index.php?showtopic=3804</link>
		<description><![CDATA[Hello<br /><br />I have got a problem with setting d6joint velocities when the physic is slow (too many actors and collissions) in the scene. It looks like nothing is done, but the function to set the velocities is called.<br />Everythig works fine, if the physic is fast enough.<br /><br />My RenderCallback looks like this:<br /><div class='codetop'>CODE</div><div class='codemain' style='height:200px;white-space:pre;overflow:auto'>void RenderCallback()<br />{<br />	if(MonitorSetRequest3DUpdate)<br />	{<br />		MonitorSetRequest3DUpdate=false;<br />		glClear(GL_COLOR_BUFFER_BIT | GL_DEPTH_BUFFER_BIT);<br />		ProcessCameraKeys();<br />		SetupCamera();<br />		if (gScene && !bPause)<br />		{<br />			GetPhysicsResults();<br />			ProcessInputs();<br />			StartPhysics();<br />		}<br />		// Display scene<br />		RenderActors(bShadows);<br />		RenderFluid();<br />		//RenderFluid();<br />		if (bForceMode)<br />			DrawForce(gSelectedActor, gForceVec, NxVec3(1,1,0));<br />		else<br />			DrawForce(gSelectedActor, gForceVec, NxVec3(0,1,1));<br />		gForceVec = NxVec3(0,0,0);<br />		glFlush();<br />		glutSwapBuffers();<br />		updateJoints();<br />	}<br />}</div><br /><br />updateJoints is a c# function which sets all joint velocities with ((NxD6Joint*)d6jointid)-&gt;setDriveLinearVelocity(NxVec3(velX*LINEAR_VELOCITY_SCALE, velY*LINEAR_VELOCITY_SCALE, velZ*LINEAR_VELOCITY_SCALE)); when all joints are updated, MonitorSetRequest3DUpdate is set to true again.<br />If I wait a few seconds to set te MonitorSetRequest3DUpdate it works again.<br /><br />Do I call updateJoints() at the wrong place or is there another error in my code?<br /><br />Thank you for your help<br /><br />Robert]]></description>
		<pubDate>Mon, 02 Nov 2009 23:58:05 -0800</pubDate>
		<guid>http://developer.nvidia.com/forums/index.php?showtopic=3804</guid>
	</item>
	<item>
		<title>Maya plugin not working</title>
		<link>http://developer.nvidia.com/forums/index.php?showtopic=3803</link>
		<description><![CDATA[Hi,<br /><br />The SDK and plugin install without complaint, but<br />I'm unable to get the Maya plugin to load<br /><br />*Maya 8.5<br />*I've tried on 2 different machines, both are running XP sp2<br />*one has a 8600 GTS the other a 7900 GS <br />(Softimage's Physx implementation runs fine on the7900 <br />but I realize the Maya plugin might not)<br />*PhysX_SDK_2.8.1_build_13_for_PC<br />*NVIDIA_PhysX_Plugin_for_Maya_1.0.4<br /><br />The driver for the 8600 is a bit old (180.6) <br />but I have the most recent XP driver (191.07) on the 7900.<br /><br />Is there a specific order for the installation of the main elements?<br />(i.e. driver, SDK, plugin)<br /><br />Are there any files I may need to copy to other locations?<br /><br />Thanks,<br /><br />Jeff<br /><br /><br />]]></description>
		<pubDate>Mon, 02 Nov 2009 19:52:00 -0800</pubDate>
		<guid>http://developer.nvidia.com/forums/index.php?showtopic=3803</guid>
	</item>
	<item>
		<title>how to free an actor free from the effect of force</title>
		<link>http://developer.nvidia.com/forums/index.php?showtopic=3802</link>
		<description><![CDATA[I have created two rigid actors A and B .they are connected with a kind of joint . the joint has spring.<br />so if I make B KINEMATIC first ,and when I swing A , A will go back .<br />but ,if B is not KINEMATIC, when I swing A, B will follow A moving. <br />it is certain that the spring force between A and B is interdepend . <br /><br />but now I want to free B from the effect of the spring force ,and B is not KINEMSTIC, and I want to set its velocity later.<br /> <br />how to do that??<br />need some help!<br /><br />thank you!<br />]]></description>
		<pubDate>Mon, 02 Nov 2009 18:01:46 -0800</pubDate>
		<guid>http://developer.nvidia.com/forums/index.php?showtopic=3802</guid>
	</item>
	<item>
		<title>CkTools.h needs CkProperty virtual destructor</title>
		<link>http://developer.nvidia.com/forums/index.php?showtopic=3801</link>
		<description><![CDATA[very small bug in  CkTools.h<br /><br />add virtual destructor to CkProperty:<br /><br />   virtual ~CkProperty()<br />  {<br />  }<br />]]></description>
		<pubDate>Mon, 02 Nov 2009 13:02:25 -0800</pubDate>
		<guid>http://developer.nvidia.com/forums/index.php?showtopic=3801</guid>
	</item>
	<item>
		<title>collision detection with cooked mesh and capsule</title>
		<link>http://developer.nvidia.com/forums/index.php?showtopic=3800</link>
		<description><![CDATA[hello physx people <img src="http://developer.nvidia.com/forums/style_emoticons/default/smile.gif" style="vertical-align:middle" emoid=":)" border="0" alt="smile.gif" />, <br /><br />1) i've cooked a mesh loaded from a ASE file (tested and working ok ) <br />2) then ... i've tested cooked mesh against simple boxes actors (tested and working ok again)<br />3) i've createa a character controller actor (capsule) and the problem is that it is NOT colliding with the previous cooked mesh .<br /><br />here is my code: <br /><br />1) cooking ase (static mesh): <br /><br />                                ....<br />                                level.numVertices		= ASE.Scene.pMesh[0].VertexCount;<br />		level.numTriangles		= ASE.Scene.pMesh[0].FacesCount;<br />		level.pointStrideBytes		= sizeof(CVector3);<br />		level.triangleStrideBytes	= 3*sizeof(int);<br />		level.points		= ASE.Scene.pMesh[0].pVertex;<br />		level.triangles	                = ASE.Scene.pMesh[0].pTriangleIndex;<br />		level.flags			= NX_SIMULATION_HW;	<br /><br />		NxTriangleMeshShapeDesc	shape;<br /><br />		NxInitCooking();<br />		MemoryWriteBuffer writeBuffer;<br />		bool status = NxCookTriangleMesh(level, writeBuffer);<br />		shape.meshData = gPhysicsSDK-&gt;createTriangleMesh(MemoryReadBuffer(writeBuffer.data));<br />		NxCloseCooking();<br /><br />		shape.group		= GROUP_COLLIDABLE_NON_PUSHABLE;<br />		if (gScene-&gt;getSimType() == NX_SIMULATION_HW) shape.meshPagingMode = NX_MESH_PAGING_AUTO;<br />		<br />		actorDesc.shapes.pushBack(&shape);<br />		gScene-&gt;createActor(actorDesc);<br />                                // all ok ... cooked <img src="http://developer.nvidia.com/forums/style_emoticons/default/smile.gif" style="vertical-align:middle" emoid=":)" border="0" alt="smile.gif" /><br /><br /><br />2) creating character controller (capsule) :<br /><br />                gAllocator = new UserAllocator;<br />	gManager = NxCreateControllerManager(gAllocator);<br /><br />	NxVec3 pos;<br />	pos.x = Player.m_Camera.m_Position.x;<br />	pos.y = Player.m_Camera.m_Position.y;<br />	pos.z = Player.m_Camera.m_Position.z;<br /><br />	gCapsuleController = new MyCharacterController(gManager, gScene, pos , 5, 35);<br />               ....<br /><br />3) collision detection between capsule and simple boxes WORKS just fine so far' ! <br /><br />4) collision detection <b>between capsule and cooked mesh</b> DO NOT work , here's the code: <br /><br />void CPhysX::UpdateCharacter(NxReal deltaTime)<br />{<br />	NxU32 collisionFlags;<br /><br />	NxVec3 vel;<br />	vel.x = Player.m_Camera.m_Velocity.x;<br />	vel.y = Player.m_Camera.m_Velocity.y;<br />	vel.z = Player.m_Camera.m_Velocity.z;<br />	<br />	gCapsuleController-&gt;Move( vel * deltaTime , collisionFlags);<br /><br />	gManager-&gt;updateControllers();<br />}<br /><br />... and here's the call of the above function (main physx processing function) :<br /><br />void CPhysX::Update(float dt)<br />{<br />	nbActors = gScene-&gt;getNbActors();<br />	NxActor** actors = gScene-&gt;getActors();<br />	while(nbActors--)<br />	{<br />		NxActor* actor = *actors++;<br />		if(!actor-&gt;userData) continue;<br /><br />		// Render actor<br />		// .... draw box <br />	}<br /><br />	<b>UpdateCharacter(dt);</b><br />	<br />	// Fetch simulation results<br />	gScene-&gt;simulate(dt);<br />	gScene-&gt;flushStream();<br />	gScene-&gt;fetchResults(NX_RIGID_BODY_FINISHED, true);<br />}<br /><br /><br />The collision between cooked ASE mesh and capsule does not work . <br />Do i miss someting ? <img src="http://developer.nvidia.com/forums/style_emoticons/default/smile.gif" style="vertical-align:middle" emoid=":)" border="0" alt="smile.gif" /> Please some indications... <br />]]></description>
		<pubDate>Mon, 02 Nov 2009 12:43:51 -0800</pubDate>
		<guid>http://developer.nvidia.com/forums/index.php?showtopic=3800</guid>
	</item>
	<item>
		<title>Kinematic - Dynamic Collision Force</title>
		<link>http://developer.nvidia.com/forums/index.php?showtopic=3798</link>
		<description><![CDATA[Hi,<br /><br />This is a question about the force generated on the dynamic actor when a kinematic actor collides with a dynamic actor.<br /><br />The PhysX documentation states the kinematic actor has the "appearance of infinite mass". I'm therefore guessing that the (infinite)momentum of the kinematic actor is not used in computing the force generated on the dynamic actor.<br /><br />However, the speed(meaning the difference between two subsequent moveGlobalPosition's) of the kinematic actor when colliding with the dynamic actor does result in a greater force on the dynamic actor.<br /><br />1. Is the force generated on the dynamic actor simply the penetration of the kinematic actor into the material of the dynamic actor? A greater speed of the kinematic actor would increase the probability of penetration into the dynamic actor material, and hence greater force on the dynamic actor. However, this could potentially be unstable?<br /><br />2. Is there any documentation available on the computed force when a kinematic actor collides with a dynamic actor?]]></description>
		<pubDate>Mon, 02 Nov 2009 06:31:24 -0800</pubDate>
		<guid>http://developer.nvidia.com/forums/index.php?showtopic=3798</guid>
	</item>
	<item>
		<title>PVD crashes with heightfields</title>
		<link>http://developer.nvidia.com/forums/index.php?showtopic=3797</link>
		<description><![CDATA[Hello,<br />Apparently I'm having issues with the Visual Debugger, it crashes each time it attempts to visualize a heightfield.<br />At first I figured it might just be my issue, so I also tried using the Visual Debugger with the SampleHeightField and I was met with the same result.<br />When debugging the Visual Debugger (ironic I know) it moans about an Access Violation.<br />I'm at quite a loss as with what to do since disabling my heightfield each time I want to debug something isn't exactly a realistic possibility.<br /><br />Any help/advice is appreciated.]]></description>
		<pubDate>Mon, 02 Nov 2009 05:51:40 -0800</pubDate>
		<guid>http://developer.nvidia.com/forums/index.php?showtopic=3797</guid>
	</item>
	<item>
		<title>total begginer q: initial camera setting</title>
		<link>http://developer.nvidia.com/forums/index.php?showtopic=3796</link>
		<description><![CDATA[i am totally new in physx ..so  i am dealing now with settings or experiment variables on the code.<br />shortly , by far , my first broblem:  <br />when i am trying to change the initial camera settings from... <br />// Camera globals<br />float	gCameraAspectRatio = 1.0f;<br />NxVec3	gCameraPos(10,50,-15  --&gt; changed to several values..);<br />NxVec3	gCameraForward(1,1,11 --&gt; changed to several values..);<br />NxVec3	gCameraRight(-11,0,0 --&gt; changed to several values..);<br />nothing happend <br />everything remain unmoved and cannot change nothing in initial positition of it.....]]></description>
		<pubDate>Mon, 02 Nov 2009 02:48:12 -0800</pubDate>
		<guid>http://developer.nvidia.com/forums/index.php?showtopic=3796</guid>
	</item>
	<item>
		<title>how to disable a joint?</title>
		<link>http://developer.nvidia.com/forums/index.php?showtopic=3795</link>
		<description><![CDATA[i have create a joint between two object A and B.<br />but sometimes I want to disable the joint. just as there is no joint between A and B before..<br />and sometimes I want to restore the joint again  .<br /><br />who can give me some ideas ? <br />best give me some code.<br /><br />thank you<br /><br />]]></description>
		<pubDate>Mon, 02 Nov 2009 00:25:55 -0800</pubDate>
		<guid>http://developer.nvidia.com/forums/index.php?showtopic=3795</guid>
	</item>
	<item>
		<title>Jittery Movement</title>
		<link>http://developer.nvidia.com/forums/index.php?showtopic=3794</link>
		<description><![CDATA[Hello, I am using a modified version of the character controller code found in the PhysX sample code. I created the game's world by grabbing visible triangles from a BSP map and cooking them.<br /><br />My problem is that whenever the player walks across two triangles on the ground, it stutters. I looked in the remote debugger very closely and right as the player walks in between them, the directional vector (in remote debugger) points down, then up again, as if it were falling in between the two triangles, but the triangles are all touching, there is no gap between them.<br /><br />A friend suggested that I use capsule sweeps, but I ran into roadblocks when implementing it because it would cause me to lose a lot of the functionality that the default character controller offers. Do any of you guys know of a better way of doing character movement than what is shown in the PhysX samples?<br /><br />I uploaded the PXD showing the issue.<br /><br />PLEASE help me, I've been stuck with this issue for months! <img src="http://developer.nvidia.com/forums/style_emoticons/default/mad.gif" style="vertical-align:middle" emoid=":angry:" border="0" alt="mad.gif" />]]></description>
		<pubDate>Mon, 02 Nov 2009 00:18:07 -0800</pubDate>
		<guid>http://developer.nvidia.com/forums/index.php?showtopic=3794</guid>
	</item>
	<item>
		<title>NVIDIA System Tools with ESA Support Trouble</title>
		<link>http://developer.nvidia.com/forums/index.php?showtopic=3793</link>
		<description><![CDATA[Hi there, I just downloaded the system tools and every time I try to open up the System Monitor it crashes and stops working. I have a Compaq Presario CQ60 AMD Athlon x2 64 bit with a NVIDIA GeForce 8200M G Graphics card. I'm not to sure whats going here but any help would be greatly appreciated. <img src="http://developer.nvidia.com/forums/style_emoticons/default/sad.gif" style="vertical-align:middle" emoid=":(" border="0" alt="sad.gif" />]]></description>
		<pubDate>Sun, 01 Nov 2009 08:18:24 -0800</pubDate>
		<guid>http://developer.nvidia.com/forums/index.php?showtopic=3793</guid>
	</item>
	<item>
		<title>nvidia geforce 6200</title>
		<link>http://developer.nvidia.com/forums/index.php?showtopic=3792</link>
		<description><![CDATA[cant see my bios when i reboot and i see dotted lines running through a lot of the simpler graphics online and on my computer.<br /><br />I originally tried reinstalling the drivers that didnt help i then updated the drivers still same probelm. <br />using a HP Compaq Presario with Windows XP Pro x86 OS]]></description>
		<pubDate>Sun, 01 Nov 2009 00:38:30 -0700</pubDate>
		<guid>http://developer.nvidia.com/forums/index.php?showtopic=3792</guid>
	</item>
	<item>
		<title>Cloth attachment in Max plugin</title>
		<link>http://developer.nvidia.com/forums/index.php?showtopic=3791</link>
		<description><![CDATA[Hi, I'm a newbie of PhysX and I want to build a demo to show some cloth effect in it.  As it is a little difficult for me to attach cloth in the code, I <br />decide to do it in the Max.  After several attempts, I found that  for now the plugin(1.06) only support auto-attach cloth to primitive type object. <br />    <br />      But I found there is some thing strange. For example, when I create a box(or a sphere) which the cloth is attached to  without change its shape, the <br />plugin can work properly, but if I just scale it to fit the cloth the auto-attach don't work. I've checked the exported xml file to find that if I scale the box then the plugin will export it as a trianglemesh. But it's very inconvient to only decide the shape box when it is created to get all these things ran properly. Is there someway I can do to hack this to let the plugin export my box as primitive after I scale it?(just like the way set "isConcave = 1" to export triangleMesh).<br /><br />    Sorry for my poor English.]]></description>
		<pubDate>Sun, 01 Nov 2009 00:20:40 -0700</pubDate>
		<guid>http://developer.nvidia.com/forums/index.php?showtopic=3791</guid>
	</item>
	<item>
		<title>Hammering task problem being a softbody problem or not?</title>
		<link>http://developer.nvidia.com/forums/index.php?showtopic=3790</link>
		<description><![CDATA[Sorry for posting this question here but i do not know where to post it otherwise. So here is my problem:<br />I want to simulate a hammering task, where an arm tries to hit a nail with a hammer in order to put the nail into a board. The nail should be deformable, the hammer obviously not. But i am not sure if it is best to simulate this task using softbody dynamics or not?<br /><br />I am new to physx but not to physics because i study physics about four years now.<br /><br />So maybe someone can help me.<br /><br />Another question: Do you think that this task is rather complicated or not (taking the dynamic constraints of the board to the nail into account during nailing).<br /><br />Thanks,<br />Thomas<br />]]></description>
		<pubDate>Sat, 31 Oct 2009 05:58:09 -0700</pubDate>
		<guid>http://developer.nvidia.com/forums/index.php?showtopic=3790</guid>
	</item>
	<item>
		<title>FX composer no loading alpha channel in textures</title>
		<link>http://developer.nvidia.com/forums/index.php?showtopic=3789</link>
		<description><![CDATA[The latest version of FX Composer 2.5 (2.52.0319.1250) loads all textures with the alpha component set to 1.0 (255).  I initially thought I had a bug with my shaders, but then realized that the material window was showing a value of 1.0 for all my texture's alpha.  I have tried to use 32-bit bitmaps as well as DDS textures and the results are always the same.  I do however get correct results with FX Composer 1.8 (1.8.0815.0850.7772). Is there a bug with texture loading or am I missing something simple?]]></description>
		<pubDate>Fri, 30 Oct 2009 23:20:10 -0700</pubDate>
		<guid>http://developer.nvidia.com/forums/index.php?showtopic=3789</guid>
	</item>
	<item>
		<title>NVIDIA 8600M GT DRIVER ISSUE ON DELL INSPIRON 1720</title>
		<link>http://developer.nvidia.com/forums/index.php?showtopic=3788</link>
		<description><![CDATA[Looking for a bit of help....<br /><br />I tried to connect my laptop to my TV via a S-VIDEO to TV-OUT connector cable and when I glanced at my laptop screen a checkerboard effect was displayed made up of blocks of blue/green pixel squares. When I delete the driver for the graphics card and restart my laptop using the integrated VGA graphics adaptor everything seems to work correctly (minus graphics capabilities). I tried installing a different driver but every time I try one and restart my computer the problem reappears. Any help would be greatly appreciated!<br /><br />Click on image below for screenshot of problem:<br /><a href="http://img691.imageshack.us/i/desktopu.jpg/" target="_blank"><img src="http://img691.imageshack.us/img691/2005/desktopu.th.jpg" border="0" class="linked-image" /></a>]]></description>
		<pubDate>Fri, 30 Oct 2009 02:39:13 -0700</pubDate>
		<guid>http://developer.nvidia.com/forums/index.php?showtopic=3788</guid>
	</item>
	<item>
		<title>NVIDIA 8600M GT DRIVER ISSUE ON DELL INSPIRON 1720</title>
		<link>http://developer.nvidia.com/forums/index.php?showtopic=3787</link>
		<description><![CDATA[Looking for a bit of help....<br /><br />I tried to connect my laptop to my TV via a S-VIDEO to TV-OUT connector cable and when I glanced at my laptop screen a checkerboard effect was displayed made up of blocks of blue/green pixel squares. When I delete the driver for the graphics card and restart my laptop using the integrated VGA graphics adaptor everything seems to work correctly (minus graphics capabilities). I tried installing a different driver but every time I try one and restart my computer the problem reappears. Any help would be greatly appreciated!<br /><br />Click on image below for screenshot of problem:<br /><a href="http://img691.imageshack.us/i/desktopu.jpg/" target="_blank"><img src="http://img691.imageshack.us/img691/2005/desktopu.th.jpg" border="0" class="linked-image" /></a>]]></description>
		<pubDate>Fri, 30 Oct 2009 02:37:24 -0700</pubDate>
		<guid>http://developer.nvidia.com/forums/index.php?showtopic=3787</guid>
	</item>
	<item>
		<title>Driving GLSL shaders with Cg runtime?</title>
		<link>http://developer.nvidia.com/forums/index.php?showtopic=3786</link>
		<description><![CDATA[Is it possible to load, compile, and drive GLSL shaders using the Cg runtime?<br /><br />Anyone got an example?<br /><br />Is there anything special you have to do to the GLSL source to get vertex attribute names bound to vertex attribute slots?<br /><br />Thanks.]]></description>
		<pubDate>Thu, 29 Oct 2009 18:19:41 -0700</pubDate>
		<guid>http://developer.nvidia.com/forums/index.php?showtopic=3786</guid>
	</item>
	<item>
		<title>Driver issues help please (nvlddmkm)</title>
		<link>http://developer.nvidia.com/forums/index.php?showtopic=3785</link>
		<description>Ok so I have a 64 bit Windows vista computer with a NVIDIA GeForce 8800 GT graphics card. I have up until recently had a nVida driver installed until it gave me the above (nvlddmkm). So I uninstalled it and now my computers working again. Only problem is now im left without a driver(Besides the basic useless one) and everytime I download a new nVida driver I get the BSOD. Help please</description>
		<pubDate>Thu, 29 Oct 2009 13:43:58 -0700</pubDate>
		<guid>http://developer.nvidia.com/forums/index.php?showtopic=3785</guid>
	</item>
	<item>
		<title>Exact Collisions</title>
		<link>http://developer.nvidia.com/forums/index.php?showtopic=3784</link>
		<description><![CDATA[Hi,<br /><br />I have a human model and when I throw a ball at it I want the point of the collision between the ball and the model on the surface of the model as exact as possible. The ball should not intersect the model or bounce against something in the air in front of the model.<br /><br />This would be no problem with a model that is not animated but with animations I cant use just one single shape to represent the human model in the physics engine. <br /><br />At the moment the only Iiea I have is using Kinematic Actors and set them in every frame on the Position of the bones of the model. But this would not always prevent the ball from intersecting with the body because the actors dont have the exact same shape as the graphic.<br /><br />My hope is anyone can provide some good Ideas or only hints that help me finding something that works better. Any answer is welcome <br /><br />greetings]]></description>
		<pubDate>Thu, 29 Oct 2009 10:20:06 -0700</pubDate>
		<guid>http://developer.nvidia.com/forums/index.php?showtopic=3784</guid>
	</item>
	<item>
		<title>Precomputing/Calculating</title>
		<link>http://developer.nvidia.com/forums/index.php?showtopic=3783</link>
		<description><![CDATA[Hello,<br /><br />I need to calculate the future Position of a flying object. To be exact I have a ball that is thrown from different positions and I want to be able to tell where the ball will be at a certain timestep assumed no other object interacts with the ball in the meantime.<br /><br />I also need the counterpart of the previous problem. I need to know the force I have to set/add on the ball to make it fly to the position I want. Even better would be if I could tell when the ball should be there. There is probably also the angle of the force important.<br /><br />Is there already any Methode/Function to do this in PhysX? If not, perhaps someone knows a 3rd party lib that helps with this kind of calculations or another forum where I can get help with this kind of problem? (I imagine it rather problematic to predict this by myself because I have to use the same algorithms that PhysX uses to get the same results. Simply opening a physics book probably won´t help) <br /><br />Any help and advice is welcome.<br /><br />greetings<br />]]></description>
		<pubDate>Thu, 29 Oct 2009 08:12:31 -0700</pubDate>
		<guid>http://developer.nvidia.com/forums/index.php?showtopic=3783</guid>
	</item>
	<item>
		<title>NXU error</title>
		<link>http://developer.nvidia.com/forums/index.php?showtopic=3782</link>
		<description><![CDATA[hi<br />i am using a .nxb file for 3d modeling in training program 101 where i used include file "NXU_Helper.h"<br /><br />my function is<br /><br /><!--c1--><div class='codetop'>CODE</div><div class='codemain'><!--ec1-->void LoadModel&#40;&#41;<br />&nbsp;&nbsp;&nbsp;&nbsp;{<br />&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;NXU&#58;&#58;NxuPhysicsCollection *c = NXU&#58;&#58;loadCollection&#40;&#34;mymodel.nxb&#34;, NXU&#58;&#58;FT_BINARY &#41;;<br /><br />&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;if &#40; c &#41;<br />&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;{<br />&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;<br />&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;NXU&#58;&#58;instantiateCollection&#40; c, *gPhysicsSDK, gScene, 0, 0 &#41;;<br />&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; <br />&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;NXU&#58;&#58;releaseCollection&#40;c&#41;;<br />&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;}<br />&nbsp;&nbsp;&nbsp;&nbsp;}<!--c2--></div><!--ec2--><br /><br />but i get a linker error..like this.....<br /><br /><!--c1--><div class='codetop'>CODE</div><div class='codemain'><!--ec1-->1&#62;------ Build started&#58; Project&#58; Lesson101_Primary_Shape, Configuration&#58; Debug Win32 ------<br />1&#62;Linking...<br />1&#62;Lesson101.obj &#58; error LNK2019&#58; unresolved external symbol &#34;void __cdecl NXU&#58;&#58;releaseCollection&#40;class NXU&#58;&#58;NxuPhysicsCollection *&#41;&#34; &#40;?releaseCollection@NXU@@YAXPAVNxuPhysicsCollection@1@@Z&#41; referenced in function &#34;void __cdecl LoadLand&#40;void&#41;&#34; &#40;?LoadLand@@YAXXZ&#41;<br />1&#62;Lesson101.obj &#58; error LNK2019&#58; unresolved external symbol &#34;bool __cdecl NXU&#58;&#58;instantiateCollection&#40;class NXU&#58;&#58;NxuPhysicsCollection *,class NxPhysicsSDK &,class NxScene *,class NxMat34 const *,class NXU_userNotify *&#41;&#34; &#40;?instantiateCollection@NXU@@YA_NPAVNxuPhysicsCollection@1@AAVNxPhysicsSDK@@PAVNxS<br />cene@@PBVNxMat34@@PAVNXU_userNotify@@@Z&#41; referenced in function &#34;void __cdecl LoadLand&#40;void&#41;&#34; &#40;?LoadLand@@YAXXZ&#41;<br />1&#62;Lesson101.obj &#58; error LNK2019&#58; unresolved external symbol &#34;class NXU&#58;&#58;NxuPhysicsCollection * __cdecl NXU&#58;&#58;loadCollection&#40;char const *,enum NXU&#58;&#58;NXU_FileType,void *,int&#41;&#34; &#40;?loadCollection@NXU@@YAPAVNxuPhysicsCollection@1@PBDW4NXU_FileType@1@PAXH@Z&#41; referenced in function &#34;void __cdecl LoadLand&#40;void&#41;&#34; &#40;?LoadLand@@YAXXZ&#41;<br />1&#62;&#46;&#46;/&#46;&#46;/&#46;&#46;/&#46;&#46;/&#46;&#46;/bin/Win32/Lesson101_Primary_Shape_DEBUG.exe &#58; fatal error LNK1120&#58; 3 unresolved externals<br />1&#62;Build log was saved at &#34;file&#58;//c&#58;&#092;NVIDIA PhysX SDKv2.8.1&#092;TrainingPrograms&#092;Programs&#092;Chapter1_Rigid_Bodies&#092;compiler&#092;vc8win32&#092;build&#092;Win32&#092;Lesson101_Primary_Shape&#092;Debug&#092;BuildLog.htm&#34;<br />1&#62;Lesson101_Primary_Shape - 4 error&#40;s&#41;, 0 warning&#40;s&#41;<br />========== Build&#58; 0 succeeded, 1 failed, 0 up-to-date, 0 skipped ==========<!--c2--></div><!--ec2--><br /><br />i cant get why these errors are?????<br /><br />thanx....<br />]]></description>
		<pubDate>Thu, 29 Oct 2009 07:25:31 -0700</pubDate>
		<guid>http://developer.nvidia.com/forums/index.php?showtopic=3782</guid>
	</item>
	<item>
		<title>Compile problems</title>
		<link>http://developer.nvidia.com/forums/index.php?showtopic=3780</link>
		<description><![CDATA[I am a newly cg learner from China. now confront with some problems. I had consulted with many searching engines, but with few achievements, now I list them as follows:<br />At platform vs 2005:<br /> I added a external tools named CG vertex compile, then get problem 1:<br />(0) : error C0000: syntax error, unexpected $end at token "&lt;EOF&gt;"<br />(0) : error C0501: type name expected at token "&lt;null atom&gt;"<br />Then, after clicking the right button,  I set the propeties of the cg file, filing the command line with:"$(CG_BIN_PATH)&#092;cgc.exe" "$(InputPath)" -profile arbvp1 -o out.vp<br />and the Outputs line with: $(OutDir)<br />finishing this, I compiled the cg file, get problem 2:<br />error PRJ0019: A tool returned an error code from "Performing Custom Build Step"<br />At platform vc 6.0:<br />the problems are:<br />problem 1:<br />(0) : fatal error C9999: Can't open file: D:&#092;C++WorkLab&#092;cg&#092;raycasting_shader.cg-profile<br />Tool returned code: 1<br /><br />problem 2:<br />Performing Custom Build Step on .&#092;raycasting_shader.cg<br />.&#092;raycasting_shader.cg<br />.&#092;raycasting_shader.cg(63) : warning C7011: implicit cast from "float4" to "float3"<br />(0) : error C3001: no program defined<br />84 lines, 1 warnings, 1 errors.<br />Error executing c:&#092;windows&#092;system32&#092;cmd.exe.<br /><br />I am extremly confused with these errors, any thanks would be greatful.<br />]]></description>
		<pubDate>Thu, 29 Oct 2009 00:17:30 -0700</pubDate>
		<guid>http://developer.nvidia.com/forums/index.php?showtopic=3780</guid>
	</item>
	<item>
		<title>OpenCL for Cuda paper</title>
		<link>http://developer.nvidia.com/forums/index.php?showtopic=3779</link>
		<description><![CDATA[Hello<br />I would like to ask if anybody has tried examples from the OpenCL (Programming for the CUDA Architecture) pdf document?<br /><br />m]]></description>
		<pubDate>Wed, 28 Oct 2009 17:04:55 -0700</pubDate>
		<guid>http://developer.nvidia.com/forums/index.php?showtopic=3779</guid>
	</item>
	<item>
		<title>Its impossible to install PhysX plugin in 3ds Max (2009 32bit)</title>
		<link>http://developer.nvidia.com/forums/index.php?showtopic=3778</link>
		<description><![CDATA[Im totally bored now :/ . . . First during installation the installer says that he can't open the file C:&#092;Windows&#092;System32&#092;nvRegDevWin32.exe and sometimes a few others argh...but strangely if a click ignore a few times it will install but then, while opening 3ds Max, I get an error message that tells that the file ...&#092;Autodesk&#092;3ds Max 2009&#092;plugins&#092;pxplugin.dlm couldn't be found, but this file is exactly there!! WTF IS THAT?!? OH MY<br /><br />Thanks for any help..]]></description>
		<pubDate>Wed, 28 Oct 2009 11:00:44 -0700</pubDate>
		<guid>http://developer.nvidia.com/forums/index.php?showtopic=3778</guid>
	</item>
	<item>
		<title>3DS Plugin revolute joints</title>
		<link>http://developer.nvidia.com/forums/index.php?showtopic=3777</link>
		<description><![CDATA[Hi,<br /><br />Using the PhysX plugin for 3D studio max I specify revolute joints on some actors.<br />When I import the actors using nxustream2 the  NXU_notifyJoint (NxJoint *joint, const char *userProperties) callback gives me a 6DOF joint.<br />Thus, joint-&gt;isRevoluteJoint() returns null.<br /><br />Why is this?<br /><br />It seems the degrees of freedom are locked into the same dof's as a revolute joint, so it does behave like one.<br />I guess it doesn't matter too much since the 6DOF joint supports the same functionality with slightly different function calls.<br /><br />]]></description>
		<pubDate>Wed, 28 Oct 2009 08:59:59 -0700</pubDate>
		<guid>http://developer.nvidia.com/forums/index.php?showtopic=3777</guid>
	</item>
	<item>
		<title>cannot locate suitable resource to bind parameter error</title>
		<link>http://developer.nvidia.com/forums/index.php?showtopic=3776</link>
		<description><![CDATA[<!--c1--><div class='codetop'>CODE</div><div class='codemain'><!--ec1-->struct Light<br />{<br />float4 Color;<br />float4 Position;<br />...<br />};<br />uniform Light lights&#91;256&#93;;<!--c2--></div><!--ec2--><br />results in "error C5041: cannot locate suitable resource to bind parameter "lights". I use gp4fp profiles.<br />AFAIK constant buffers in DX10 are unlimited size. So should be OpenGL uniform buffer objects. Doesn't Cg use them?]]></description>
		<pubDate>Wed, 28 Oct 2009 04:12:20 -0700</pubDate>
		<guid>http://developer.nvidia.com/forums/index.php?showtopic=3776</guid>
	</item>
	<item>
		<title>need some help to simulate plane flying</title>
		<link>http://developer.nvidia.com/forums/index.php?showtopic=3775</link>
		<description><![CDATA[there is some examples about vehicles in physx . <br />so , the physx give us some examples to simulate car driving .<br /><br />but now , I want to do something about flying.  I want to use physx to simulate a plane flying ..and I can control it fly up or down, turn left or right . the question is that,  when you control a plane to turn left, it is not like a simple rigid body just turning left stiff .  but the plane should have some radian in yaw angle or pitch attitude or rotation..<br /><br />I need some suggestion .   thank you all.<br />]]></description>
		<pubDate>Tue, 27 Oct 2009 20:10:13 -0700</pubDate>
		<guid>http://developer.nvidia.com/forums/index.php?showtopic=3775</guid>
	</item>
	<item>
		<title>NxSceneDesc::groundPlane = ?</title>
		<link>http://developer.nvidia.com/forums/index.php?showtopic=3774</link>
		<description><![CDATA[I can't tell the difference between a scene created with NxSceneDesc::groundPlane = true/false. The doc only states it's a parameter for using a "default ground plane". Does anyone have any detail on what the default ground plane actually is? Is it an automatically created static plane shaped actor? In my minimal application groundPlane=true doesn't seem to produce anything.<br />Thanks,<br />-n]]></description>
		<pubDate>Tue, 27 Oct 2009 18:10:40 -0700</pubDate>
		<guid>http://developer.nvidia.com/forums/index.php?showtopic=3774</guid>
	</item>
	<item>
		<title>help nvidia plugins photoshop dds</title>
		<link>http://developer.nvidia.com/forums/index.php?showtopic=3773</link>
		<description><![CDATA[hi i am new here and i broken my head all the time!my problem is in the nvidia photoshop plug-ins....<br />when i save my texture in photoshop cs3 256x256 on format dxt1 no alpha the size up to 42kb and no 32kb.<br />i need the file dds will be 32 kb but its allways 42kb...what to do what the problem???<br />someone can&#65279; help me please...]]></description>
		<pubDate>Tue, 27 Oct 2009 09:01:22 -0700</pubDate>
		<guid>http://developer.nvidia.com/forums/index.php?showtopic=3773</guid>
	</item>
	<item>
		<title>cgValidateTechnique fails with Cg 2.2</title>
		<link>http://developer.nvidia.com/forums/index.php?showtopic=3772</link>
		<description><![CDATA[Hi all,<br /><br />I had to upgrade to the latest Cg Toolkit (v2.2) to get the bug fix which makes cgGLGetTextureEnum work with effect params.  Unfortunately, now my shader won't validate.  Cg gives me no errors at all.  I can switch back and forth between 2.2 and 2.0 and it validates in the older version with no problems.<br /><br />Can anyone tell me what I might be doing wrong or some way to get more info about why it is failing?<br /><br />Here is a snippit of the technique...<br /><div class='codetop'>CODE</div><div class='codemain' style='height:200px;white-space:pre;overflow:auto'>technique Main<br />{<br />	pass MakeShadow<br />	{<br />		VertexProgram = compile vp40 shadVS( gWorldITXf, gWorldXf, gViewIXf, gWvpXf,<br />			gLampViewXf, mul(gLampProjXf, gLampViewXf), gNearClip, gClipSize );<br />		DepthTestEnable = true;<br />		DepthMask = true;<br />		CullFaceEnable = false;<br />		BlendEnable = false;<br />		DepthFunc = LEqual;<br />		FragmentProgram = compile arbfp1 shadPS();<br />	}<br />	pass DrawShadow<br />	{<br />		VertexProgram = compile arbvp1 drawShadVS(gWvpXf);<br />		DepthTestEnable = true;<br />		DepthMask = true;<br />		CullFaceEnable = false;<br />		BlendEnable = false;<br />		DepthFunc = LEqual;<br />		FragmentProgram = compile arbfp1 drawShadPS( gShadSampler );<br />	}<br />	. . .<br />}</div><br /><br />It is serious crunch time here and these Cg issues are killing my productivity.<br /><br />Randy Croucher<br />Disney Interactive Studios]]></description>
		<pubDate>Tue, 27 Oct 2009 08:54:35 -0700</pubDate>
		<guid>http://developer.nvidia.com/forums/index.php?showtopic=3772</guid>
	</item>
	<item>
		<title>Help Me.. Where do i begin?</title>
		<link>http://developer.nvidia.com/forums/index.php?showtopic=3771</link>
		<description><![CDATA[hello im a 3d artist and i saw the real time animation "Medusa" from nvidia coolstuff i just wanted to know wer to start to make that cool realtime animation and what other things i should be studying about. thanks in advance]]></description>
		<pubDate>Tue, 27 Oct 2009 08:35:58 -0700</pubDate>
		<guid>http://developer.nvidia.com/forums/index.php?showtopic=3771</guid>
	</item>
	<item>
		<title>Question regarding heightfield.</title>
		<link>http://developer.nvidia.com/forums/index.php?showtopic=3770</link>
		<description><![CDATA[Hello,<br /><br />First off, if this has been covered before, please excuse me.<br />I tried searching before, of course, but there is always the possibility that I missed it.<br /><br />Anyway, I was wondering about what the correct method would be to implement heightfield(s) in a tool such as a map editor that has dynamic terrain?<br />Maybe my english terminology is off,  but I assume this to be dynamic heightfields?<br />But from what I've read online people are saying that dynamic heightfields with PhysX are impossible.<br />So I'm guessing that my best option would be to create multiple heightfields which I then create/destroy as the terrain is being edited?<br />Is there a more efficient or even simpler way of achieving this, or am I overlooking another option entirely?<br /><br />Any advice is appreciated.]]></description>
		<pubDate>Tue, 27 Oct 2009 05:41:17 -0700</pubDate>
		<guid>http://developer.nvidia.com/forums/index.php?showtopic=3770</guid>
	</item>
	<item>
		<title>PTX</title>
		<link>http://developer.nvidia.com/forums/index.php?showtopic=3769</link>
		<description><![CDATA[I am looking to write a high performance library utilizing PTX. It is possible for me to use an emulator, but for performance reasons I would prefer to know the details of cubin.bin, including its opcodes.<br />Is this possible?]]></description>
		<pubDate>Sun, 25 Oct 2009 17:59:07 -0700</pubDate>
		<guid>http://developer.nvidia.com/forums/index.php?showtopic=3769</guid>
	</item>
	<item>
		<title>propeller CCD</title>
		<link>http://developer.nvidia.com/forums/index.php?showtopic=3768</link>
		<description><![CDATA[If I got a helicopter propeller, should I use circle shape for CCD while it turn around?<br />And should I disable the base shape actor collision to improve performance?<br />How could I do it well?]]></description>
		<pubDate>Sat, 24 Oct 2009 11:06:53 -0700</pubDate>
		<guid>http://developer.nvidia.com/forums/index.php?showtopic=3768</guid>
	</item>
	<item>
		<title>geometry and physx</title>
		<link>http://developer.nvidia.com/forums/index.php?showtopic=3766</link>
		<description><![CDATA[hello, <br />Im a newbie <img src="http://developer.nvidia.com/forums/style_emoticons/default/smile.gif" style="vertical-align:middle" emoid=":)" border="0" alt="smile.gif" /> and i have just succeded to make a simple boxes app with physx and<br />please escuse me if i ask something stupid or i make mistakes.<br /><br />I have in my app a mesh (triangles) with nearly 50000 poly's. The mesh is subdivided for faster rendering (octree subdibision) with frusutm culling.<br />The character (player) moves in this level (mesh/map) like in a 3d first person shooter type camera. <br />I want collision dectection and response (and simple phyisics like bounce off the wall) between the character and this mesh (geometry / level).<br /><br />Now, my question: <br /><br /> - i cannot pass to physx all 50000 poly's because i think it will be slow (too many polygons to test collision and simulate and too many calculations) <br />   so i would prefer to pass only the current node poly's (player's current node) wich has only 1000 - 2000 polys max. <br />  <br />Is this possible ? (note: the player's position changes and new polys must be feed to the physx engine to test collision against)<br /><br />thank you.]]></description>
		<pubDate>Sat, 24 Oct 2009 00:01:19 -0700</pubDate>
		<guid>http://developer.nvidia.com/forums/index.php?showtopic=3766</guid>
	</item>
	<item>
		<title>Textures and PhysX</title>
		<link>http://developer.nvidia.com/forums/index.php?showtopic=3765</link>
		<description><![CDATA[Hi,<br /><br />I am very new to PhysX and there are a couple of questions that are making things a little hard for me and I couldn't find any solutions online. Do try to help keeping in mind that I am very new to this.<br /><br />a. I am aware that PhysX is a "middleware" and that it is to be used to model the physical behaviour of an object. Hence, textures do not fall into it's purview. However, I would then like to know what one needs to do to solve the following problem.<br /><br />How can I render an image (say done by openGL) and then add phyics to it using PhysX? Say I define a box with very little thickness (NxBoxShape) and also an openGL cube, how do I "tag/attach" the physics to the openGL textured object?<br /><br />Any help would be really appreciated. <br /><br />Thanks for your time]]></description>
		<pubDate>Fri, 23 Oct 2009 14:09:54 -0700</pubDate>
		<guid>http://developer.nvidia.com/forums/index.php?showtopic=3765</guid>
	</item>
	<item>
		<title>Continuous angular limit</title>
		<link>http://developer.nvidia.com/forums/index.php?showtopic=3764</link>
		<description><![CDATA[I would like some info on where in the PhysX documentation I should look to figure out how to make a ANGULAR_LIMITED joint that its limit is greater than -PI and PI?<br /><br />I have a robot arm where the main wrist is able to rotate in either direction 10 revolutions.  It is important for the software to know where within its rotation the wrist is (a scale between -1 and 1 from -10 revolution to +10 revolution).  How should I go about doing this?<br /><br />Thanks for any suggestions/directions/help.<br /><br />Regards,<br /><br />Bob2u]]></description>
		<pubDate>Thu, 22 Oct 2009 12:27:24 -0700</pubDate>
		<guid>http://developer.nvidia.com/forums/index.php?showtopic=3764</guid>
	</item>
	<item>
		<title>Cg uniform int compiled to float register</title>
		<link>http://developer.nvidia.com/forums/index.php?showtopic=3763</link>
		<description><![CDATA[I am using Cg with DX minimal interface (compiling to assembly then using DX for setting up the shader and passing parameters).  I ran into a problem where I would pass a 'uniform int' to the shader, but when I used SetVertexShaderConstantI() (based on the fact that cgGetParameterType() returns a parameter type of CG_INT after compiling), but the value wasn't being set.  Looking closer at the assembly that Cg provided, it is putting all uniform ints into float registers.  Is this as-expected (I can work around it easily by using SetVertexShaderConstantF as long as I can count on this), or is this controllable, or is there a way to get the correct register type that was used for the parameter?<br /><br />Thanks.<br /><br />Here is sample Pixel shader:<br /><br /><!--fonto:Courier New--><span style="font-family:Courier New"><!--/fonto--> struct PixelIn<br /> {<br />   float4 Position : POSITION;<br />   float2 TexCoord0 : TEXCOORD0; <br /> };<br /> <br /> struct PixelOut<br /> {<br />   float4 Color : COLOR;<br /> };<br /> <br /> PixelOut main(PixelIn pixelIn, uniform sampler2D sampler0, uniform float pixelHeight, uniform int rowAdd)<br /> {<br />   PixelOut pixelOut;<br />  int row = pixelIn.TexCoord0.y / pixelHeight;<br />  row += rowAdd;<br />  pixelIn.TexCoord0.y = row * pixelHeight;<br />   pixelOut.Color = tex2D(sampler0, pixelIn.TexCoord0.xy);<br /> <br />   return pixelOut;<br /> }<br /><!--fontc--></span><!--/fontc--><br /><br />and the assembly produced (with profile ps_3_0)<br /><!--fonto:Courier New--><span style="font-family:Courier New"><!--/fonto-->// cgc version 2.2.0010, build date Sep 29 2009<br />// command line args: -profile ps_3_0<br />// source file: test.txt<br />//vendor NVIDIA Corporation<br />//version 2.2.0.10<br />//profile ps_3_0<br />//program main<br />//semantic main.sampler0<br />//semantic main.pixelHeight<br />//semantic main.rowAdd<br />//var float2 pixelIn.TexCoord0 : $vin.TEXCOORD0 : ATTR0 : 0 : 1<br />//var sampler2D sampler0 :  : texunit 0 : 1 : 1<br />//var float pixelHeight :  : c[0] : 2 : 1<br />//var int rowAdd :  : c[1] : 3 : 1<br />//var float4 main.Color : $vout.COLOR : COL : -1 : 1<br />dcl_2d s0<br />dcl_texcoord0 v0.xy<br />rcp r0.x, c0.x<br />mul r0.x, v0.y, r0<br />abs r0.y, r0.x<br />frc r0.z, r0.y<br />add r0.y, r0, -r0.z<br />cmp r0.x, r0, r0.y, -r0.y<br />add r0.y, r0.x, c1.x<br />mov r0.x, v0<br />mul r0.y, r0, c0.x<br />texld r0, r0, s0<br />mov oC0, r0<br /><!--fontc--></span><!--/fontc-->]]></description>
		<pubDate>Thu, 22 Oct 2009 10:10:37 -0700</pubDate>
		<guid>http://developer.nvidia.com/forums/index.php?showtopic=3763</guid>
	</item>
	<item>
		<title>PhysX in windows 7</title>
		<link>http://developer.nvidia.com/forums/index.php?showtopic=3762</link>
		<description>Currently there is no drivers that are supporting windows 7 it seems couse i dl the latest ones and windows 7 doesnt accept them=(</description>
		<pubDate>Thu, 22 Oct 2009 00:59:14 -0700</pubDate>
		<guid>http://developer.nvidia.com/forums/index.php?showtopic=3762</guid>
	</item>
	<item>
		<title>Bat Swing</title>
		<link>http://developer.nvidia.com/forums/index.php?showtopic=3761</link>
		<description><![CDATA[I'm currently trying to  create a nice smooth bat swinging motion that is controlled by a player. While the currently the character moves around the bat will swing fine but when I rotate the character the bat gets stuck facing forward.    I've implemented the system by:<br /><br />Create a box shaped actor around the bat,<br />The Player is controlled by a Capsule Controller which controls the movement of a the camera.<br />There is another   actor which is positioned with an offset from the camera, and is given the camera's rotation.<br /><br />I then create an Revolution Joint with an X-Axis Rotation, so that it pitches up and down.   The joint has motor to keep it swinging, and limits to keep it only moving from 0 to 2.4 Radians. When it has moved its full distance I reset the bat's position to point back up.<br /><br />to update everything<br />1. Move the capsule controller, and rotate the camera.  (The capsule controller moves the camera).<br />2.  Set the camera actor with the new camera information<br />3. Set the position of the bat to reflect any movement and rotation, so that it stays in the same place relative to the player. <br /><br /><br />Thanks]]></description>
		<pubDate>Wed, 21 Oct 2009 16:26:03 -0700</pubDate>
		<guid>http://developer.nvidia.com/forums/index.php?showtopic=3761</guid>
	</item>
	<item>
		<title>moveGlobalPose() is like setGlobalPose()?</title>
		<link>http://developer.nvidia.com/forums/index.php?showtopic=3760</link>
		<description><![CDATA[Currently, it seems these Functions do the same thing: <br />Positioning an actor in the world, regardless of collision. Is that correct? <br /><br />So it is not possible to move an kinematic actor using moveGlobalPose() ie. on a spline and have collision for it? <br />Thanks in advance!<br />]]></description>
		<pubDate>Wed, 21 Oct 2009 09:36:57 -0700</pubDate>
		<guid>http://developer.nvidia.com/forums/index.php?showtopic=3760</guid>
	</item>
	<item>
		<title>Quadro SDI Capture SDK</title>
		<link>http://developer.nvidia.com/forums/index.php?showtopic=3759</link>
		<description><![CDATA[Hi,<br /><br />I can't seem to find the Nvidia SDI SDK. I would like to understand it in advance as I will be purchasing the Quadro SDI Capture card. Could someone point me to the link?<br /><br /><br />Thanks in advance.<br /><br />[I understand how to get it done. Thanks]]]></description>
		<pubDate>Wed, 21 Oct 2009 01:52:59 -0700</pubDate>
		<guid>http://developer.nvidia.com/forums/index.php?showtopic=3759</guid>
	</item>
	<item>
		<title>Moving an actor bound by fixed joint.</title>
		<link>http://developer.nvidia.com/forums/index.php?showtopic=3758</link>
		<description><![CDATA[Hi, I'm moving 2 actors bound by 2 respective fixed joints.<br /><br />To move the actors, what I did was releasing their respective joints, move the actors and bound them up by fixed joint again.<br /><br />The 2 actors move away and towards each other.<br /><br />I'm facing a problem though.<br /><br />One of my actor, I called it the right actor, tend to rotate as when I move. I'm sure the vector I'm adding isn't the problem as this vector is use to move other similar actors/shapes and they all looks correct.<br /><br />Is there any other issue why this can happen?<br /><br />I didn't set the anchor point though but i don't think that's critical.]]></description>
		<pubDate>Tue, 20 Oct 2009 21:08:49 -0700</pubDate>
		<guid>http://developer.nvidia.com/forums/index.php?showtopic=3758</guid>
	</item>
	<item>
		<title>Bug in OpenGL driver ?!</title>
		<link>http://developer.nvidia.com/forums/index.php?showtopic=3757</link>
		<description><![CDATA[Hello!<br />There is opinion, that function glClearBufferfv doesnt work at desktop cards in Windows x64. This func. doesnt work for me at 3 different desktop cards(for example 9800GT) and works properly at my laptop with 8700GT. I've sent "bug report", but not sure, that i did it rightt(because i have not received any response)  <img src="http://developer.nvidia.com/forums/style_emoticons/default/blink.gif" style="vertical-align:middle" emoid=":blink:" border="0" alt="blink.gif" /> <br /><br />Has anybody the same problem ?<br /><br />Dima.]]></description>
		<pubDate>Tue, 20 Oct 2009 16:06:09 -0700</pubDate>
		<guid>http://developer.nvidia.com/forums/index.php?showtopic=3757</guid>
	</item>
	<item>
		<title>Cloth collision response</title>
		<link>http://developer.nvidia.com/forums/index.php?showtopic=3756</link>
		<description><![CDATA[I'm doing a project involving per-finger force feedback and a data glove.<br /><br />Lets say for the sake of discussion that the hand is modelled in PhysX by kinematic spheres on the fingertips.<br />This works good for rigid bodies in the sense that I can set actor pair flags and get the normal force of collision and exert the force on the fingers of the user.<br /><br />However, it seems not that easy to get the force response for anything other than rigid bodies, like cloth or soft bodies.<br /><br />I want the user to be able to feel the force response when pressing down on a piece of stretched cloth or squeezing a soft body.<br /><br />What are your suggestions on achieving this?<br /><br />Thanks.<br /><br />PS: I only have access to the public 2.8.1 SDK.]]></description>
		<pubDate>Tue, 20 Oct 2009 10:30:01 -0700</pubDate>
		<guid>http://developer.nvidia.com/forums/index.php?showtopic=3756</guid>
	</item>
	<item>
		<title>CgFx Performance</title>
		<link>http://developer.nvidia.com/forums/index.php?showtopic=3755</link>
		<description><![CDATA[Hello! I was profiling my app and I came to conclusion that CgFx is my hot spot.<br /><br />I read some threads here about optimizing CgFx and I did the following:<br />- Disabled shared parameters (except textures, at least I want Cg to manage those for me) - gave big boost.<br />- I never call cgResetPassState() - this gives huge boost.<br />- I also did some other stuff like disabling parameter shadowing and etc. - gave tiny boost.<br /><br />But still Cg is taking about 42% of my frame time!<br /><br />Take a look at this profiler output:<br /><br />VS profiler:<br /><img src="http://img99.imageshack.us/img99/351/cgperf.png" border="0" class="linked-image" /><br /><br />PerfHUD:<br /><img src="http://img196.imageshack.us/img196/5845/cgperf2.jpg" border="0" class="linked-image" /><br />(If your application is CPU-bound, you’ll see a big gap between the yellow line (“Total Frame Time”) and the red line (“Driver Time”). - PerfHUD manual)<br /><br />What does Cg do all this time? What's going on?]]></description>
		<pubDate>Tue, 20 Oct 2009 08:22:37 -0700</pubDate>
		<guid>http://developer.nvidia.com/forums/index.php?showtopic=3755</guid>
	</item>
	<item>
		<title>Cg profile support for Direct3D 10.1 Fearture Levels for D3D9</title>
		<link>http://developer.nvidia.com/forums/index.php?showtopic=3754</link>
		<description><![CDATA[Hi,<br /><br /><br />Is there any way to enable support for Direct3D 10.1 Fearture Level like:<br />D3D10_FEATURE_LEVEL_9_1<br />D3D10_FEATURE_LEVEL_9_2<br />D3D10_FEATURE_LEVEL_9_3<br /><br />so in shader with:<br />ps/vs_4_0_level_9_3<br />ps/vs_4_0_level_9_2<br />ps/vs_4_0_level_9_1<br /><br />profiles.<br /><br />or is there other way to get Cg work with this direct3d10.1 Feature Levels?<br /><br />Please Help.<br />Best Regards<br />Marcin Musia&#322;<br /><br /><br />]]></description>
		<pubDate>Tue, 20 Oct 2009 06:59:54 -0700</pubDate>
		<guid>http://developer.nvidia.com/forums/index.php?showtopic=3754</guid>
	</item>
	<item>
		<title>GLSLv and GLSLf profiles produce uniform name collisions</title>
		<link>http://developer.nvidia.com/forums/index.php?showtopic=3753</link>
		<description><![CDATA[Using the GLSLv and GLSLf profile for a Cg vertex shader and a Cg fragment shader can lead to aliasing between uniform parameters when using both GLSL results in OpenGL. Aliasing between uniform parameters means, that there is only one uniform location in OpenGL for two independent parameters. A simple sample:<br /><br />Cg shader (vs: vertex shader, fs: fragment shader):<br /><!--c1--><div class='codetop'>CODE</div><div class='codemain'><!--ec1-->void vs&#40;<br />&nbsp;&nbsp;float4 pos &#58; POSITION,<br />&nbsp;&nbsp;out float4 output &#58; POSITION,<br />&nbsp;&nbsp;uniform float a,<br />&nbsp;&nbsp;uniform float b&#41;<br />{<br />&nbsp;&nbsp;output = a * b * pos;<br />}<br /><br />void fs&#40;<br />&nbsp;&nbsp;out float4 color &#58; COLOR,<br />&nbsp;&nbsp;uniform float a,<br />&nbsp;&nbsp;uniform float _b&#41;<br />{<br />&nbsp;&nbsp;color = a * _b;<br />}<!--c2--></div><!--ec2--><br />the compiled vertex shader:<br /><!--c1--><div class='codetop'>CODE</div><div class='codemain'><!--ec1-->uniform float _a1;<br />uniform float _b1;<br /><br /> // main procedure, the original name was vs<br />void main&#40;&#41;<br />{<br />&nbsp;&nbsp;&nbsp;&nbsp;vec4 _output;<br />&nbsp;&nbsp;&nbsp;&nbsp;_output = &#40;_a1*_b1&#41;*gl_Vertex;<br />&nbsp;&nbsp;&nbsp;&nbsp;gl_Position = _output;<br />} // main end<!--c2--></div><!--ec2--><br />and the compiled fragment shader:<br /><!--c1--><div class='codetop'>CODE</div><div class='codemain'><!--ec1-->uniform float _a1;<br />uniform float _b1;<br /><br /> // main procedure, the original name was fs<br />void main&#40;&#41;<br />{<br />&nbsp;&nbsp;&nbsp;&nbsp;vec4 _color;<br />&nbsp;&nbsp;&nbsp;&nbsp;_color = vec4&#40;_a1*_b1, _a1*_b1, _a1*_b1, _a1*_b1&#41;;<br />&nbsp;&nbsp;&nbsp;&nbsp;gl_FragColor = _color;<br />} // main end<!--c2--></div><!--ec2--><br /><br />when linking these two shaders in OpenGL into a program there will be only one location for the uniforum _a1 and only one location for the uniform _b1. For the uniform parameter "a" I could (maybe) accept it since its called "a" in Cg for vertex and fragment shader. But "b" and "_b" are clearly different in Cg but get mapped to one single uniform "_b1". (Cf. <a href="http://www.opengl.org/wiki/GLSL_:_common_mistakes#Uniform_Names_in_VS.2C_GS_and_FS" target="_blank">Uniform Names in VS, GS and FS</a>)<br /><br />Is there a way aorund this problem, or can this be fixed in the Cg compiler (maybe by producing names that do never collide between vertex shader and fragment shader).]]></description>
		<pubDate>Tue, 20 Oct 2009 04:41:27 -0700</pubDate>
		<guid>http://developer.nvidia.com/forums/index.php?showtopic=3753</guid>
	</item>
	<item>
		<title>Textures. Suggestions for a beginner please</title>
		<link>http://developer.nvidia.com/forums/index.php?showtopic=3752</link>
		<description><![CDATA[Hi <br /><br />I'm working on a research project using physx and am quite inexperienced with any sort of graphics programming. I'm having a lot of trouble getting textures loaded and working, if anyone can either point me in the direction of some documentation (I have been searching for said documentation for a long time and haven't found anything specific to physx) or give me some pointers I would be very grateful. <br /><br />For sake of simplicity, let's just say I'm working with lesson101 from the training programs that come with the sdk. I have a bitmap image, how would I go about applying that texture to one face of a box? <br /><br />I know this is really simple, but any help at all would really help...<br /><br />Thanks<br /><br />Alex]]></description>
		<pubDate>Mon, 19 Oct 2009 14:44:00 -0700</pubDate>
		<guid>http://developer.nvidia.com/forums/index.php?showtopic=3752</guid>
	</item>
	<item>
		<title>Identifying physical ports on adapter</title>
		<link>http://developer.nvidia.com/forums/index.php?showtopic=3751</link>
		<description><![CDATA[Hi,<br /><br />I've been using nvapi and nvcpl API's in what has been up till now an unsuccessful effort to identify which physical DVI port on Quadro cards like the FX380 is associated with the system logical monitors.    For some reason, it seems this is well hidden.   We have a problem with this situation, as end users of our system who might inadvertantly unplug the monitor 1 cable will find the monitors reversed if they boot up with monitor 2 followed by a second bootup upon restoring the monitor 1 connection - which would be natural sequence of events.   This reversal can be a real problem as the assigned modes get swapped along with the display contents.  The only workaround I have is to identify the properties (such as supported modes) of the plugged in monitors to try to infer which is the real physical adapter (or at least track them in a relative sense) - but this is a cludge, as the monitors might be identical.<br /><br />So my question is: is there any way to programatically identify the true physical output port (amongst the 2) on an adapter?<br /><br />Thanks.]]></description>
		<pubDate>Mon, 19 Oct 2009 08:55:14 -0700</pubDate>
		<guid>http://developer.nvidia.com/forums/index.php?showtopic=3751</guid>
	</item>
	<item>
		<title>Pool Friction and Velocity</title>
		<link>http://developer.nvidia.com/forums/index.php?showtopic=3750</link>
		<description><![CDATA[Hello.<br /><br />I´m making a pool game in Director using the physX Xtra engine.<br /><br />I´ve adjust all physX values end the game behaves very similar to reality. But I,ve noticed that when I apply a linear Impulse to the white ball in a X-Y direction, The ball is deflected as the speed decreases. This doesn´t happen when I apply the impulse only in one coordinate. Moreover, I´ve noticed that when I put the ball an felt frictions to zero the ball doesn´t deviate.<br /><br />For example:<br /><br />If I apply a linear impulse in X axis the ball rolls Ok and doesn´t deviate.<br />But if I apply on X-Y impulse, for example, (0.5,0.75,0.0) the balls tends to deviate, like the y axis takes more force.<br /><br />It seems as if the linear X velocity and linear Y velocity don´t decrease proportionally. In this case, Y linear velocity takes longer to reach zero.<br /><br />Note: When I apply an impulse in X and Y axis and this impulse take the same value in the two axis (i.e. (0.5,0.5,0.5)) the ball rolls OK.<br /><br />I hope you can help me.<br /><br />Thanks]]></description>
		<pubDate>Mon, 19 Oct 2009 03:52:25 -0700</pubDate>
		<guid>http://developer.nvidia.com/forums/index.php?showtopic=3750</guid>
	</item>
	<item>
		<title>Problems with 3D Vision</title>
		<link>http://developer.nvidia.com/forums/index.php?showtopic=3749</link>
		<description><![CDATA[I have developed a 3D game,now i want to play it with 3D Vision ,but I have some problems :when i start it in full screen model,the 3D glass started,but the game state have not change,so i can&#96;t have the 3D Vision Effect,but some games is very good, i want change my game to adapt 3D Vision buy revise my C++ code ,how should I do ?thanks.<br />i want to know how to develop the 3D program to adapt the 3D Vision ,why some game can not adapt it very well. (all the game was developed buy DirectX),thanks]]></description>
		<pubDate>Sun, 18 Oct 2009 22:38:23 -0700</pubDate>
		<guid>http://developer.nvidia.com/forums/index.php?showtopic=3749</guid>
	</item>
	<item>
		<title>Assignment of Uniform Array</title>
		<link>http://developer.nvidia.com/forums/index.php?showtopic=3748</link>
		<description><![CDATA[<!--fonto:Courier New--><span style="font-family:Courier New"><!--/fonto--><!--sizeo:4--><span style="font-size:14pt;line-height:100%"><!--/sizeo-->There is quite a lot of data and I’d like to assign them to an array,therefore, I can use them through index.<br />In order to simplified the problem, an example is presented as listed bellows:<br />(1)	In vertex shader, I define an uniform array ch with two elements. If its second element ch[1] assigned a value greater than 2 ,then the vertex is rendered green;otherwise the vertex red.<br />(2)	In .material file,we test a cube using the vertex shader above.<!--coloro:#006400--><span style="color:#006400"><!--/coloro-->The problem is: what’s format should I choose to assign values to the array ch?<!--colorc--></span><!--/colorc--><br /><br />  <!--coloro:#9932CC--><span style="color:#9932CC"><!--/coloro--> &#9675;1  param_named ch  float 5 8 <br />   &#9675;2  param_named ch[2] float 5 8 <br />  &#9675;3   param_named []ch  float 5 8<br />  &#9675;4   param_named ch[]  float 5 8<br />  &#9675;5   param_named ch[0]  float 5<br />            param_named ch[1] float 8<!--colorc--></span><!--/colorc--><br /><br />I’ve  try them all, but none is correct.  Expecting anyone who could provide some advice.  Thank you~<br /><br /><b>Vertex Shader[V_green_array.cg]</b><br /><div class='codetop'>CODE</div><div class='codemain' style='height:200px;white-space:pre;overflow:auto'>void green_array(float3 position:POSITION,<br /><br />                   out float4 oPosition:POSITION,<br />                   out float4 oColor:COLOR,<br /><br />	  uniform float4x4 worldViewProj,<br />	  <!--coloro:#8B0000--><span style="color:#8B0000"><!--/coloro--><b>uniform float ch</b>[2]<!--colorc--></span><!--/colorc-->&#65289;<br />{<br />       <br />       oPosition=mul(worldViewProj,float4(position,1));<br />       float4 color;<br />       if(ch[1]&gt;2)<br />         color=float4(0,1,0,1);<br />       else<br />         color=float4(1,0,0,1);<br />        <br /> oColor=color;<br />          <br />}<br /></div><br /><br /><b>Material File[test.material]</b><br /><div class='codetop'>CODE</div><div class='codemain' style='height:200px;white-space:pre;overflow:auto'>vertex_program ver_green_array cg<br />{<br />	source v_green_array.cg<br />	entry_point green_array<br />	profiles vs_3_0<br /><br />}<br />material lifangti_mesh_material<br />{<br />	technique<br />	{<br />		pass<br />		{<br />			vertex_program_ref ver_green_array<br />			{			<br />			 param_named_auto worldViewProj worldviewproj_matrix <br />			<b><!--coloro:#9932CC--><span style="color:#9932CC"><!--/coloro-->param_named ch  float  5  8   <!--colorc--></span><!--/colorc--><!--coloro:#006400--><span style="color:#006400"><!--/coloro-->// The problem here:what’s format should be used?<!--colorc--></span><!--/colorc--></b>		 			 			<br />			 }<br />		}<br /><br />	}<br />}<br /></div><!--sizec--></span><!--/sizec--><!--fontc--></span><!--/fontc-->]]></description>
		<pubDate>Sun, 18 Oct 2009 21:14:00 -0700</pubDate>
		<guid>http://developer.nvidia.com/forums/index.php?showtopic=3748</guid>
	</item>
	<item>
		<title>Setting uniform parameters</title>
		<link>http://developer.nvidia.com/forums/index.php?showtopic=3747</link>
		<description><![CDATA[I'm trying to use Cg to get my HLSL shaders working in my OpenGL renderer and I haven't experienced any problems except that I can't find any other way to set uniform parameters except explicitly by name. This is a problem for me because my convention requires that I set data directly to constant registers. Doing such a thing in Direct3D is trivial (SetVertexShaderConstantF(), etc.), but OpenGL and Cg seem to have me beating my head on walls.<br /><br />So, my question is, in Cg, can I set parameters by register index?]]></description>
		<pubDate>Sun, 18 Oct 2009 18:59:55 -0700</pubDate>
		<guid>http://developer.nvidia.com/forums/index.php?showtopic=3747</guid>
	</item>
	<item>
		<title>Ageia PPU (yes, im too asking for it)</title>
		<link>http://developer.nvidia.com/forums/index.php?showtopic=3746</link>
		<description><![CDATA[Hello nvidia groub/humans. <br /><br />I can play/use examples of physx on 8.04.25 PhysX Drivers, but when i install any newer versions (8.09.04.. etc) of it i cant play anything!<br />First it asking "nvcuda.dll". Im tried almost all version of CUDA tool kit and they comes with that *.dll but... WHY CANT LOAD LIBRARY FROM IT?<br /><br /><br />That "test or developers" program which comes with newest CUDA tool kit has same problem. Why? <br />Im tried copy ( cuda.dll renaimed to nvcuda.dll whit oldest CUDA drivers and orginal new nvcuda.dll ) to system/WOW64 (?) -folder, but then its says<br /><br />"Entry Point Not Found"<br /><br />*The procedure entry point cuCtxDestroy could not be located in the dynamic link library nvcuda.dll."<br /><br />Bullsh*t!  <img src="http://developer.nvidia.com/forums/style_emoticons/default/dry.gif" style="vertical-align:middle" emoid="&lt;_&lt;" border="0" alt="dry.gif" /><br /><br />Help me please?  <img src="http://developer.nvidia.com/forums/style_emoticons/default/sad.gif" style="vertical-align:middle" emoid=":(" border="0" alt="sad.gif" /> What im doing wrong? Need a buy nvidia geforce for CUDA which needed be open freeware?]]></description>
		<pubDate>Sun, 18 Oct 2009 06:24:10 -0700</pubDate>
		<guid>http://developer.nvidia.com/forums/index.php?showtopic=3746</guid>
	</item>
	<item>
		<title><![CDATA[CG & Fragment Parameters... AARRRGH!]]></title>
		<link>http://developer.nvidia.com/forums/index.php?showtopic=3745</link>
		<description><![CDATA[Hi apologies for the cross-post from another forum but this problem is really doing my nut...  literally cannot continue work until this issue is resolved. Here goes...<br /><br />I'm no CG expert but have used OGl for some years now and have done some fairly complex OGL/CG apps in recent months with no major issues which only serves to add even more mystery to my problem, which is thus:<br /><br />I've been struggling for all of yesterday trying to get round this simple yet show-stopping situation: I can't seem to access the 4th texture stages' (and beyond) tex coords (stuffed with values passed from the vert shader). I implemented a fully interactive mesh deformation app using texture splatting with shaders, no problem. Recently I started researching normal mapping and got a rough prototype loading the meshes exported from the first app, no problem. I then start investigating tri-planar texturing and this is where my problem lies. Even though my laptop GPU supports &gt;8 (16 I think?) texture stages, I can't even get the tri-planar app off the ground rendering some test data  <img src="http://developer.nvidia.com/forums/style_emoticons/default/sad.gif" style="vertical-align:middle" emoid=":(" border="0" alt="sad.gif" /> <br /><br />Here's the fragment shader, stripped down to the bare essentials:<br /><br /><div class='codetop'>CODE</div><div class='codemain' style='height:200px;white-space:pre;overflow:auto'>/*************************************<br />*	<br />*	    Fragment Shader<br />*	    ---------------<br />*<br />**************************************/<br /><br />void main(in float4	       iColour			: COLOR,<br />	  in float3 	       iVertex			: TEXCOORD0, // Fine<br />	  in float3 	       iNormal			: TEXCOORD1, // Fine<br />	  in float2            iXZCoords                : TEXCOORD2, // Fine<br />          in float2            iXYCoords                : TEXCOORD3, // No!<br />          in float2            iZYCoords                : TEXCOORD4, // No!<br /><br />	  in uniform sampler2D iTexXZ			: TEX0,<br /><br />	  out float4	       oColour			: COLOR)<br />{	<br />        float2 xzPlaneCoords = iXZCoords;<br />	float2 xyPlaneCoords = iXYCoords;<br />	float2 zyPlaneCoords = iZYCoords;<br /><br />        // Nothing... or with the other 2 coords<br />	float4 diffuse = tex2D(iTexXZ,  xyPlaneCoords);<br /><br />	oColour =  diffuse;<br /><br />}<br /></div><br /><br />Now, I've used up to the 4th texture coords stage before in another app and had no problem. I've tried loading in the other 2 textures into tex stage 1 & 2 but still nothing. I've even tried compounding the vert, norm and tex coords into the 1st 3 tex coordinates (stuffed in float4s) but still nothing. I've tried putting the tex coords in TEXCOORD0, TEXCOORD1 & TEXCOORD2 respectively and put the vert and norm after but still nothing! If I leave the normal and vert out it's fine, but as soon as I start using the latter texture stage coords or compounding the vectors and coords in 4 float long tex coords it goes * up.<br /><br />My card supports at least 8 texture stages (16 I think?) and I've used all 8 texture stages and up to the 4th (TEXCOORD3) stage's texture coords before and I'm at a complete loss as to why this is.<br /><br /><br />It's this and other issues (such as incorrect results unless I sample all the texture stages or whole system crash if I assign the output colour to certain values) that have really been bugging me. Is it something schoolboy that I;m missing? Dodgy OpenGL driver (mediocre laptop)?<br /><br />Here's another (successful) fragment shader I wrote a few months back that works absolutely fine, just so you get an idea of scope (i.e. that my laptop can actually do some stuff alright, certainly more complex & demanding than the above stipped down example). Notice some odd coding conventions (such as calculating the output colour by storing and altinating between the oColour and tempory result vector). I've had to do things like this to get it not to crash or display incorrect results... there has to be something big I've missed in my understanding, surely? or is the laptop not up to the task (opengl 1.5, HW fragment but vertex runs in SW)<br /><br /><div class='codetop'>CODE</div><div class='codemain' style='height:200px;white-space:pre;overflow:auto'>/*************************************<br />*	<br />*			Fragment Shader<br />*			---------------<br />*<br />**************************************/<br /><br />void main(in float4	       iFragColour	: COLOR0,	<br />	  in float2	       iTexCoord	: TEXCOORD0,<br />	  in float4	       iSplatWeight1to4	: TEXCOORD1,<br />	  in float4	       iSplatWeight5to7	: TEXCOORD2,<br />	  in float3	       iFragPosition	: TEXCOORD3,	// Frag position interpolated from vertex position<br /><br />	  in uniform sampler2D iTexBase		: TEX0,<br />	  in uniform sampler2D iTexGrass	: TEX1,<br />	  in uniform sampler2D iTexRock		: TEX2,<br />	  in uniform sampler2D iTexSand		: TEX3,<br />	  in uniform sampler2D iTexWater	: TEX4,<br />	  in uniform sampler2D iTexDetail1	: TEX5,		// Detail textures 1-4 (ground, grass, rock, sand)<br />	  in uniform sampler2D iTexDetail2	: TEX6,		// Detail textures 5-8 (water)<br /><br />	  in uniform bool      useSplat,			// Toggle hi res textures on/off<br />	  in uniform bool      useDetail,			// Toggle hi res textures on/off<br />	  in uniform bool      useColour,			// Toggle vertex colours on/off<br />	  in uniform bool      useFog,				// Toggle fog on/off<br /><br />	  in uniform float     maxDist,				// Max distance at which the detail splats will be blended over the low re splats<br />	  in uniform float     medDist,				// Distance at which the blending o hi res splats rapidly falls off for smooth transition<br /><br />	  in uniform float4    iFogColour	: state.fog.color,<br />	  in uniform float4    iFogParams	: state.fog.params,<br /><br />	  out float4		oColour		: COLOR)<br />{<br />	float4 base;<br />	float4 grass;<br />	float4 rock; <br />	float4 sand; <br />	float4 water; <br />	float4 detail1;<br />	float4 detail2;<br /><br />	float4 finalBase;<br />	float4 finalGrass;<br />	float4 finalRock;<br />	float4 finalSand;<br />	float4 finalWater;<br /><br />	base  = tex2D(iTexBase,  iTexCoord);<br />	grass = tex2D(iTexGrass, iTexCoord);<br />	rock  = tex2D(iTexRock,  iTexCoord);<br />	sand  = tex2D(iTexSand,  iTexCoord);<br />	water = tex2D(iTexWater, iTexCoord);<br />	detail1  = tex2D(iTexDetail1,  iTexCoord);<br />	detail2 = tex2D(iTexDetail2, iTexCoord);<br /><br />	// Cam is at origin as vert/frag pos is in projected world space<br />	float3 CamPos = float3(0.0,0.0,0.0);<br /><br />	// Cut-off distance for the high res detail texture<br />	//float maxDist = 50.0;<br /><br />	// Calculate the frag's distance from the viewer<br />	float fragDist = distance(CamPos, iFragPosition);<br /><br />	// Are we close enough to use the hi-res texture?<br />	if(true == useDetail && fragDist &lt; maxDist)<br />	{<br />		// Amount of hi-res detail texture blended onto grass texture<br />		float hiresAmount;<br /><br />		// Max distance = 15. Therefore, any frags within 0 to 15 units is hi-res. However, in order to stop the<br />		// hi-res texture "popping" into place, we will blend the hi res amount over the last 2 thirds of the distance<br />		// range. The further away the frag, the less hi res detail blended in. This gives us a smooth transition<br />		// as we zoom in and out, making the hi-res "popping" non-existent<br />		if(fragDist &gt; medDist)<br />		{<br />			hiresAmount = 0.35 * (((maxDist - fragDist) / maxDist) + 0.5);<br />		}<br />		else<br />		{<br />			hiresAmount = 0.35;<br />		}<br /><br />		// Yes. Add the high-detail texture in the 2nd texture unit to the base texture from the appropriate colour channel<br />		finalBase  = ((1 - hiresAmount) * base) + (hiresAmount * detail1.r);<br />		finalGrass = ((1 - hiresAmount) * grass) + (hiresAmount * detail1.g);<br />		finalRock = ((1 - hiresAmount) * rock) + (hiresAmount * detail1.<img src="http://developer.nvidia.com/forums/style_emoticons/default/cool.gif" style="vertical-align:middle" emoid="B)" border="0" alt="cool.gif" />;<br />		finalSand = ((1 - hiresAmount) * sand) + (hiresAmount * detail1.a);<br />		finalWater = ((1 - hiresAmount) * water) + (hiresAmount * detail2.r);<br />	}<br />	else<br />	{<br />		// No. Only use the lo res textures<br />		finalBase = base;<br />		finalGrass = grass;<br />		finalRock = rock;<br />		finalSand = sand;  <br />		finalWater = water;<br />	}<br /><br />	float4 result = lerp(finalBase,   finalGrass, iSplatWeight1to4.x);   <br />	oColour       = lerp(result, finalRock,  iSplatWeight1to4.y);   <br />	result        = lerp(oColour,finalSand,  iSplatWeight1to4.z);   <br />	oColour       = lerp(result, finalWater, iSplatWeight1to4.w);  <br /><br />	if(true == useSplat)<br />	{<br />		if(true == useColour)<br />		{<br />			oColour *= iFragColour;<br />		}<br />	}<br />	else<br />	{<br />		oColour = iFragColour;<br />	}<br /><br />	// Linear fog equation<br />	if(true == useFog)<br />	{<br />		// iFogParams.y = fog start<br />		// iFogParams.z = fog end<br />		float fogFactor = (iFogParams.z - iFragPosition.z) / (iFogParams.z - iFogParams.y);<br /><br />		fogFactor = clamp(fogFactor, 0.0, 1.0);<br /><br />		float4 texColour   = oColour;<br /><br />		// Blend fog in with coloured & lit splat fragment<br />		oColour = lerp(iFogColour, texColour, fogFactor);<br />	}<br />}<br /><br /></div><br /><br />Also notice that in the above shader I've used TEXCOORD3, no probs.<br /><br />The following code crashes my system 100%... have to take the battery out the laptop to get it to boot up again:<br /><br /><div class='codetop'>CODE</div><div class='codemain' style='height:200px;white-space:pre;overflow:auto'>/ Vert shader<br />cgGLEnableProfile(CG_PROFILE_ARBVP1);<br />cgGLBindProgram(vertProgram);<br /><br />// Frag shader<br />cgGLEnableProfile(CG_PROFILE_ARBFP1);<br />cgGLBindProgram(fragProgram);<br /><br />m3dMesh.Draw();<br /><br /><br />// Disable shaders<br />cgGLDisableProfile(CG_PROFILE_ARBFP1);<br />cgGLDisableProfile(CG_PROFILE_ARBVP1);<br /></div><br /><div class='codetop'>CODE</div><div class='codemain' style='height:200px;white-space:pre;overflow:auto'>bool SGLMesh3d::Draw()<br />{<br />	if(!meshVertData || !meshIndexData) return false;<br /><br />	// Filled rendering<br />	glEnableVertexAttribArrayARB( 0 );<br />	glEnableVertexAttribArrayARB( 1 );<br />	glEnableVertexAttribArrayARB( 2 );<br />	glEnableVertexAttribArrayARB( 3 );<br />	glEnableVertexAttribArrayARB( 4 );<br />	glEnableVertexAttribArrayARB( 5 );<br />	glEnableVertexAttribArrayARB( 6 );<br />	glEnableVertexAttribArrayARB( 7 );<br /><br />		glVertexAttribPointerARB(0, 3, GL_FLOAT, GL_FALSE, sizeof(SMeshVertex), &meshVertData-&gt;xyzPos.x );	// x,y,z<br />		glVertexAttribPointerARB(1, 3, GL_FLOAT, GL_FALSE, sizeof(SMeshVertex), &meshVertData-&gt;xyzTangent.x );	// tx,ty,tz<br />		glVertexAttribPointerARB(2, 3, GL_FLOAT, GL_FALSE, sizeof(SMeshVertex), &meshVertData-&gt;xyzBiNormal.x );	// bnx,bny,bnz<br />		glVertexAttribPointerARB(3, 3, GL_FLOAT, GL_FALSE, sizeof(SMeshVertex), &meshVertData-&gt;xyzNormal.x );	// nx,ny,nz<br />		glVertexAttribPointerARB(4, 4, GL_FLOAT, GL_FALSE, sizeof(SMeshVertex), &meshVertData-&gt;rgbaColour.r);	// r,g,b,a<br />		glVertexAttribPointerARB(5, 2, GL_FLOAT, GL_FALSE, sizeof(SMeshVertex), &meshVertData-&gt;texCoord0.u );	// texU0, texV0<br />		glVertexAttribPointerARB(6, 2, GL_FLOAT, GL_FALSE, sizeof(SMeshVertex), &meshVertData-&gt;texCoord1.u );	// texU1, texV1<br />		glVertexAttribPointerARB(7, 4, GL_FLOAT, GL_FALSE, sizeof(SMeshVertex), &meshVertData-&gt;customData );	// custom data<br /><br />		glDrawElements(GL_TRIANGLES, numIndices, GL_UNSIGNED_INT, meshIndexData);<br />		glPopMatrix();<br /><br />	glDisableVertexAttribArrayARB( 0 );<br />	glDisableVertexAttribArrayARB( 1 );<br />	glDisableVertexAttribArrayARB( 2 );<br />	glDisableVertexAttribArrayARB( 3 );<br />	glDisableVertexAttribArrayARB( 4 );<br />	glDisableVertexAttribArrayARB( 5 );<br />	glDisableVertexAttribArrayARB( 6 );<br />	glDisableVertexAttribArrayARB( 7 );<br /><br />	return true;<br />}<br /></div><br /><div class='codetop'>CODE</div><div class='codemain' style='height:200px;white-space:pre;overflow:auto'>/*************************************<br />*	<br />*			Vertex Shader<br />*			-------------<br />*<br />**************************************/<br /><br />void main(in float4  iPosition				: ATTR0,<br />		  in float4  iTangent			: ATTR1,<br />		  in float4  iBiNormal			: ATTR2,<br />		  in float4  iNormal			: ATTR3,<br />		  in float4  iColour			: ATTR4,<br />		  in float2  iTexCoord0			: ATTR5,<br />		  in float2  iTexCoord1			: ATTR6,<br />		  in float4  iCustomData		: ATTR7,<br /><br />		  out float4 oPosition			: POSITION,				<br />		  out float4 oColour			: COLOR,<br />		  out float4 t0				: TEXCOORD0,<br />		  out float4 t1				: TEXCOORD1,<br />		  out float4 t2				: TEXCOORD2,<br /><br />		  uniform float4x4 ModelViewProj	: state.matrix.mvp)		<br />{<br />	oPosition =  mul(ModelViewProj, iPosition);<br /><br />	t0 = float4(1,0,0,1);<br />	t1 = float4(0,1,0,1);<br />	t2 = float4(0,0,1,1);<br /><br />	oColour	= iColour;<br /><br />}<br /></div><br /><div class='codetop'>CODE</div><div class='codemain' style='height:200px;white-space:pre;overflow:auto'>/*************************************<br />*	<br />*			Fragment Shader<br />*			---------------<br />*<br />**************************************/<br /><br />void main(in float4		iColour		: COLOR,<br />		  in float4	t0		: TEXCOORD0,<br />		  in float4	t1		: TEXCOORD1,<br />		  in float4	t2		: TEXCOORD2,<br />		  out float4	oColour		: COLOR)<br />{	<br />	<br />	oColour = t0;<br /><br />}<br /></div><br /><br />The following code doesn't crash, but only shows a garbled coloured mesh when using TEXCOORD2:<br /><div class='codetop'>CODE</div><div class='codemain' style='height:200px;white-space:pre;overflow:auto'>// Load shaders, as before...<br /><br />...<br /><br />glBegin(GL_TRIANGLES);<br /><br />	glColor3f(1.0f,1.0f,1.0f);<br /><br />	for(UInt32 ctr = 0; ctr &lt; m3dMesh.numIndices; ctr++)<br />	{<br />	     glVertex3fv(&m3dMesh.meshVertData[m3dMesh.meshIndexData[ctr]].xyzPos.x);<br />	}<br /><br />glEnd();<br /><br />// Unload shaders, as before...<br /></div><br /><div class='codetop'>CODE</div><div class='codemain' style='height:200px;white-space:pre;overflow:auto'>/*************************************<br />*	<br />*			Vertex Shader<br />*			-------------<br />*<br />**************************************/<br /><br />void main(<br />		  in float4  iPosition				: POSITION,<br />		  in float4  iColour				: COLOR,<br /><br />		  out float4 oPosition				: POSITION,			// Translated and projected vertex position	<br />		  out float4 oColour				: COLOR,<br />		  out float4 t0				: TEXCOORD0,<br />		  out float4 t1			: TEXCOORD1,<br />		  out float4 t2				: TEXCOORD2,<br /><br />		  uniform float4x4 ModelViewProj	: state.matrix.mvp)			// ModelViewPorojection matrix<br />{<br />	// Translate and project the vertex<br />	oPosition =  mul(ModelViewProj, iPosition);<br /><br />	t0 = float4(1,0,0,1);<br />	t1 = float4(0,1,0,1);<br />	t2 = float4(0,0,1,1);<br /><br />	oColour	= iColour;<br />}<br /></div><br /><div class='codetop'>CODE</div><div class='codemain' style='height:200px;white-space:pre;overflow:auto'>/*************************************<br />*	<br />*			Fragment Shader<br />*			---------------<br />*<br />**************************************/<br /><br />void main(in float4		iColour		: COLOR,<br />		  in float4		t0			: TEXCOORD0,<br />		  in float4		t1			: TEXCOORD1,<br />		  in float4		t2			: TEXCOORD2,<br />		  out float4	oColour		: COLOR)<br />{	<br />	// output is white... should be blue?<br />	oColour = t2;<br /><br />}<br /></div><br />t0 and t1 show red ang green meshes as expected, t2 shows black with wierd artefacts. Does anyone have any ideas as to what is going on? Any help would be much appreciated.]]></description>
		<pubDate>Sun, 18 Oct 2009 04:37:37 -0700</pubDate>
		<guid>http://developer.nvidia.com/forums/index.php?showtopic=3745</guid>
	</item>
	<item>
		<title>PerfHUD 6.6: Strange GPU time values</title>
		<link>http://developer.nvidia.com/forums/index.php?showtopic=3744</link>
		<description><![CDATA[Hi all,<br />I have been faced with the following porblem.<br />PerfHUD shows very large time value for GPU.<br /><br /><a href='http://developer.nvidia.com/forums/index.php?act=attach&type=post&id=282'>http://developer.nvidia.com/forums/index.php?act=attach&type=post&id=282</a><br /><br />Bug occures on the following PC configuration.<br />OS: Windows Vista 32.<br />Video: GeForce 8600 GT.<br />Driver: 190.62<br />DX SDK: August 2009<br /><br />...but does not on:<br />OS: Windows Vista 64.<br />Video: GeForce GTX 260<br />Driver: 190.62<br />DX SDK: August 2009<br /><br />Samples work well on both computers.<br /><br />How to solve this problem?<br /><br />Thanks beforehand <img src="http://developer.nvidia.com/forums/style_emoticons/default/smile.gif" style="vertical-align:middle" emoid=":)" border="0" alt="smile.gif" />]]></description>
		<pubDate>Sun, 18 Oct 2009 01:28:49 -0700</pubDate>
		<guid>http://developer.nvidia.com/forums/index.php?showtopic=3744</guid>
	</item>
	<item>
		<title>3d modeling in physx..</title>
		<link>http://developer.nvidia.com/forums/index.php?showtopic=3743</link>
		<description><![CDATA[hi,i am new to nvidia physx....<br />i had done some basic physx trining programs....now i want to load a .3ds and md2 models and handle it using physx.......<br />how can i do????<br />thanx...... <br /><br />i have posted this question in wrong topic....in cq..sorry for my noobness...<br /><a href="http://developer.nvidia.com/forums/index.php?showtopic=3728" target="_blank">http://developer.nvidia.com/forums/index.php?showtopic=3728</a><br /><br />i have search in google and in this site also but cant find a good solution....pls give me some guidelines.......... <br />thanx.... <img src="http://developer.nvidia.com/forums/style_emoticons/default/smile.gif" style="vertical-align:middle" emoid=":)" border="0" alt="smile.gif" />]]></description>
		<pubDate>Fri, 16 Oct 2009 21:49:12 -0700</pubDate>
		<guid>http://developer.nvidia.com/forums/index.php?showtopic=3743</guid>
	</item>
	<item>
		<title>Cg Geometry shaders and Macbook Pro</title>
		<link>http://developer.nvidia.com/forums/index.php?showtopic=3742</link>
		<description><![CDATA[Hey everybody<br /><br />My name is Elvar Orn Unnthorsson and I am currently working on my final project which uses the Panda3d Game engine (panda3d.org) and Cg shaders.<br />I am having problems and I seriously need some info from you guys.<br /><br />Can I or can't I use the Cg Geometry shader with my Macbook Pro which has the GeForce 8600M GT card?]]></description>
		<pubDate>Fri, 16 Oct 2009 16:47:21 -0700</pubDate>
		<guid>http://developer.nvidia.com/forums/index.php?showtopic=3742</guid>
	</item>
	<item>
		<title>Switching Textures mid-Shader</title>
		<link>http://developer.nvidia.com/forums/index.php?showtopic=3741</link>
		<description><![CDATA[If I do these steps...<br /><br /><ul>* Bind Shader A (cgSetPassState)<br />** Set texture parameter to Texture A (cgGLSetTextureParameter)<br />*** Draw face A<br />** Set texture parameter to Texture B (cgGLSetTextureParameter)<br />*** Draw face B<br />* Unbind Shader A (cgResetPassState)</li></ul><br /><br />...both faces ( A & B ) end up with Texture B on them!<br /><br />I have tried every permutation I can think of to get this to work.  Setting up the texture parameters outside of the binding of the shaders does work, but I am trying to minimize state changes.  This seems like shader usage 101, but I am stumped as to why I can't get it to work.  Is this not a legal thing to do?<br /><br />I am a little reluctant to post source because I have tried so many variations I am not sure which one to post, but here is the simplest one that I think should work...<br /><br /><div class='codetop'>CODE</div><div class='codemain' style='height:200px;white-space:pre;overflow:auto'>// the shader is bound at this point via cgSetPassState<br />// CGparameter param = a sampler parameter supplied from elsewhere<br />// GLuint glTextureA & glTextureB are textures supplied from elsewhere<br />	cgGLSetTextureParameter( param, glTextureA );<br />	cgGLEnableTextureParameter( param );<br /> <br />	glBegin( GL_QUADS );<br />		glTexCoord2f( 0, 0 );<br />		glVertex3f( 0, 0, 0 );<br />		glTexCoord2f( 0, 1 );<br />		glVertex3f( 0, 1, 0 );<br />		glTexCoord2f( 1, 1 );<br />		glVertex3f( 1, 1, 0 );<br />		glTexCoord2f( 1, 0 );<br />		glVertex3f( 1, 0, 0 );<br />	glEnd();<br /> <br />	cgGLSetTextureParameter( param, glTextureB );<br />	cgGLEnableTextureParameter( param );<br /> <br />	glBegin( GL_QUADS );<br />		glTexCoord2f( 0, 0 );<br />		glVertex3f( 1, 0, 0 );<br />		glTexCoord2f( 0, 1 );<br />		glVertex3f( 1, 1, 0 );<br />		glTexCoord2f( 1, 1 );<br />		glVertex3f( 2, 1, 0 );<br />		glTexCoord2f( 1, 0 );<br />		glVertex3f( 2, 0, 0 );<br />	glEnd();</div><br /><br />Thanks in advance,<br />-Randy]]></description>
		<pubDate>Fri, 16 Oct 2009 09:32:26 -0700</pubDate>
		<guid>http://developer.nvidia.com/forums/index.php?showtopic=3741</guid>
	</item>
	<item>
		<title>MFC view doc and PerfHud</title>
		<link>http://developer.nvidia.com/forums/index.php?showtopic=3740</link>
		<description><![CDATA[Hello, <br /><br />I'm a newbie with perfHud.<br /><br />Is it possible to use PerfHud with an MFC view/doc application? <br /><br />I'm creating my D3d9 device during the OnInitialUpdate of a View.<br />This view is created only after a click on "New 3d scene" button. <br /><br />When I run the app with PerfHud it silently closes without showing a thing.<br /><br />Are there some logs, or error messages I could look for? Where are them?<br /><br />What happens when I run PerfHud??<br /><br />Thanks a lot for every reply, please ask me more if I was not clear in my explanation.<br /><br />Best regards,<br />Ema.]]></description>
		<pubDate>Fri, 16 Oct 2009 08:08:26 -0700</pubDate>
		<guid>http://developer.nvidia.com/forums/index.php?showtopic=3740</guid>
	</item>
	<item>
		<title>Trying to push two actors connected with spherical joints away from each other programmically</title>
		<link>http://developer.nvidia.com/forums/index.php?showtopic=3739</link>
		<description><![CDATA[Hi,<br /><br />I have 2 actors connected by a spherical joint.<br /><br />I need to push them apart or move them towards about this spherical joints.<br /><br />Of course I just realize updating the spherical joints descriptior doesn't work.<br /><br />I don't know how I could explain it better but if you need more information I would gladly provides.<br /><br />Thanks in advance! <img src="http://developer.nvidia.com/forums/style_emoticons/default/smile.gif" style="vertical-align:middle" emoid=":)" border="0" alt="smile.gif" />]]></description>
		<pubDate>Fri, 16 Oct 2009 01:59:19 -0700</pubDate>
		<guid>http://developer.nvidia.com/forums/index.php?showtopic=3739</guid>
	</item>
	<item>
		<title>overlapAABBTetrahedra seems to give false positive</title>
		<link>http://developer.nvidia.com/forums/index.php?showtopic=3738</link>
		<description><![CDATA[<br />I know that there is only ONE tetrahedra in my softbody with the specific NxBounds that I am testing.  Even if there is more, I tried to scale the bounds several times, and I keep receiving three tetrahedras in the numberTetra variable. Finally I removed all tetrahedras from the scene, except the one that really contains the bounds, and tested again the same point: got 1 tetrahedra on the numberTetra. <br /><br /><div class='codetop'>CODE</div><div class='codemain' style='height:200px;white-space:pre;overflow:auto'><br /><br />NxVec3 center;<br />NxVec3 extents;<br />NxBounds3 bounds;<br /><br />center.x = point.x;<br />center.y = point.y;<br />center.z = point.z;<br />extents.x = 1;<br />extents.y = 1;<br />extents.z = 1;<br /><br />bounds.setCenterExtents(center, extents);<br /><br />NxU32 numberTetras;<br />NxU32 const**indices;<br />indices = (NxU32 const**) new NxU32[maxIndices];<br /><br />pSoftBody-&gt;-&gt;overlapAABBTetrahedra(bounds, numberTetras, *indices);<br /><br />//Overlapping is  3, should be 1<br /><br /></div><br /><br />Scaling the NxBounds with a very small factor:<br /><br /><div class='codetop'>CODE</div><div class='codemain' style='height:200px;white-space:pre;overflow:auto'><br /><br />bounds.setCenterExtents(center, extents);<br />bounds.scale(0.000001);<br /><br />NxU32 numberTetras;<br />NxU32 const**indices;<br />indices = (NxU32 const**) new NxU32[maxIndices];<br /><br />pSoftBody-&gt;-&gt;overlapAABBTetrahedra(bounds, numberTetras, *indices);<br /><br />//Overlapping still 3, should be 1 or 0!<br /><br /></div><br /><br />I saw a similar topic: <a href="http://developer.nvidia.com/forums/index.php?showtopic=3570&pid=10418&mode=threaded&start=#entry10418" target="_blank">http://developer.nvidia.com/forums/index.p...art=#entry10418</a><br /><br />Am I doing something wrong? Is this a bug?<br />TYA.<br />]]></description>
		<pubDate>Thu, 15 Oct 2009 16:21:03 -0700</pubDate>
		<guid>http://developer.nvidia.com/forums/index.php?showtopic=3738</guid>
	</item>
	<item>
		<title>Switch displays between clone and dualview with NVAPI?</title>
		<link>http://developer.nvidia.com/forums/index.php?showtopic=3737</link>
		<description><![CDATA[Dear NVIDIA Developers:<br /><br />Is it possible to use NVAPI to switch the two displays between clone and dualview? the operating system I use is Vista.<br /><br />any suggestion or example codes are highly appreciated.<br /><br />-Deliang]]></description>
		<pubDate>Thu, 15 Oct 2009 11:56:50 -0700</pubDate>
		<guid>http://developer.nvidia.com/forums/index.php?showtopic=3737</guid>
	</item>
	<item>
		<title>Simplest Way to make a force field?</title>
		<link>http://developer.nvidia.com/forums/index.php?showtopic=3736</link>
		<description><![CDATA[I tried to make my own force sphere, with simple kernel but it doesn't work. I'm using linux... so do i have to set any compile flags???<br /> but i get all confused about the coordinates of the force, i.e. spherical etc. it makes me wonder how do i make a sphere that forces stuff outwards???<br /><br />could someone post the simplest example possible and explain to me why the forces are 3d vectors and not floats????]]></description>
		<pubDate>Thu, 15 Oct 2009 10:49:41 -0700</pubDate>
		<guid>http://developer.nvidia.com/forums/index.php?showtopic=3736</guid>
	</item>
	<item>
		<title>Triangle Meshes</title>
		<link>http://developer.nvidia.com/forums/index.php?showtopic=3735</link>
		<description><![CDATA[Hi all,<br />I am a newbie to SDK Physics in general, but I decided to learn the NVIDIA's SDK because I think that is the best choice.<br />I have built a basic mesh, using my favorite modeling application. This mesh will serve as a terrain for characters of my game.<br />I don't want to make use of a height field.<br />I suppose that I need to build a triangleMesh, but I don't know how to load the mesh data from the modeling package to the PhysX SDK.<br />I know that first I have to cook the mesh data but the question remains: How I will feed the SDK with the mesh data from the modeling application?<br /><br />Thanks in advance.]]></description>
		<pubDate>Thu, 15 Oct 2009 09:00:33 -0700</pubDate>
		<guid>http://developer.nvidia.com/forums/index.php?showtopic=3735</guid>
	</item>
	<item>
		<title>glDrawElements and CG</title>
		<link>http://developer.nvidia.com/forums/index.php?showtopic=3734</link>
		<description><![CDATA[When I use glVertex3f with CG there aren't any problem, but when I draw with glDrawElements, it's like I don't use CG. <br />For example<br />If a draw a triangle with glVertex3f and then I change the color with CG, it works<br />If a draw a triangle with glDrawElements, then I change the color, the color don't change.<br /><br />Are there problems with GC using glDrawElements ?]]></description>
		<pubDate>Thu, 15 Oct 2009 01:38:56 -0700</pubDate>
		<guid>http://developer.nvidia.com/forums/index.php?showtopic=3734</guid>
	</item>
	<item>
		<title>Help Wanted</title>
		<link>http://developer.nvidia.com/forums/index.php?showtopic=3733</link>
		<description>I am in need of some assistance in developing some custom vehicle handling and would be able to pay someone as an independent contractor for their time.  If interested please respond and we can discuss further.  Thanks</description>
		<pubDate>Wed, 14 Oct 2009 17:29:45 -0700</pubDate>
		<guid>http://developer.nvidia.com/forums/index.php?showtopic=3733</guid>
	</item>
	<item>
		<title>Pool physX</title>
		<link>http://developer.nvidia.com/forums/index.php?showtopic=3731</link>
		<description><![CDATA[Hi everybody,<br /><br />I´m using Director´s physX Xtra to develope a Pool Game. I don´t know how Ageia calculates the friction/restitution factors.<br />For example:<br /><br />I know this values: <br />Coeff. of restitution between balls = 0.98<br />Coeff. of restitution between ball and felt = 0.065<br /><br />But I have to set a ball restitution value and a felt restitution value, not a ball/ball restituion value or a ball/felt restitution value...<br /><br />Taking the coefficient of restitution, how do I calculate the restitution value of each rigid body?<br /><br />Thanks.]]></description>
		<pubDate>Wed, 14 Oct 2009 03:52:25 -0700</pubDate>
		<guid>http://developer.nvidia.com/forums/index.php?showtopic=3731</guid>
	</item>
	<item>
		<title>PhysX-2.8.1 Debian amd64</title>
		<link>http://developer.nvidia.com/forums/index.php?showtopic=3730</link>
		<description><![CDATA[Hi, <br /><br />I'm unable to install PhysX 2.8.1 on my debian system, which is a 64-system. <br /><br />How do I install PhysX on my amd-64 bit system? <br /><br />Am I required to install Debian Lenny i386 to make PhysX 2.8.1 to work? <br />I hope not...<br /><br />Ole <br /><br /><br /><br /><br /><br />]]></description>
		<pubDate>Wed, 14 Oct 2009 01:17:49 -0700</pubDate>
		<guid>http://developer.nvidia.com/forums/index.php?showtopic=3730</guid>
	</item>
	<item>
		<title>Get program source from effect?</title>
		<link>http://developer.nvidia.com/forums/index.php?showtopic=3729</link>
		<description>Given an effect, I can get the CGprograms for a given pass.  Is there anyway to get the source code from these CGprogram objects?  I can call cgGetProgramString( x, CG_PROGRAM_SOURCE ), but this just returns the original contents of the effect file.  I can swap in CG_COMPILED_PROGRAM, and get the disassembled code to the actual program.  Is there a way I can get the cg code that was compiled into this disassembly?</description>
		<pubDate>Tue, 13 Oct 2009 19:18:41 -0700</pubDate>
		<guid>http://developer.nvidia.com/forums/index.php?showtopic=3729</guid>
	</item>
	<item>
		<title>.3ds modeling</title>
		<link>http://developer.nvidia.com/forums/index.php?showtopic=3728</link>
		<description><![CDATA[hi,i am new to nvidia physx....<br />i had done some basic physx trining programs....now i want to load a .3ds and md2 models and handle it using physx.......<br />how can i do????<br />thanx......]]></description>
		<pubDate>Tue, 13 Oct 2009 18:52:53 -0700</pubDate>
		<guid>http://developer.nvidia.com/forums/index.php?showtopic=3728</guid>
	</item>
	<item>
		<title>Need more documentation</title>
		<link>http://developer.nvidia.com/forums/index.php?showtopic=3727</link>
		<description><![CDATA[Is there any more documentation other than the training programs and samples anywhere? Tutortials? Examples? Help?  The training programs and samples are great but I need more help, I can't just come up with stuff out of nowhere.  Not everything is explained.  How do I know what everything does in NxTireFunctionDesc, how do I find out how to add analog and steering wheel support?  How do I increase the velocity of a vehicle?  I've got a million questions like these.  If I become a licensed develeper will they answer these questions for me?  Is there a cost to that?  I appreciate any help.]]></description>
		<pubDate>Tue, 13 Oct 2009 16:53:34 -0700</pubDate>
		<guid>http://developer.nvidia.com/forums/index.php?showtopic=3727</guid>
	</item>
	<item>
		<title>Depht map from object (obj)</title>
		<link>http://developer.nvidia.com/forums/index.php?showtopic=3726</link>
		<description><![CDATA[I have to do a depth map from a object (obj). But, when I try to use CG for example to change the color, the object&#96;color didn't change.<br />Somebody knows how to do a depth map from and object or someone can tell me what I doing wrong ?<br /><br />This is the code to viewer de object<br /><br /><!--c1--><div class='codetop'>CODE</div><div class='codemain'><!--ec1-->void DrawModelUsingProgrammablePipeline&#40;&#41;<br />{<br />&nbsp;&nbsp;&nbsp;&nbsp;const ModelOBJ&#58;&#58;Mesh *pMesh = 0;<br />&nbsp;&nbsp;&nbsp;&nbsp;const ModelOBJ&#58;&#58;Material *pMaterial = 0;<br />&nbsp;&nbsp;&nbsp;&nbsp;const ModelOBJ&#58;&#58;Vertex *pVertices = 0;<br />&nbsp;&nbsp;&nbsp;&nbsp;ModelTextures&#58;&#58;const_iterator iter;<br />&nbsp;&nbsp;&nbsp;&nbsp;GLuint texture = 0;<br /><br />&nbsp;&nbsp;&nbsp;&nbsp;glEnable&#40;GL_BLEND&#41;;<br />&nbsp;&nbsp;&nbsp;&nbsp;glBlendFunc&#40;GL_SRC_ALPHA, GL_ONE_MINUS_SRC_ALPHA&#41;;<br /><br />&nbsp;&nbsp;&nbsp;&nbsp;for &#40;int i = 0; i &#60; g_model.getNumberOfMeshes&#40;&#41;; ++i&#41;<br />&nbsp;&nbsp;&nbsp;&nbsp;{<br />&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;pMesh = &g_model.getMesh&#40;i&#41;;<br />&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;pMaterial = pMesh-&#62;pMaterial;<br />&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;pVertices = g_model.getVertexBuffer&#40;&#41;;<br /><br />&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;glMaterialfv&#40;GL_FRONT_AND_BACK, GL_AMBIENT, pMaterial-&#62;ambient&#41;;<br />&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;glMaterialfv&#40;GL_FRONT_AND_BACK, GL_DIFFUSE, pMaterial-&#62;diffuse&#41;;<br />&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;glMaterialfv&#40;GL_FRONT_AND_BACK, GL_SPECULAR, pMaterial-&#62;specular&#41;;<br />&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;glMaterialf&#40;GL_FRONT_AND_BACK, GL_SHININESS, pMaterial-&#62;shininess * 128.0f&#41;;<br /><br />&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;if &#40;pMaterial-&#62;bumpMapFilename.empty&#40;&#41;&#41;<br />&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;{<br />&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;// Per fragment Blinn-Phong code path.<br /><br />&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;glUseProgram&#40;g_blinnPhongShader&#41;;<br />&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;glAttachShader&#40;g_blinnPhongShader,cgVertexProfile&#41;;<br /><br />&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; Bind the color map texture.<br /><br />&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;texture = g_nullTexture;<br /><br />&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;if &#40;g_enableTextures&#41;<br />&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;{<br />&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;iter = g_modelTextures.find&#40;pMaterial-&#62;colorMapFilename&#41;;<br /><br />&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;if &#40;iter != g_modelTextures.end&#40;&#41;&#41;<br />&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;texture = iter-&#62;second;<br />&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;}<br /><br />&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;glActiveTexture&#40;GL_TEXTURE0&#41;;<br />&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;glEnable&#40;GL_TEXTURE_2D&#41;;<br />&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;glBindTexture&#40;GL_TEXTURE_2D, texture&#41;;<br /><br />&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;// Update shader parameters.<br /><br />&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;glUniform1i&#40;glGetUniformLocation&#40;<br />&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;g_blinnPhongShader, &#34;colorMap&#34;&#41;, 0&#41;;<br />&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;glUniform1f&#40;glGetUniformLocation&#40;<br />&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;g_blinnPhongShader, &#34;materialAlpha&#34;&#41;, pMaterial-&#62;alpha&#41;;<br />&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;}<br />&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;else<br />&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;{<br />&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;// Normal mapping code path.<br /><br />&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;glUseProgram&#40;g_normalMappingShader&#41;;<br /><br />&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;// Bind the normal map texture.<br /><br />&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;iter = g_modelTextures.find&#40;pMaterial-&#62;bumpMapFilename&#41;;<br /><br />&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;if &#40;iter != g_modelTextures.end&#40;&#41;&#41;<br />&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;{<br />&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;glActiveTexture&#40;GL_TEXTURE1&#41;;<br />&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;glEnable&#40;GL_TEXTURE_2D&#41;;<br />&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;glBindTexture&#40;GL_TEXTURE_2D, iter-&#62;second&#41;;<br />&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;}<br /><br />&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;// Bind the color map texture.<br /><br />&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;texture = g_nullTexture;<br /><br />&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;if &#40;g_enableTextures&#41;<br />&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;{<br />&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;iter = g_modelTextures.find&#40;pMaterial-&#62;colorMapFilename&#41;;<br /><br />&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;if &#40;iter != g_modelTextures.end&#40;&#41;&#41;<br />&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;texture = iter-&#62;second;<br />&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;}<br /><br />&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;glActiveTexture&#40;GL_TEXTURE0&#41;;<br />&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;glEnable&#40;GL_TEXTURE_2D&#41;;<br />&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;glBindTexture&#40;GL_TEXTURE_2D, texture&#41;;<br /><br />&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;// Update shader parameters.<br /><br />&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;glUniform1i&#40;glGetUniformLocation&#40;<br />&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;g_normalMappingShader, &#34;colorMap&#34;&#41;, 0&#41;;<br />&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;glUniform1i&#40;glGetUniformLocation&#40;<br />&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;g_normalMappingShader, &#34;normalMap&#34;&#41;, 1&#41;;<br />&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;glUniform1f&#40;glGetUniformLocation&#40;<br />&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;g_normalMappingShader, &#34;materialAlpha&#34;&#41;, pMaterial-&#62;alpha&#41;;<br />&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;}&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;<br /><br />&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;// Render mesh.<br /><br />&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;if &#40;g_model.hasPositions&#40;&#41;&#41;<br />&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;{<br />&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;glEnableClientState&#40;GL_VERTEX_ARRAY&#41;;<br />&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;glVertexPointer&#40;3, GL_FLOAT, g_model.getVertexSize&#40;&#41;,<br />&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;g_model.getVertexBuffer&#40;&#41;-&#62;position&#41;;<br />&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;}<br /><br />&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;if &#40;g_model.hasTextureCoords&#40;&#41;&#41;<br />&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;{<br />&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;glClientActiveTexture&#40;GL_TEXTURE0&#41;;<br />&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;glEnableClientState&#40;GL_TEXTURE_COORD_ARRAY&#41;;<br />&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;glTexCoordPointer&#40;2, GL_FLOAT, g_model.getVertexSize&#40;&#41;,<br />&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;g_model.getVertexBuffer&#40;&#41;-&#62;texCoord&#41;;<br />&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;}<br /><br />&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;if &#40;g_model.hasNormals&#40;&#41;&#41;<br />&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;{<br />&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;glEnableClientState&#40;GL_NORMAL_ARRAY&#41;;<br />&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;glNormalPointer&#40;GL_FLOAT, g_model.getVertexSize&#40;&#41;,<br />&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; g_model.getVertexBuffer&#40;&#41;-&#62;normal&#41;;<br />&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;}<br /><br />&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;if &#40;g_model.hasTangents&#40;&#41;&#41;<br />&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;{<br />&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;glClientActiveTexture&#40;GL_TEXTURE1&#41;;<br />&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;glEnableClientState&#40;GL_TEXTURE_COORD_ARRAY&#41;;<br />&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;glTexCoordPointer&#40;4, GL_FLOAT, g_model.getVertexSize&#40;&#41;,<br />&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;g_model.getVertexBuffer&#40;&#41;-&#62;tangent&#41;;<br />&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;}<br /><br />&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;<br />&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;glDrawElements&#40;GL_TRIANGLES, pMesh-&#62;triangleCount * 3, GL_UNSIGNED_INT,<br />&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;g_model.getIndexBuffer&#40;&#41; + pMesh-&#62;startIndex&#41;;<br /><br />&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;if &#40;g_model.hasTangents&#40;&#41;&#41;<br />&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;{<br />&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;glClientActiveTexture&#40;GL_TEXTURE1&#41;;<br />&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;glDisableClientState&#40;GL_TEXTURE_COORD_ARRAY&#41;;<br />&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;}<br /><br />&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;if &#40;g_model.hasNormals&#40;&#41;&#41;<br />&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;glDisableClientState&#40;GL_NORMAL_ARRAY&#41;;<br /><br />&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;if &#40;g_model.hasTextureCoords&#40;&#41;&#41;<br />&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;{<br />&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;glClientActiveTexture&#40;GL_TEXTURE0&#41;;<br />&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;glDisableClientState&#40;GL_TEXTURE_COORD_ARRAY&#41;;<br />&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;}<br /><br />&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;if &#40;g_model.hasPositions&#40;&#41;&#41;<br />&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;glDisableClientState&#40;GL_VERTEX_ARRAY&#41;;<br />&nbsp;&nbsp;&nbsp;&nbsp;}<br /><br />&nbsp;&nbsp;&nbsp;&nbsp;glBindTexture&#40;GL_TEXTURE_2D, 0&#41;;<br />&nbsp;&nbsp;&nbsp;&nbsp;glUseProgram&#40;0&#41;;<br />&nbsp;&nbsp;&nbsp;&nbsp;glDisable&#40;GL_BLEND&#41;;<br />}<!--c2--></div><!--ec2--><br /><br /><br />I'm not good with english. Sorry]]></description>
		<pubDate>Tue, 13 Oct 2009 12:02:07 -0700</pubDate>
		<guid>http://developer.nvidia.com/forums/index.php?showtopic=3726</guid>
	</item>
	<item>
		<title>moving PhysX entirely on GPU (no CPU usage)</title>
		<link>http://developer.nvidia.com/forums/index.php?showtopic=3725</link>
		<description><![CDATA[Hi everybody,<br /><br />I think this is a question mostly for PhysX developers.<br /><br /><br />Is there a chance to get such feature from PhysX as setting and getting PhysX data from the client GPU data area to the PhysX GPU data, eliminating CPU participating?<br /><br />At least such data as positions, orientations, velocities, forces, torques, masses and moments of inertia.<br />I think it is not too hard to realize on a buffer object basis, for example, under OpenGL.<br />Of course CPU stays be involved for a while, but at minimal level.<br /><br />I'm interested to have such feature and, I hope, I'm not alone.<br /><br /><br />Thanks]]></description>
		<pubDate>Tue, 13 Oct 2009 09:21:59 -0700</pubDate>
		<guid>http://developer.nvidia.com/forums/index.php?showtopic=3725</guid>
	</item>
	<item>
		<title>Question about Cloth Stretch</title>
		<link>http://developer.nvidia.com/forums/index.php?showtopic=3724</link>
		<description><![CDATA[Hi all<br />      I have a problem about cloth stretch.As SDK doc says that stretching stiffness defines the strength of the constraint along a triangle edge that maintains the distance between the edge end points.I hope my cloth is very difficult to stretch so I set this value is one.Then I attach my cloth to some points,and when attach points move fast,I found that cloth also have been stretched.What shall I do? Thanks!!!]]></description>
		<pubDate>Tue, 13 Oct 2009 09:15:37 -0700</pubDate>
		<guid>http://developer.nvidia.com/forums/index.php?showtopic=3724</guid>
	</item>
	<item>
		<title>Understanding output of PerfHUD 6.6</title>
		<link>http://developer.nvidia.com/forums/index.php?showtopic=3723</link>
		<description><![CDATA[Hi all,<br />   I had done a GPU based ray caster and wanted to profile it using perfhud.I need to understand what the output shows to me which is shown below<br /><img src="http://www3.ntu.edu.sg/home2007/mova0002/snap.PNG" border="0" class="linked-image" /><br />q1) Why is it so that i get pulses in the lower hud is it due to the frame rate limiting that my application does? <br />q2) Where can i find the detials about all the metrics and signals shown in PerfHUD?<br />q3) Since this is mostly a fragment shader code in this application, could any body let me know how to improve the performance of my application currently at the close in position, I get &lt;10 fps and at the startup position i get roughly 30 fps. I need to improve on this. Any suggestions would be appreciated.<br /><br />Thanks,<br />Mobeen<br /><br /><br />]]></description>
		<pubDate>Tue, 13 Oct 2009 06:39:40 -0700</pubDate>
		<guid>http://developer.nvidia.com/forums/index.php?showtopic=3723</guid>
	</item>
	<item>
		<title>Low quality of gp4fp code</title>
		<link>http://developer.nvidia.com/forums/index.php?showtopic=3722</link>
		<description><![CDATA[gp4fp profile generates bad quality code and require many manual optimizations.<br />For example:<br /><!--c1--><div class='codetop'>CODE</div><div class='codemain'><!--ec1-->if &#40;a&#60;5&#41; b=0;<!--c2--></div><!--ec2--><br />generates IF...ENDIF<br />while<br /><!--c1--><div class='codetop'>CODE</div><div class='codemain'><!--ec1-->b = &#40;a&#60;5&#41;?0&#58;b;<!--c2--></div><!--ec2--><br />generates single SLT instruction.<br /><br />This is not a problem in other profiles, for example in fp40 if (a&lt;5) also generates SLT instruction.]]></description>
		<pubDate>Mon, 12 Oct 2009 14:15:19 -0700</pubDate>
		<guid>http://developer.nvidia.com/forums/index.php?showtopic=3722</guid>
	</item>
	<item>
		<title>Planet gravity</title>
		<link>http://developer.nvidia.com/forums/index.php?showtopic=3721</link>
		<description><![CDATA[Hello everybody,<br />I've read something about disable gravity. But I couldn't find anything about enable  <img src="http://developer.nvidia.com/forums/style_emoticons/default/rolleyes.gif" style="vertical-align:middle" emoid=":rolleyes:" border="0" alt="rolleyes.gif" /> <br />We have a planet with a defined mass, does PhysX create a gravity itself? Or have I to create force fields? I hope you can help!<br /><br />with best regards<br /><br />MasterLord]]></description>
		<pubDate>Mon, 12 Oct 2009 10:18:23 -0700</pubDate>
		<guid>http://developer.nvidia.com/forums/index.php?showtopic=3721</guid>
	</item>
	<item>
		<title>Cg 2.3 wishes</title>
		<link>http://developer.nvidia.com/forums/index.php?showtopic=3720</link>
		<description><![CDATA[as was mentioned before, what comes next might be influenced by user desires... although the forum doesn't capture feedback nvidia receives from their developer contacts (would be nice to hear what they wish for Cg's future)<br /><br />- OpenGL forward compatible context (and related GLSL version issues <a href="http://developer.nvidia.com/forums/index.php?showtopic=3607)" target="_blank">http://developer.nvidia.com/forums/index.php?showtopic=3607)</a><br />- on the GLSL side it would be nice to get needed functionality to deal with EXT/ARB_transform_feedback which sadly require low-level program and parameter interaction.]]></description>
		<pubDate>Mon, 12 Oct 2009 06:20:06 -0700</pubDate>
		<guid>http://developer.nvidia.com/forums/index.php?showtopic=3720</guid>
	</item>
	<item>
		<title>NV_gpu_program4 issues with Leopard, Snow Leopard and Apple NVIDIA drivers in general</title>
		<link>http://developer.nvidia.com/forums/index.php?showtopic=3719</link>
		<description><![CDATA[I am writing this post to complain about the general lack of support for NVIDIA OpenGL 2.x features in any capacity under Mac OS.<br /><br />For some time now, NVIDIA has offered an excellent set of features such as the NV_gpu_program4 shader profiles and the texture buffer object extensions, which allow programmers to pull off some pretty amazing tricks with Cg. These features have been available on Windows and Linux for the past 3 years. However, on Mac OS X, developers have seen multiple driver updates and even new OS's with no such implementation. As a result, Cg is all but useless on Mac OS, to developers, researchers and even hobbyists.<br /><br />This would not be an issue if Apple had implemented OpenGL 3.0 in Snow Leopard. However, as far as graphics is concerned, Apple is stuck in the dark ages, barely supporting OpenGL 2 with extensions.<br /><br />From NVIDIA's perspective this should be shameful. Apple is often one of the first OEM's to receive new NVIDIA chips (particularly mobile GPU's) and it is a travesty to have to use Windows on these systems just to access GPU features. <br /><br />I realize Apple is not the most open company with its software, and developing for Mac OS is not as simple as writing a new Linux driver. Perhaps Apple writes their own NVIDIA drivers from scratch, and I should complain to them instead. However, NVIDIA should take responsibility for their product on all platforms. <br /><br />I am begging NVIDIA: please write your own Mac drivers and make them available via your website, alongside Windows, Linux and BSD offerings. I don't care if they are buggy or unsupported -- I have Apple Care in case I fry my GPU. If this is not considered before Snow Leopard is done, I will seriously consider moving away from both NVIDIA and Apple hardware on my next purchase.]]></description>
		<pubDate>Mon, 12 Oct 2009 04:00:24 -0700</pubDate>
		<guid>http://developer.nvidia.com/forums/index.php?showtopic=3719</guid>
	</item>
	<item>
		<title>3dsmax 2010</title>
		<link>http://developer.nvidia.com/forums/index.php?showtopic=3718</link>
		<description><![CDATA[How can one get the beta?<br />this plugin is the ONLY TOOL in our pipeline holding us back all the time.<br />I cannot understand how it can take this long.<br /><br />We are developing high-end simulators, and we are gettting into serious trouble<br />developing because physx plugin is not max 2010 yet.]]></description>
		<pubDate>Mon, 12 Oct 2009 01:04:08 -0700</pubDate>
		<guid>http://developer.nvidia.com/forums/index.php?showtopic=3718</guid>
	</item>
	<item>
		<title>Softimage PhysX</title>
		<link>http://developer.nvidia.com/forums/index.php?showtopic=3717</link>
		<description><![CDATA[Hi, Max & Maya plugins a great, but what with Softimage? When you add this cool new features in Softimage?]]></description>
		<pubDate>Mon, 12 Oct 2009 00:06:12 -0700</pubDate>
		<guid>http://developer.nvidia.com/forums/index.php?showtopic=3717</guid>
	</item>
	<item>
		<title>How to adjust field of view?</title>
		<link>http://developer.nvidia.com/forums/index.php?showtopic=3716</link>
		<description><![CDATA[It would be really nice if I could adjust the FOV of the camera in FX Composer.  Is there a way to do this?  It seems like I've looked everywhere.]]></description>
		<pubDate>Sun, 11 Oct 2009 15:27:22 -0700</pubDate>
		<guid>http://developer.nvidia.com/forums/index.php?showtopic=3716</guid>
	</item>
	<item>
		<title>How many Actors is too many</title>
		<link>http://developer.nvidia.com/forums/index.php?showtopic=3714</link>
		<description><![CDATA[I have been playing around with PhysX for a little while now as part of a hobby FPS game but have run into a bit of a problem. <br />I have a rather large outdoor scene with 8000 odd trees and was looking to be able to add some static actors to represent them so that my character controller does not walk through them and any other objects can bounce off them. The trees don't need to be able to move or anything basically just get in the way.<br /><br />But when I create 8000 actors I am no longer able to create a cube to throw at them, the cube just never appears. When I try with less trees say 1000 it works fine I can throw the cube at the tree and it bounces off correctly. Each Tree is implemented as a single capsule shape. Also with 8000 trees the charcter controller bounces off all of them correctly (well all the ones I have tried running into randomly at least)<br /><br />So two questions is there a limit to the number of actors I can add to a scene? and if there is a limit does anybody have a suggestion for implementing a large number of static obstacles?<br />Any help will be much appreciated.]]></description>
		<pubDate>Sun, 11 Oct 2009 00:52:25 -0700</pubDate>
		<guid>http://developer.nvidia.com/forums/index.php?showtopic=3714</guid>
	</item>
	<item>
		<title>Problems with Shadow mapping</title>
		<link>http://developer.nvidia.com/forums/index.php?showtopic=3713</link>
		<description><![CDATA[In the last two weeks I've been working on a shadow mapping application, but there is a problem I don't know how to solve : polygon offset. I've tried lots of things but none of them worked well. <br /><br /><div class='codetop'>CODE</div><div class='codemain' style='height:200px;white-space:pre;overflow:auto'><br />struct VertOut{<br />	float4 Position: POSITION;<br />	float4 Color: COLOR;<br />	float4 ShadowTexCoord: TEXCOORD0;<br /><br />	};<br />uniform float4x4 WorldViewProj;<br />uniform float4x4 WorldView;<br />uniform float4x4 TextureMatrix;<br />uniform float4x4 ModelViewIT;<br />uniform float4 GlobalAmbient;<br />uniform float4 Ma, Md;<br />uniform float4 LightPos;<br />uniform float4 La,Ld;<br /><br /><br /><br />VertOut mainVS(float4 pos : POSITION,/* float2 TC:TEXCOORD1,*/<br />			   float4 Color:COLOR,<br />			   float4 Normal: NORMAL){<br />	VertOut vo;<br />	vo.Position = mul(WorldViewProj, pos);<br /><br />	<br />	vo.ShadowTexCoord = mul(TextureMatrix,pos);<br />	pos.z += .5f;<br />	<br />	vo.ShadowTexCoord *= 1.0f/vo.ShadowTexCoord.w;<br />	float4 EyeVertex = mul(WorldView, pos);<br />	float4 LightDir = normalize(LightPos - pos);<br />	float Cosine = max(dot(mul(ModelViewIT,Normal),LightDir),0);	<br />	vo.Color = (GlobalAmbient + Ld) * Ma + Md * Ld * Cosine;<br />	<br />	return vo;<br />}<br /><br />uniform sampler2D ShadowMap;<br />const float Resolution = 1.0f / 512.0f;	<br />#define SHADOW_COMPARISON 0.04<br /><br />float ShadowStep(float2 TexCoords, float FragmentDepth)<br />{<br />	float Result = 0;<br />	float Depth = tex2D(ShadowMap, TexCoords).r;<br />	<br />	return (FragmentDepth &lt; Depth + SHADOW_COMPARISON);<br />}<br />float4 mainPS(float4 ShadowTexCoord: TEXCOORD0, <br />			   float4 Color:COLOR) : COLOR {<br />	<br />	float2 ShadowCoord = ShadowTexCoord.xy;<br />	float FragmentDepth = ShadowTexCoord.z;<br />	<br />	<br />	float ShadowAccum = ShadowStep(ShadowCoord + float2(-Resolution,0),FragmentDepth);<br />	ShadowAccum += ShadowStep(ShadowCoord + float2(Resolution,0),FragmentDepth);<br />	<br />	ShadowAccum += ShadowStep(ShadowCoord + float2(0,-Resolution),FragmentDepth);<br />	<br />	ShadowAccum += ShadowStep(ShadowCoord + float2(0,Resolution)  ,FragmentDepth);<br />	ShadowAccum += ShadowStep(ShadowCoord,FragmentDepth);<br />	ShadowAccum *= 0.2;<br />	<br />	return  ShadowAccum* Color;<br />}<br /><br />technique technique0 {<br />	pass p0 {<br />		CullFaceEnable = false;	<br />		VertexProgram = compile vp30 mainVS();<br />		FragmentProgram = compile fp30 mainPS();<br />	}<br />}</div><br /><br /><br />Ma and Md: Material's ambient and diffuse color.<br />La and Ld: Light's ambient and diffuse color. <br />LightPos: Light Position already in eye space. <br />WorldView: ModelView matrix.<br />WorldView: Modelview inverse transpose.<br />TextureMatrix: Matrix concatenation for standard projective texturing. It's calculated in the same way as The CG tutorial does.<br />WorldViewProj: ModelviewProjection matrix.<br /><br /><br />This is my actual code. My depth map texture has 512x512. With this code I'm averaging the four closest neighbors and the current pixel to determine the final shadow contribution. I've also tried with the one sample version and the result is a very aliased scene, and if I take SHADOW_COMPARISON too small then the scene looks ugly due to shadow acne.<br />I don't use OpenGL's polygon offset functionality because I don't know how to properly set the values factor and units, instead in the shadow step function I use a constant offset to compare the depths, but due to the nature of the depth buffer it's probably wrong. Yesterday I found an article on Game Programming Gems I, section 4.1: Tweaking a vertex's projected depth value and I've implemented the OffsetProjectionMatrix method but the result of the maximum effective value for the eye space offset I calculated with the equation 1.1.10 to offset a polygon away from the camera is very very very small. So....right now I'm lost with this shadow mapping implementation so, any information, link, tutorial or anything will be well received. One thing more: sorry my English......Thanks in advance.]]></description>
		<pubDate>Sun, 11 Oct 2009 00:03:53 -0700</pubDate>
		<guid>http://developer.nvidia.com/forums/index.php?showtopic=3713</guid>
	</item>
	<item>
		<title>Vehicle drift</title>
		<link>http://developer.nvidia.com/forums/index.php?showtopic=3712</link>
		<description><![CDATA[Can someone throw me in the right direction on how to implement drifting?  <br />Also, maybe I am not understanding everything but whats the purpose of having a "wheel" and a "wheelshape"?  It seems to be redundant when they go through a wheeldesc set values then go through a wheelshapdesc and basically just send everything from wheeldesc to it.]]></description>
		<pubDate>Sat, 10 Oct 2009 09:12:02 -0700</pubDate>
		<guid>http://developer.nvidia.com/forums/index.php?showtopic=3712</guid>
	</item>
	<item>
		<title><![CDATA[Kinematic rigid bodies & the max plugin]]></title>
		<link>http://developer.nvidia.com/forums/index.php?showtopic=3711</link>
		<description><![CDATA[I'm having some serious issues with kinematic rigid bodies in my simulations. I'm using max 2009 + 1.06 of the physx plugin.<br /><br /><a href="http://www.youtube.com/watch?v=by2XS1IL72U" target="_blank">Here's a link to a youtube video</a> showing what's going wrong. (The max scene is also attached if you want to have a look at how it's build)<br /><br />- On the left you see a simple dynamic box falling onto a static object, this goes just fine.<br />- Next to that is the same box falling onto a kinematic body. As you can see the box seems to go to sleep and ignores the kinematic body's movement completely until it has fully passed when it suddenly remembers that it is supposed to fall down.<br />- In the third situation I raised the box a bit so it does not stabilize before the kinematic body starts to move. This seems to prevent it from going to sleep but the box still behaves strangely and does not properly follow the kinematic body even though they have very high friction.<br />- On the right we then have the behavior that I would expect. This was achieved by making the kinematic body a dynamic body and constraining it to a dummy outside of the simulation.<br /><br />Is there any way to make a kinematic body work properly? I'm glad I at least found the workaround with the dummy, but I'm currently working with a sim involving hundreds of kinematic bodies and tens of thousands of rigid bodies, so this extra layer of complexity is not really helping to keep max from constant out of memory crashes.*<br /><br />Thanks for the help,<br />Arne Bezuijen<br /><br />* Yes this is also a not so subtle hint that I would really really like a 64 bit version of the plugin <img src="http://developer.nvidia.com/forums/style_emoticons/default/smile.gif" style="vertical-align:middle" emoid=":)" border="0" alt="smile.gif" />]]></description>
		<pubDate>Sat, 10 Oct 2009 05:33:31 -0700</pubDate>
		<guid>http://developer.nvidia.com/forums/index.php?showtopic=3711</guid>
	</item>
	<item>
		<title><![CDATA[nvPerfHUD 6.6 : error launching application "cannot initialize OLE"]]></title>
		<link>http://developer.nvidia.com/forums/index.php?showtopic=3710</link>
		<description><![CDATA[Since I have installed version 6.6, I can't start my application anymore, I get this "cannot initialize OLE" error message.<br /><br />I am using Vista 64, latest Forceware 191.07, and Direct3D 10.<br /><br />Any idea what could be causing this issue (6.51 was OK) ?<br /><br />Cheers,<br />Gregory]]></description>
		<pubDate>Fri, 09 Oct 2009 05:44:22 -0700</pubDate>
		<guid>http://developer.nvidia.com/forums/index.php?showtopic=3710</guid>
	</item>
	<item>
		<title>Core Dump</title>
		<link>http://developer.nvidia.com/forums/index.php?showtopic=3709</link>
		<description><![CDATA[Hello<br /><br />I'm using the Coredump Functions of Lesson 503 with standard actors, convex meshes and joints. <br /><br />When I try to load the exported file (attached), I get a debug error: "<b>HEAP CORRUPTION DETECTED": after Normal block (#292) at 0x035F5A70. CRT detected that the application wrote to memory after end of heap buffer.</b>"<br />If I click Ignore, it seems to work correctly.<br />But how can I avoid this Error Message?. Is it possible, that not enough memory is allocated or where can I begin to find the Error?<br /><br />Thanks in Advance.<br />Mada]]></description>
		<pubDate>Fri, 09 Oct 2009 02:08:30 -0700</pubDate>
		<guid>http://developer.nvidia.com/forums/index.php?showtopic=3709</guid>
	</item>
	<item>
		<title>wglCreateAffinityDCNV only works on GPU #1</title>
		<link>http://developer.nvidia.com/forums/index.php?showtopic=3708</link>
		<description><![CDATA[I've integrated wglCreateAffinityDCNV() into a project in the hopes of creating an OpenGL window with a render context affinity mask locking everything to GPU #2, drawing to a display attached to GPU #2. When I set the affinity exclusively to GPU #1, everything works correctly. I can draw to my window and everything looks fine except, of course, I'm running off my primary GPU, instead of #2.<br /><br />When I set the affinity exclusively to GPU #2, the call to wglMakeCurrent() fails, with GetLastError() returning 0xc00720d0. I can't find a reference to this error code anywhere.<br /><br />I've enabled and disabled Aero and the Desktop Windows Manager. I've tried changing all the 3D settings I could find in the NVidia control panel. I haven't seen any applications or sample code that does what I'm trying to do except Equalizer, which demonstrates exactly the same problems I've been experiencing with my own code.<br /><br />Interestingly, on another note, my pair of Quadro FX 4800 cards don't seem to want to enable SLI. I'm definitely running on a motherboard that supports it. Various tools have reported that SLI is not an option for me, even with the bridge attached. I'm not too worried, as I cannot use SLI for this particular project anyway, but is it possible that this is symptomatic of a bigger problem? <br /><br />When I remove the primary GPU from the slot, GPU #2 works just fine.<br /><br />Machine stats:<br />Windows 7 Ultimate 64-bit<br />Two Quadro FX 4800 cards<br />191.07 GeForce video drivers<br />ASUS P6T6 WS Revolution Motherboard<br /><br />Where else can I look for clues? I'm running out of ideas. Any thoughts? Leads?<br /><br />Thanks!<br />]]></description>
		<pubDate>Thu, 08 Oct 2009 16:08:19 -0700</pubDate>
		<guid>http://developer.nvidia.com/forums/index.php?showtopic=3708</guid>
	</item>
	<item>
		<title>Cg 2.2 October 2009 released</title>
		<link>http://developer.nvidia.com/forums/index.php?showtopic=3707</link>
		<description><![CDATA[Cg 2.2 October 2009 is now available at <a href="http://developer.nvidia.com/object/cg_toolkit.html" target="_blank">http://developer.nvidia.com/object/cg_toolkit.html</a><br /><br />The main focus of this release is on fixing bugs, many of which were reported on this forum.  (We do listen to you guys.   <img src="http://developer.nvidia.com/forums/style_emoticons/default/smile.gif" style="vertical-align:middle" emoid=":)" border="0" alt="smile.gif" /> )<br /><br />On that note, I want to say how much we appreciate the feedback from everyone on the forum.  You are helping us make Cg a better product, and that's our primary goal.  So keep the suggestions and bug reports coming; we'll do our best to address them.<br /><br />Scott]]></description>
		<pubDate>Thu, 08 Oct 2009 10:37:15 -0700</pubDate>
		<guid>http://developer.nvidia.com/forums/index.php?showtopic=3707</guid>
	</item>
	<item>
		<title>GLSL compiler bug</title>
		<link>http://developer.nvidia.com/forums/index.php?showtopic=3706</link>
		<description><![CDATA[I get a "fatal error C9999" "error during compilation" with the attached example (I'm using the driver version 191.07 on Windows 7 with a 8800 GTS). The first program compiles correctly. The second one gives the error, although it is identical (and shares a common shader object with the first program). The error disappears if I remove the "vec3 f(vec3 p);" declaration in "shaderCode1", or if replace the body of the "g" function with "return texture1D(mySampler, r);"<br />]]></description>
		<pubDate>Thu, 08 Oct 2009 09:40:59 -0700</pubDate>
		<guid>http://developer.nvidia.com/forums/index.php?showtopic=3706</guid>
	</item>
	<item>
		<title>GTX260m</title>
		<link>http://developer.nvidia.com/forums/index.php?showtopic=3705</link>
		<description><![CDATA[im not 100% sure if this is the place for this forum post but i have a quick question. I have a g71 with the bios update, and it has a 260m. is it safe to overclock this gpu and if so what software should i use to overclock it.<br /><br />Thank you <img src="http://developer.nvidia.com/forums/style_emoticons/default/biggrin.gif" style="vertical-align:middle" emoid=":D" border="0" alt="biggrin.gif" />]]></description>
		<pubDate>Thu, 08 Oct 2009 02:38:35 -0700</pubDate>
		<guid>http://developer.nvidia.com/forums/index.php?showtopic=3705</guid>
	</item>
	<item>
		<title>OpenGL Context on a secondary GPU - and other problems</title>
		<link>http://developer.nvidia.com/forums/index.php?showtopic=3704</link>
		<description><![CDATA[I'm having a problem where I've been unable get any OpenGL application to take advantage of the second GPU on my system.<br /><br />I have two Quadro FX 4800 cards, no SLI, Win7 x64 and, as far as I can tell, all the latest drivers. I somehow got it into my head that I could do the following: open a window entirely within a display attached to my secondary graphics adapter, create a context for that window, and then all subsequent OpenGL calls would be routed to that GPU, leaving the primary GPU largely untouched, and available for near-exclusive use by another process. (I understand that OSX works something like that.) This has not turned out to be the case.<br /><br />I cannot get my app to run on the second GPU. I call CreateWindowEx() with parameters to open the window [and then keep it] entirely on one of these secondary monitors. No luck. I've tried disabling Windows Aero... still no luck.<br /><br />I should mention that I'm using the RivaTuner monitor to determine which GPU is active at any given time, using the Voltage Regulator Current readout for each GPU.<br /><br />The only partial success I've had is by using wglCreateAffinityDCNV() with a mask to permit activity only on the secondary GPU. Then, my app does seem to run [mostly] on the secondary GPU! (The primary is still a little busy doing... something.) But I can no longer draw to the window because Affinity-DCs cannot be bound to a window. I think. This extension is NOT well-documented, and I have found no examples of its use. I've not looked at the source for Equalizer, as I suspect it's just using affinity-DCs for rendering to FBOs, not to an actual window. Maybe it would still be useful?<br /><br />The whole point of this exercise is to draw at 60fps to all 4 graphics outputs on the machine. As it is, I can create a single window that spans my 3840x2160 quad 1920x1080 desktop and get a solid 30 fps with VSync on. This would actually be fine for my application, except for one problem: Even though I see 30fps, OpenGL reports that it's drawing 60fps. I'm calling SwapBuffers() 60 times per second, but only every-other frame is visible on screen. I can't tell programmatically WHICH frames are visible, which means that later on, I'll have trouble frame-syncing my application across machines.<br /><br />Now, when I first execute my application in a window stretched across two monitors, both attached to the primary adapter, I get a smooth 60fps. Even with a window stretched across the two displays on the secondary adapter, I get a smooth 60 for a second or two before it falls to 30. (Currently, I'm just rendering a few rectangles, so it's not that my scene is overly complex.) It would be acceptable to run two entirely separate instances of the application, one for each GPU... if I could be sure that the second instance would run on the second GPU, and I'm hoping: thus run at 60fps.<br /><br />Anyway, anybody have any ideas? Hopefully there's something simple that I'm overlooking, but I've searched around for days now to no avail.<br /><br />Also - I've been holding off on posting about this issue anywhere because there seem to be so many different facets to the problem... Would I be more likely to get assistance by breaking down my problems into smaller posts?<br /><br />Thanks for treading to the end. Any help would be REALLY appreciated!]]></description>
		<pubDate>Wed, 07 Oct 2009 14:57:05 -0700</pubDate>
		<guid>http://developer.nvidia.com/forums/index.php?showtopic=3704</guid>
	</item>
	<item>
		<title>friction between soft body and surface</title>
		<link>http://developer.nvidia.com/forums/index.php?showtopic=3703</link>
		<description><![CDATA[I'm having trouble increasing the friction between my soft body and the ground plane.  I am assuming that the ground plane inherits the default material properties:<br /><br />NxMaterialDesc       m; <br />	m.restitution        = 0.00f; // was 0.2<br />	m.staticFriction     = 9999; // was 0.2<br />	m.dynamicFriction    = 0.99; // was 0.2 should not be higher than 1<br />	m.frictionCombineMode = NX_CM_MAX; //the maximum of two surface frictions<br />	NxMaterial * mat = gScene-&gt;getMaterialFromIndex(0);<br />	mat-&gt;loadFromDesc(m); <br /><br /><br />and here is my softbody description:<br /><br />	NxSoftBodyDesc softBodyDesc;<br />	softBodyDesc.density = 0.5f; //31.5 g/cm3 = 0.004 g/unit^3<br />	softBodyDesc.globalPose.t = NxVec3(0.0f, 2.6f, 0.0f);<br />	softBodyDesc.particleRadius = 0.01f; //was 0.2<br />	softBodyDesc.volumeStiffness = 1.f;<br />	softBodyDesc.stretchingStiffness =1.f;<br />	softBodyDesc.friction = 1.0f;<br />	softBodyDesc.attachmentResponseCoefficient = 0.1f;<br />	softBodyDesc.dampingCoefficient = 0.99;<br />	softBodyDesc.solverIterations = 7;<br />	softBodyDesc.relativeGridSpacing = 0.1;<br /><br />	if (gHardwareSimulation)<br />		softBodyDesc.flags |= NX_SBF_HARDWARE;<br /><br />	//softBodyDesc.flags |= NX_CLF_COMDAMPING; // only damp intra-body motion, not translation and rotation<br />//	softBodyDesc.flags |= NX_SBF_VOLUME_CONSERVATION; //enable volume conservations<br />	softBodyDesc.flags |= NX_SBF_DAMPING; //enable damping<br />	softBodyDesc.flags |= NX_SBF_SELFCOLLISION;<br />	softBodyDesc.flags |= NX_SBF_COLLISION_TWOWAY;<br /><br /><br />How can I increase friction?  Again, my code is largely based upon the sampleSoftBody code, and I am running the remote debugger.<br />]]></description>
		<pubDate>Wed, 07 Oct 2009 12:59:54 -0700</pubDate>
		<guid>http://developer.nvidia.com/forums/index.php?showtopic=3703</guid>
	</item>
	<item>
		<title>2.8.3 Training Program Executables Crash</title>
		<link>http://developer.nvidia.com/forums/index.php?showtopic=3702</link>
		<description><![CDATA[I downloaded and installed the latest System Software and 2.8.3 SDK; didn't uninstall 2.8.1 or the older System Software.<br /><br />When I run any of the training program executables it crashes almost immediately. Compiling them in debug I see that NxCreatePhysicsSDK(NX_PHYSICS_SDK_VERSION) is returning NULL, thus causing a crash later. Has anyone been able to run with 2.8.3?]]></description>
		<pubDate>Wed, 07 Oct 2009 12:16:18 -0700</pubDate>
		<guid>http://developer.nvidia.com/forums/index.php?showtopic=3702</guid>
	</item>
	<item>
		<title>Soft Bodies behave differently in hardware mode</title>
		<link>http://developer.nvidia.com/forums/index.php?showtopic=3701</link>
		<description><![CDATA[I am implementing a softbody simulator in PhysX.   My framework is based largely around the SampleSoftBody demo, and in fact I haven't changed much of anything in the underlying code.<br /><br />So far, when I run the simulation in software mode only it behaves as expected.  However, my simulation behaves drastically differently (and and bizarrely) when I turn hardware support on (I have three GX 280s in triple SLI mode, and ought to be able to use them for *something*!).  Could anyone give advice about where to begin understanding where this problem may arise?<br /><br />I thought it might be a problem with cooking, but that is all done by the provided MySoftBody code, which I haven't touched at all.<br /><br />(I should add that I have the code set up for remote debugging, should anyone wish...)<br /><br />regards,<br />jr]]></description>
		<pubDate>Wed, 07 Oct 2009 10:17:45 -0700</pubDate>
		<guid>http://developer.nvidia.com/forums/index.php?showtopic=3701</guid>
	</item>
	<item>
		<title>A8R8G8B8 DDS file generation...?</title>
		<link>http://developer.nvidia.com/forums/index.php?showtopic=3700</link>
		<description><![CDATA[I'm currently trying to use nvcompress.exe to create a quick tool to convert some old textures to the *.dds format. Some of the textures need to be DXT5 compressed, while others need to be lossless DDS (A8R8G8B8). Is there a way to get nvcompress.exe to generate an uncompressed *.dds from a given *.png?<br /><br />I originally thought that the "-rgb" option would do this, however the resulting texture is certainly <b>not</b> lossless, so I'm guessing that it's not what I thought it was?  <img src="http://developer.nvidia.com/forums/style_emoticons/default/blink.gif" style="vertical-align:middle" emoid=":blink:" border="0" alt="blink.gif" />]]></description>
		<pubDate>Wed, 07 Oct 2009 08:12:03 -0700</pubDate>
		<guid>http://developer.nvidia.com/forums/index.php?showtopic=3700</guid>
	</item>
	<item>
		<title>How to apply external forces on rigid bodies in PhysX maya plugin?</title>
		<link>http://developer.nvidia.com/forums/index.php?showtopic=3699</link>
		<description><![CDATA[Hi,<br />I want to manipulate rigid bodies in maya based on external forces. I am using  physX plugin for maya. I have managed to relocate active rigid bodies by MEL scipting, but how to apply forces/torques on them I could figure out. When I create an active rigid body, among others it will have the following attributes:<br />inputForce, generalForce, outputForce, outputTorque. Could anyone explain to me what are these attributes? Which one should be used for applying external forces on rigid bodies, and how? I would really appreciate an example MEL script.<br />Thanks<br />Zoltan]]></description>
		<pubDate>Wed, 07 Oct 2009 00:31:12 -0700</pubDate>
		<guid>http://developer.nvidia.com/forums/index.php?showtopic=3699</guid>
	</item>
	<item>
		<title><![CDATA[PhysX 'floor' one world unit offset from real geometry]]></title>
		<link>http://developer.nvidia.com/forums/index.php?showtopic=3698</link>
		<description>Edit:  Typing up the post was apparently enough.  Please delete.</description>
		<pubDate>Tue, 06 Oct 2009 15:15:17 -0700</pubDate>
		<guid>http://developer.nvidia.com/forums/index.php?showtopic=3698</guid>
	</item>
	<item>
		<title>how to debug vetex shaders?</title>
		<link>http://developer.nvidia.com/forums/index.php?showtopic=3697</link>
		<description><![CDATA[I used to use PIX, but it's off line, someone can help ?]]></description>
		<pubDate>Mon, 05 Oct 2009 19:07:20 -0700</pubDate>
		<guid>http://developer.nvidia.com/forums/index.php?showtopic=3697</guid>
	</item>
	<item>
		<title>PhysX SDK for PS3</title>
		<link>http://developer.nvidia.com/forums/index.php?showtopic=3696</link>
		<description><![CDATA[Hi, <br /><br />Nowadays I've had to create physics simulation using Physx  in a PS3 platform. can anyone help me? <br /><br />Where can i find documentation to start?<br />How can i get the PhysX SDK for PS3?<br /><br />Thank you.<br />]]></description>
		<pubDate>Mon, 05 Oct 2009 11:23:59 -0700</pubDate>
		<guid>http://developer.nvidia.com/forums/index.php?showtopic=3696</guid>
	</item>
	<item>
		<title>raycast groups?</title>
		<link>http://developer.nvidia.com/forums/index.php?showtopic=3695</link>
		<description><![CDATA[Hi<br /><br />Could someone explain the simplest possible way to perform a ray cast on all shapes (referably using raycastClosestShape(...)) and, for example, exclude all actor/shapes of group id 2 and 3?<br />I'm trying to use this for line of sight testing in my game but I just can't seem to get my head around it, either the raycast only returns shapes in the default group or the group I send to raycastClosestShape(). But what I want to do is to tell it to exclude a couple of groups instead of raycast a single group. How have others solved this problem?]]></description>
		<pubDate>Sun, 04 Oct 2009 12:49:41 -0700</pubDate>
		<guid>http://developer.nvidia.com/forums/index.php?showtopic=3695</guid>
	</item>
	<item>
		<title>List of supported internalformat and type for hardware based PCF.</title>
		<link>http://developer.nvidia.com/forums/index.php?showtopic=3694</link>
		<description><![CDATA[With the following code I can set up the GPU to do a logic operation on texture samples before filtering<br /><br /><br />glTexParameteri( mTarget, GL_TEXTURE_COMPARE_MODE_ARB, GL_COMPARE_R_TO_TEXTURE_ARB );<br />glTexParameteri( mTarget, GL_TEXTURE_COMPARE_FUNC_ARB, GL_LEQUAL );<br /><br /><br />Lets say I call tex2D( texDepth, P ), then the third component of P will be compared with the texture value before filtering stage, effectively allowing us to do 2x2 PCF in hardware.<br /><br />This has been supported by nVidia for ages and lately also ATI? Is there a list somewhere that pairs GPU's with supported depth texture format and types? Here some the possible values:<br /><br /><br />internalformats:<br />GL_DEPTH_COMPONENT<br />GL_DEPTH_COMPONENT16<br />GL_DEPTH_COMPONENT24<br />GL_DEPTH_COMPONENT32<br /><br />types:<br />GL_FLOAT<br />GL_UNSIGNED_BYTE<br />GL_UNSIGNED_INT<br /><br /><br />RANDOM QUESTIONS:<br /><br />1. What combination of hardware and formats enables PCF in hardware?<br /><br />2. Does using FLOAT actually mean that the depth value is not being normalized to ( 0, 1 ).<br /><br />3. Obviously GL_DEPTH_COMPONENT24 and GL_UNSIGNED_INT is not a valid combination. How can I check what format OpenGL settles on?<br /><br />4. Would I be better of just dropping hardware based PCF and stick with my shader 2x2 PCF.<br /><br /><br />If you have an answer to any of the above you will be of great help! Thank you.]]></description>
		<pubDate>Sun, 04 Oct 2009 07:47:20 -0700</pubDate>
		<guid>http://developer.nvidia.com/forums/index.php?showtopic=3694</guid>
	</item>
	<item>
		<title>program not loading when using tex2D() for shadow comparison</title>
		<link>http://developer.nvidia.com/forums/index.php?showtopic=3693</link>
		<description><![CDATA[I want to use tex2D( texDepth, P ) where texDepth is a depth map and P is projected texture coordinates in light space. The third component of P is the z component. This is the standard setup for hardware based 2x2 PCF. My problem is that when loading the shader with cgGLLoadProgram() I get an error "the program could not load". Natually the screen also go blank. If I simply use tex2D( tex, P.xy ) for normal projected texturing the shader loads fine.<br /><br />I am using OpenGL and my hardware is nVidia GTX 260 and ofcourse supports shader model 4 as required (according to Cg reference). I also check for the appropriate extensions, shader versions, depth texture. I am pretty sure my texture setup and all that on the client side is fine too, but that shouldnt matter anyway since it seems to be a problem of loading the shader.<br /><br />So whats going wrong here? What can I possibly test now to get to the core of this?]]></description>
		<pubDate>Sun, 04 Oct 2009 07:30:56 -0700</pubDate>
		<guid>http://developer.nvidia.com/forums/index.php?showtopic=3693</guid>
	</item>
	<item>
		<title>cgD3D9SetTextureParameter with NULL</title>
		<link>http://developer.nvidia.com/forums/index.php?showtopic=3692</link>
		<description><![CDATA[Direct3D Debug mode says:<br /><blockquote>Direct3D9: VS-&gt;PS Linker: X442: (Error) Current pixel shader declares that a cube texture must be at sampler [6]. However, sampler [6] currently has a 2d texture set.</blockquote><br />Before <b>DrawIndexedPrimitive()</b> call I set cube effect parameter with <b>cgD3D9SetTextureParameter</b> to <b>NULL</b>. Then I call <b>cgSetPassState()</b>, which should set sampler 6 to <b>NULL</b>.<br />Why do I get this error? What's going on?]]></description>
		<pubDate>Sun, 04 Oct 2009 02:52:51 -0700</pubDate>
		<guid>http://developer.nvidia.com/forums/index.php?showtopic=3692</guid>
	</item>
	<item>
		<title>Fail to run HDR with 2x FP16 MRTs</title>
		<link>http://developer.nvidia.com/forums/index.php?showtopic=3691</link>
		<description><![CDATA["HDR with 2x FP16 MRTs", the demo in NVIDIA SDK 9.5 .I remeber I had ever successed running it when I was using 7300GT video card. Now , I'm using 9300GS. But the demo display noting ,just a black ground. when I debug this demo .I found It's failed to DrawPrimitiveUp().But I can't find the reason ,Dose My 9300GS don't support it. I also find that many demo which I download from gamedev cant't running .the same black ground. my driver doesn't have any problem.]]></description>
		<pubDate>Sat, 03 Oct 2009 19:21:07 -0700</pubDate>
		<guid>http://developer.nvidia.com/forums/index.php?showtopic=3691</guid>
	</item>
	<item>
		<title>Faulty fan on Nvidia fx5200</title>
		<link>http://developer.nvidia.com/forums/index.php?showtopic=3690</link>
		<description>I am currently experiencing fan failure on my FX5200 video card. I was seeking a way to rectify this problem. The fan has increased the noise that it produces and performance is lagging. I believe this to be a wonderful product of your corporation and I am a long time Nvidia product user. Please direct me in any method to solve my dilemma. Thank you for your time and consideration.</description>
		<pubDate>Sat, 03 Oct 2009 04:23:03 -0700</pubDate>
		<guid>http://developer.nvidia.com/forums/index.php?showtopic=3690</guid>
	</item>
	<item>
		<title>CgFX : Problem with parmeter array on vp40 profile ?</title>
		<link>http://developer.nvidia.com/forums/index.php?showtopic=3689</link>
		<description><![CDATA[Hi, <br /><br />I have some strange problems with parameters array... I've something classical like that :<br /><br /><div class='codetop'>CODE</div><div class='codemain' style='height:200px;white-space:pre;overflow:auto'>float4x4 mvp;<br />float4 bones[80 * 3];<br /><br />struct SkinningVertIn<br />{<br />	float3 position : POSITION;<br />	float2 tcoord : TEXCOORD0;<br />	float4 weights : TEXCOORD4;<br />	float4 boneIndices : TEXCOORD5;<br />};<br /><br />struct SkinningVertOut<br />{<br />	float4 hpos : HPOS;<br />	float3 color : COLOR0;<br />	float2 tcoord : TEXCOORD0;	<br />};<br /><br />float3x4 getBoneMatrix(int index)<br />{<br />	return float3x4(bones[index * 3 + 0],<br />					bones[index * 3 + 1],<br />					bones[index * 3 + 2]);<br />}<br /><br />// Skinning vertex program<br />SkinningVertOut skinningVP(SkinningVertIn IN)<br />{<br />	SkinningVertOut OUT;<br /><br />	OUT.tcoord = IN.tcoord;<br /><br />            float4 pos = mul(getBoneMatrix(indices.x), float4(IN.position,1)).xyz;<br /><br />	OUT.color = 1;<br /><br />	OUT.hpos = mul(mvp,float4(pos,1));<br />	return OUT;<br />}<br /><br />technique Skinning<br />{<br />	pass<br />	{<br />		VertexProgram = compile gp4vp skinningVP();<br />	}<br />}</div><br /><br />I put identity matrices in the bones array, and indices are in the good range...<br />And it seems to work only with vp30, gp4vp and glslv profiles. Not working at all with the vp40 profile, sometimes nothing apears, sometimes object is completely distorted. Looking in assembly code between vp30 and vp40 version does not show any noticeable differences, but the vp40 version does not work at all... (not working on vp20 due to large number of constants).<br /><br />Passing constant directly to the getBoneMatrices function work fine in all the profiles....<br /><br />Is any one else have this kind of problems ? or perhaps I'm doing something wrong ?<br /> <br /><br />]]></description>
		<pubDate>Fri, 02 Oct 2009 16:54:45 -0700</pubDate>
		<guid>http://developer.nvidia.com/forums/index.php?showtopic=3689</guid>
	</item>
	<item>
		<title>3D particle system</title>
		<link>http://developer.nvidia.com/forums/index.php?showtopic=3688</link>
		<description><![CDATA[Hello everyone,<br /><br />Does anyone know of some online resource which shows haw to implement a simple 3D particle system. I looked at the 2D example in the CG tutorial book and would like to find out how one can create 3D particles and have the perception of depth, alpha blending etc.<br /><br />I did some google search but did not find anything. There is an example in GPU gems (3, I think). However, they are too complicated and I would like something simple so that it is easier to learn the key concepts.<br /><br />Many thanks,<br /><br />xarg]]></description>
		<pubDate>Fri, 02 Oct 2009 15:02:45 -0700</pubDate>
		<guid>http://developer.nvidia.com/forums/index.php?showtopic=3688</guid>
	</item>
	<item>
		<title>Neutrally Buoyant Rigidbody</title>
		<link>http://developer.nvidia.com/forums/index.php?showtopic=3687</link>
		<description><![CDATA[I am new to PhysX and these forums, so first thing: Hi,<br /><br />second, I am trying to simulate (as close as possible) real object buoyancy using the density and calculated volumes of shapes attached to an actor (not concerned with surface buoyancy where a ration of the volume is sticking out).  This is my current implementation:<br /><br />for every shape in the water<br />1. compute mass of water displaced (shape volume * water density) m = d * v<br />2. apply it to a vector in world up direction Fb = NxVec3(0, m, 0) - i tried F = m * a with "a" being -9.8 but the force seems to be too much so I am not sure whats going on here but just using the mass alone makes it seem a bit more realistic!<br />3. Final buoyancy force = Fb * 0.1 (centimeter to meter conversion) - ( part-&gt;getLinearVelocity() * 2 ) - my system is in meters so a sphere with radius 0.5 is 50 cm.<br />4. finally in a separate thread actor-&gt;addForceAtLocalPos( final buoyancy force, local position of shape on actor, NX_FORCE (default) )<br /><br />This seems to work, but is very frame rate dependent and I can not get an object of 1000 density to stay still in the middle of water volume due to this.  Also, I am confused on why I get massive numbers when I calculate the force using Fb = mass water displayed * -9.8?<br /><br />Any hints, suggestions, ideas would be greatly appreciated (perhaps I am doing this all wrong? and should use a different force application mode?)<br /><br />Bob2u<br /><br /><br />**** EDIT<br /><br />I think I found an OK solution.  I do "Fb = NxVec3(0, - fluid density * obj volume * -9.8 * timeStep, 0) - (object velocity * 1.09)" and it seems to act just about right.  I did however change to Fixed time step and now use addForceAtPos( Fb, pos, NX_SMOOTH_IMPULSE ).  It does what I need for now but I sure wanted to figure out how to get this working with variable timestep.<br /><br />Reference used to derive solution <a href="http://www.gamedev.net/community/forums/topic.asp?topic_id=498214" target="_blank">http://www.gamedev.net/community/forums/to...topic_id=498214</a>]]></description>
		<pubDate>Fri, 02 Oct 2009 07:38:38 -0700</pubDate>
		<guid>http://developer.nvidia.com/forums/index.php?showtopic=3687</guid>
	</item>
	<item>
		<title>nTune... SUCKS!!!</title>
		<link>http://developer.nvidia.com/forums/index.php?showtopic=3685</link>
		<description><![CDATA[I just wanted to thank everyone at nVidia for your wonderful software that now inexplicably crashes my computer anytime I go near it. BTW include that info in your "warning" about the ntune software...]]></description>
		<pubDate>Thu, 01 Oct 2009 20:55:06 -0700</pubDate>
		<guid>http://developer.nvidia.com/forums/index.php?showtopic=3685</guid>
	</item>
	<item>
		<title>vehicle pushed by character controllers</title>
		<link>http://developer.nvidia.com/forums/index.php?showtopic=3684</link>
		<description><![CDATA[I have a vehicle that uses NxWheelShapes.  Everything is working alright except this one scenario and I have been unable to determine why it is happening.  When the vehicle is in motion (not sleeping, not kinematic), the character controller is able to push it around.  This happens at extremely low velocities, and also when the vehicle is moving rather quickly. <br /><br />I am explicitly returning NX_ACTION_NONE in the onShapeHit callback in my controller.  The behavior is as expected for all other dynamic objects in my scene.  My character controller runs into them, and because I'm indicating to perform no actions, nothing happens, they remain unmoved.  It shouldn't even matter according to the docs, as they claim that the other option, NX_ACTION_PUSH isn't even implemented as of 2.8.1.<br /><br />I've also confirmed that the character bounding volumes (non-character controller volumes attached to characters arms/legs/torso/weapons etc.) are NOT colliding with the vehicle.  No contact reports are being generated, as I have verified in code and using visualizations in the VRD.<br /><br />It's as if actors using wheeled shapes ignore the NX_ACTION_NONE and somehow forces are being applied causing motor torque to be set into the wheels.<br /><br />Has anyone else seen a similar issue?  Anyone have ideas as to what might be happening?<br /><br />Thanks!]]></description>
		<pubDate>Thu, 01 Oct 2009 16:16:17 -0700</pubDate>
		<guid>http://developer.nvidia.com/forums/index.php?showtopic=3684</guid>
	</item>
	<item>
		<title>win7 32bit driver wont install</title>
		<link>http://developer.nvidia.com/forums/index.php?showtopic=3683</link>
		<description><![CDATA[Hello all i'm new,<br /><br />I have a problem with my driver's. They just won't install even not in safe mode in admin mode there is just no way to get it working. I'm running windows 7 32 bits and trying to install driver : 191.03 BETA. This problem happens with EVERY driver except microsoft's driver wich they suply for win7 wich is outdated now. I tried drive sweeper and many things but none helped. Also my nvidia-control panel doesn't show up, when I want to install it it says driver doesn't match. Any solution?]]></description>
		<pubDate>Thu, 01 Oct 2009 14:48:23 -0700</pubDate>
		<guid>http://developer.nvidia.com/forums/index.php?showtopic=3683</guid>
	</item>
	<item>
		<title>trouble converting psd to dds with alpha</title>
		<link>http://developer.nvidia.com/forums/index.php?showtopic=3682</link>
		<description><![CDATA[I'm working in siumulation.<br />Using photoshop and the dds plugin(downloaded from this site) when I save a file with alpha channel there is a lot of banding, ie there isn't a smooth gradient its very choppy. I've tried many different settings, but the issue remains.<br />I suspect there's a simple solution, but I just can't seem to find it on my own.<br />Thanks<br />Dave<br /><br />ps I've tried saving out as dxt1-5 with a variety of image option settings<br /><br />UpDATE: Using the lateset version of photoshop cs3 solved it.]]></description>
		<pubDate>Thu, 01 Oct 2009 13:44:36 -0700</pubDate>
		<guid>http://developer.nvidia.com/forums/index.php?showtopic=3682</guid>
	</item>
	<item>
		<title>help with psd to dds</title>
		<link>http://developer.nvidia.com/forums/index.php?showtopic=3681</link>
		<description><![CDATA[I'm working in siumulation.<br />Using photosho and the dds plugin when I save a file with alpha channel there is a lot of banding, ie there isn't a smooth gradient its very choppy. I've tried many different settings, but the issue remains.<br />I suspect there's a simple solution, but I just can't seem to find it on my own.<br />Thanks<br />Dave<br /><br />P.s. probably not the best place for the topic...oops]]></description>
		<pubDate>Thu, 01 Oct 2009 13:22:44 -0700</pubDate>
		<guid>http://developer.nvidia.com/forums/index.php?showtopic=3681</guid>
	</item>
	<item>
		<title>Dependent 1D texture lookup with cgc</title>
		<link>http://developer.nvidia.com/forums/index.php?showtopic=3680</link>
		<description><![CDATA[Hello everybody!<br /><br />I am quite new to Cg-programming, but I have searched the internet for a few hours now without a solution for my problem.<br />So, I have a very simple fragment program (I think):<br /><br /><!--c1--><div class='codetop'>CODE</div><div class='codemain'><!--ec1-->FragOutput main&#40;float2 texCoord &#58; TEXCOORD0,<br />&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;uniform sampler2D tex &#58; TEXTURE_2D,<br />&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;uniform sampler1D map &#58; TEXTURE_1D&#41;<br />{<br />&nbsp;&nbsp;FragOutput OUT;<br />&nbsp;&nbsp;OUT.color = tex2D&#40;tex,texCoord&#41;;<br />&nbsp;&nbsp;OUT.color = tex1D&#40;map, OUT.color.r&#41;;<br />&nbsp;&nbsp;return OUT;<br />}<!--c2--></div><!--ec2--><br /><br />tex is a 2D-grayscale image, map is a 1D-RGB image, which is supposed to be a lookup table, to have nice colors instead of the gray values on the screen.<br />Both textures are working quite fine on their own, but the dependent texture lookup results in a black screen.<br /><br />I already know the problem about this thing, I found exactly the same case here:<br /><a href="http://developer.nvidia.com/forums/index.php?showtopic=995" target="_blank">Dependent 1D Texture lookup (Found in this Forum)</a><br /><br />Here is my question: The solution in the mentioned thread was to call "cgGLEnableTextureParameter(lookup)". (I guess, the cause is some kind of race condition inside the pipeline.)<br />In my case, this is not possible. I am using Java and OpenGL, don't wonder about it, it is some science-stuff. Since there is no Cg-Runtime for Java, I am using the Standalone Cg-Compiler (cgc) with the arbfp1 profile. So, how can I adopt the given solution? I already found the profile option "NoDependentReadLimit" in the manual, but my cgc (Version 2.2.0006) does not support that option somehow ("warning C6504: Unknown profile option 'NoDependentReadLimit' ignored")<br /><br />I would be very happy, if anyone has an idea!<br /><br />Best Regards<br />Markus]]></description>
		<pubDate>Thu, 01 Oct 2009 12:31:51 -0700</pubDate>
		<guid>http://developer.nvidia.com/forums/index.php?showtopic=3680</guid>
	</item>
	<item>
		<title>Rendering/Restoring Z-Buffer to and from Texture</title>
		<link>http://developer.nvidia.com/forums/index.php?showtopic=3679</link>
		<description><![CDATA[I want to render a scene and it's depthstencil to textures, and then be able to render that static scene by drawing the textures. I am then trying to render these texures to restore  a static scene that I can move objects through. The scene draws fine, but the zbuffer does not.<br /><br /><br />Here is the code for creating the texures:<br /><br /><!--c1--><div class='codetop'>CODE</div><div class='codemain'><!--ec1-->hr = Gfx&#58;&#58;CreateTexture&#40;m_width, m_height, 1, D3DUSAGE_RENDERTARGET, D3DFMT_A8R8G8B8, D3DPOOL_DEFAULT, &m_renderTarget&#41;;<br />m_renderTarget-&#62;GetSurfaceLevel&#40;0, &m_renderSurface&#41;;<br /><br />hr = Gfx&#58;&#58;CreateTexture&#40;m_width, m_height,1, D3DUSAGE_DEPTHSTENCIL, D3DFMT_D32, D3DPOOL_DEFAULT,&m_depthTexture&#41;;<br />m_depthTexture-&#62;GetSurfaceLevel&#40;0,&m_depthStencil&#41;;<!--c2--></div><!--ec2--><br /><br />I push and set the rendertarget and depth stencil, render my scene, and then pop the rendertarget and depthstencil.<br /><br />I then set the rendertarget and depthstencil textures as inputs for a DetailSampler and ZSampler respectively, clear the rendertarget and zbuffer, then render a fullwindow quad using the following pixelshader:<br /><br /><!--c1--><div class='codetop'>CODE</div><div class='codemain'><!--ec1-->void zbufferPS&#40; in float2 Texture&nbsp;&nbsp;&nbsp;&nbsp;&#58; TEXCOORD0,<br />&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;out float4 color&#58; COLOR0,<br />&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;out float depth &#58; DEPTH0&#41;<br />{<br />&nbsp;&nbsp;&nbsp;&nbsp;color = tex2D&#40;DetailSampler, Texture&#41;;<br />&nbsp;&nbsp;&nbsp;&nbsp;depth = tex2D&#40;ZSampler, Texture&#41;.r;<br />}<!--c2--></div><!--ec2--><br /><br /><br />Long story short, the scene draws fine, but the zbuffer is empty. Is it possible to read from a depth stencil texture, and if not how can I get around this?  I am using a GFX 280 video card and DX9. Thanks.]]></description>
		<pubDate>Thu, 01 Oct 2009 11:46:06 -0700</pubDate>
		<guid>http://developer.nvidia.com/forums/index.php?showtopic=3679</guid>
	</item>
	<item>
		<title>how can i become licensed developer?</title>
		<link>http://developer.nvidia.com/forums/index.php?showtopic=3678</link>
		<description><![CDATA[Hi, how can i become licensed developer?<br /><br />I sent a mail to nv-PhysXDevSupport@supportcenteronline.com, got the mail back with<br /><br />"Your request to open a new customer account has been received.<br />We will notify you with your support login information as soon as your information is approved and your account is created"<br /><br />And another some months there is no more answers, even if my request has been denied i have to know about it.<br /><br />After SDK 2.8.3 availability i tryed to make it one more time September 28 with my another email.<br />I know it have to take (1-3 days) but still no answer.<br /><br />As i understand anyone can became licensed developer. Am i mistaken?<br /><br />Thanks.]]></description>
		<pubDate>Thu, 01 Oct 2009 07:13:02 -0700</pubDate>
		<guid>http://developer.nvidia.com/forums/index.php?showtopic=3678</guid>
	</item>
	<item>
		<title><![CDATA[[Feature request] Dynamic body shape change]]></title>
		<link>http://developer.nvidia.com/forums/index.php?showtopic=3677</link>
		<description><![CDATA[I was wondering if PhysX could apply some boolean operations on already created NxActor body shapes.<br /> This feature might be called like "destructable objects", "deformable rigid bodies" or smthng like that. For example, if we have a wall and someone shot this wall from rocket launcher, we just divide some pieces of wall away making 'em different actors falling down. Is it possible to add such feature to the closest PhysX release? It would be a great one.... <img src="http://developer.nvidia.com/forums/style_emoticons/default/huh.gif" style="vertical-align:middle" emoid=":huh:" border="0" alt="huh.gif" />]]></description>
		<pubDate>Thu, 01 Oct 2009 03:33:56 -0700</pubDate>
		<guid>http://developer.nvidia.com/forums/index.php?showtopic=3677</guid>
	</item>
	<item>
		<title>Forum threading issue</title>
		<link>http://developer.nvidia.com/forums/index.php?showtopic=3676</link>
		<description><![CDATA[This may not be the right place for this topic but since it's the PhysX forum I'm most interested in, I'm posting it here.<br /><br />It seems the forum has been changed so that individual replies all require a click to show them.<br />either..<br /><br />1. Is there a way to disable this in the member profile that I haven't seen?<br /><br />or<br /><br />2. Can this be changed back to the previous message format showing all the replies.<br /><br />]]></description>
		<pubDate>Wed, 30 Sep 2009 14:28:33 -0700</pubDate>
		<guid>http://developer.nvidia.com/forums/index.php?showtopic=3676</guid>
	</item>
</channel>
</rss>