Re: [osg-users] Volume image crash on the latest version of Nvidia driver
Hi Robert, Yes. I am using 3.2.3 since I got a problem to show image in the center on 3.4 (My program is using MFC). Then I used 3.2.3. Regards, Clement From: osg-users [osg-users-boun...@lists.openscenegraph.org] on behalf of Robert Osfield [robert.osfi...@gmail.com] Sent: Monday, 17 August 2015 17:06 To: OpenSceneGraph Users Subject: Re: [osg-users] Volume image crash on the latest version of Nvidia driver HI Clement, Curious... but what I don't understand is how the Intel compiler was progressing on to the if (num_iteration!=num_iterators) as this should only be run on NVidia cards thank to the #idef NVIDIA_Corportion guard around it, Are you backporting the shaders to 3,2,x? Robert. On 17 August 2015 at 07:33, clement@csiro.aumailto:clement@csiro.au wrote: Hi Robert, I got funny volume display problem on intel card (it lost some opacity values). I checked the shader code and I found the problem on handling NaN with using if statement. Your code is : if (num_iterationsmin_iterations) num_iterations = min_iterations;\n else if (num_iterationsmax_iterations) num_iterations = max_iterations;\n else if (num_iterations!=num_iterations) num_iterations = max_iterations;\n I checked NaN on wiki https://en.wikipedia.org/wiki/NaN. The non-signaling value won't do standard comparison, so I changed the code as below. if (num_iterations!=num_iterations) num_iterations = max_iterations;\n else if (num_iterationsmin_iterations) num_iterations = min_iterations;\n else if (num_iterationsmax_iterations) num_iterations = max_iterations;\n Check the value whether it is NaN on first condition and it works as normal on intel card now. Regards, Clement From: osg-users [osg-users-boun...@lists.openscenegraph.orgmailto:osg-users-boun...@lists.openscenegraph.org] on behalf of Robert Osfield [robert.osfi...@gmail.commailto:robert.osfi...@gmail.com] Sent: Wednesday, 12 August 2015 23:58 To: OpenSceneGraph Users Subject: Re: [osg-users] Volume image crash on the latest version of Nvidia driver On 12 August 2015 at 14:47, clement@csiro.aumailto:clement@csiro.aumailto:clement@csiro.au wrote: Hi Robert, I have tested on both Intel and Nvidia machines. The solution works without any issue. Thanks. I have applied this fix to OSG-svn/trunk, OSG-3.2.3 and OSG-3.4.0 so we should now be good to go. Robert. ___ osg-users mailing list osg-users@lists.openscenegraph.orgmailto:osg-users@lists.openscenegraph.org http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org ___ osg-users mailing list osg-users@lists.openscenegraph.org http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org
Re: [osg-users] Volume image crash on the latest version of Nvidia driver
Hi Robert, I got funny volume display problem on intel card (it lost some opacity values). I checked the shader code and I found the problem on handling NaN with using if statement. Your code is : if (num_iterationsmin_iterations) num_iterations = min_iterations;\n else if (num_iterationsmax_iterations) num_iterations = max_iterations;\n else if (num_iterations!=num_iterations) num_iterations = max_iterations;\n I checked NaN on wiki https://en.wikipedia.org/wiki/NaN. The non-signaling value won't do standard comparison, so I changed the code as below. if (num_iterations!=num_iterations) num_iterations = max_iterations;\n else if (num_iterationsmin_iterations) num_iterations = min_iterations;\n else if (num_iterationsmax_iterations) num_iterations = max_iterations;\n Check the value whether it is NaN on first condition and it works as normal on intel card now. Regards, Clement From: osg-users [osg-users-boun...@lists.openscenegraph.org] on behalf of Robert Osfield [robert.osfi...@gmail.com] Sent: Wednesday, 12 August 2015 23:58 To: OpenSceneGraph Users Subject: Re: [osg-users] Volume image crash on the latest version of Nvidia driver On 12 August 2015 at 14:47, clement@csiro.aumailto:clement@csiro.au wrote: Hi Robert, I have tested on both Intel and Nvidia machines. The solution works without any issue. Thanks. I have applied this fix to OSG-svn/trunk, OSG-3.2.3 and OSG-3.4.0 so we should now be good to go. Robert. ___ osg-users mailing list osg-users@lists.openscenegraph.org http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org
Re: [osg-users] Volume image crash on the latest version of Nvidia driver
HI Clement, Curious... but what I don't understand is how the Intel compiler was progressing on to the if (num_iteration!=num_iterators) as this should only be run on NVidia cards thank to the #idef NVIDIA_Corportion guard around it, Are you backporting the shaders to 3,2,x? Robert. On 17 August 2015 at 07:33, clement@csiro.au wrote: Hi Robert, I got funny volume display problem on intel card (it lost some opacity values). I checked the shader code and I found the problem on handling NaN with using if statement. Your code is : if (num_iterationsmin_iterations) num_iterations = min_iterations;\n else if (num_iterationsmax_iterations) num_iterations = max_iterations;\n else if (num_iterations!=num_iterations) num_iterations = max_iterations;\n I checked NaN on wiki https://en.wikipedia.org/wiki/NaN. The non-signaling value won't do standard comparison, so I changed the code as below. if (num_iterations!=num_iterations) num_iterations = max_iterations;\n else if (num_iterationsmin_iterations) num_iterations = min_iterations;\n else if (num_iterationsmax_iterations) num_iterations = max_iterations;\n Check the value whether it is NaN on first condition and it works as normal on intel card now. Regards, Clement From: osg-users [osg-users-boun...@lists.openscenegraph.org] on behalf of Robert Osfield [robert.osfi...@gmail.com] Sent: Wednesday, 12 August 2015 23:58 To: OpenSceneGraph Users Subject: Re: [osg-users] Volume image crash on the latest version of Nvidia driver On 12 August 2015 at 14:47, clement@csiro.aumailto: clement@csiro.au wrote: Hi Robert, I have tested on both Intel and Nvidia machines. The solution works without any issue. Thanks. I have applied this fix to OSG-svn/trunk, OSG-3.2.3 and OSG-3.4.0 so we should now be good to go. Robert. ___ osg-users mailing list osg-users@lists.openscenegraph.org http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org ___ osg-users mailing list osg-users@lists.openscenegraph.org http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org
Re: [osg-users] Volume image crash on the latest version of Nvidia driver
Hi Robert, I have tested on both Intel and Nvidia machines. The solution works without any issue. Thanks. Regards, Clement From: osg-users [osg-users-boun...@lists.openscenegraph.org] on behalf of Robert Osfield [robert.osfi...@gmail.com] Sent: Tuesday, 11 August 2015 18:07 To: OpenSceneGraph Users Subject: Re: [osg-users] Volume image crash on the latest version of Nvidia driver Hi Clement, On 11 August 2015 at 08:02, clement@csiro.au wrote: I found opengl api https://www.opengl.org/sdk/docs/man/. I think most of systems should support it. On my search I came across suggestion of lack of support on OSX and GLES. I don't know how correct/up to date this is. Could you try replacing the if (isnan(num_iterations)) num_iterations = 2.0; With: if (num_iterations!=num_iterations) num_iterations = 2.0; I would try myself by my system that had the latest NVIdia drivers that show this bug has died due to a disk failure. I will need to buy and new disk and get everything back installed on it. This isn't something I can tackle today though. Robert. ___ osg-users mailing list osg-users@lists.openscenegraph.org http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org ___ osg-users mailing list osg-users@lists.openscenegraph.org http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org
Re: [osg-users] Volume image crash on the latest version of Nvidia driver
On 12 August 2015 at 14:47, clement@csiro.au wrote: Hi Robert, I have tested on both Intel and Nvidia machines. The solution works without any issue. Thanks. I have applied this fix to OSG-svn/trunk, OSG-3.2.3 and OSG-3.4.0 so we should now be good to go. Robert. ___ osg-users mailing list osg-users@lists.openscenegraph.org http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org
Re: [osg-users] Volume image crash on the latest version of Nvidia driver
HI Clement, Good investigation. This suggested workaround would sure be cleaner than my approach of toggling between float and int types. The only problem with it would be how widely the isnan is supported. I've done a quick internet search and haven't come up with a clear answer, it would seem there are systems that don't support it. Robert. On 11 August 2015 at 06:51, clement@csiro.au wrote: Hi Robert, Finally, I found out where causes the crash problem. The problem is not related to the type casting. See the source code below: float num_iterations = ceil(length((te-t0).xyz)/SampleDensityValue); The problem is the value of num_iterations is a NaN float value (not a number), so that why it makes an infinite running in the loop and then crashed. I think the problem is on the method length(). In some condition (different values of t0 and te) makes the method length() to return a NaN value. I added a line of code to check variable num_iterations. if (isnan(num_iterations)) num_iterations = 2.0; So far it works on both nvidia and intel cards. Regards, Clement From: osg-users [osg-users-boun...@lists.openscenegraph.org] on behalf of Robert Osfield [robert.osfi...@gmail.com] Sent: Monday, 10 August 2015 17:42 To: OpenSceneGraph Users Subject: Re: [osg-users] Volume image crash on the latest version of Nvidia driver Hi Clement, I have been looking GLSL for what might help us determine when to use float vs ints. The follow info might be useful: https://github.com/mattdesl/lwjgl-basics/wiki/GLSL-Versions Perhaps we should just add a #version 130 (OpenGL Version = 3.0) to the top of the fragment shaders to make sure they support ints. This may mean that the Intel driver fails though. What Intel hardware, drivers and OpenGL version support by the drivers do you have on your system? Robert. ___ osg-users mailing list osg-users@lists.openscenegraph.org http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org ___ osg-users mailing list osg-users@lists.openscenegraph.org http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org
Re: [osg-users] Volume image crash on the latest version of Nvidia driver
Hi Robert, I found opengl api https://www.opengl.org/sdk/docs/man/. I think most of systems should support it. Regards, Clement From: osg-users [osg-users-boun...@lists.openscenegraph.org] on behalf of Robert Osfield [robert.osfi...@gmail.com] Sent: Tuesday, 11 August 2015 16:53 To: OpenSceneGraph Users Subject: Re: [osg-users] Volume image crash on the latest version of Nvidia driver HI Clement, Good investigation. This suggested workaround would sure be cleaner than my approach of toggling between float and int types. The only problem with it would be how widely the isnan is supported. I've done a quick internet search and haven't come up with a clear answer, it would seem there are systems that don't support it. Robert. On 11 August 2015 at 06:51, clement@csiro.aumailto:clement@csiro.au wrote: Hi Robert, Finally, I found out where causes the crash problem. The problem is not related to the type casting. See the source code below: float num_iterations = ceil(length((te-t0).xyz)/SampleDensityValue); The problem is the value of num_iterations is a NaN float value (not a number), so that why it makes an infinite running in the loop and then crashed. I think the problem is on the method length(). In some condition (different values of t0 and te) makes the method length() to return a NaN value. I added a line of code to check variable num_iterations. if (isnan(num_iterations)) num_iterations = 2.0; So far it works on both nvidia and intel cards. Regards, Clement From: osg-users [osg-users-boun...@lists.openscenegraph.orgmailto:osg-users-boun...@lists.openscenegraph.org] on behalf of Robert Osfield [robert.osfi...@gmail.commailto:robert.osfi...@gmail.com] Sent: Monday, 10 August 2015 17:42 To: OpenSceneGraph Users Subject: Re: [osg-users] Volume image crash on the latest version of Nvidia driver Hi Clement, I have been looking GLSL for what might help us determine when to use float vs ints. The follow info might be useful: https://github.com/mattdesl/lwjgl-basics/wiki/GLSL-Versions Perhaps we should just add a #version 130 (OpenGL Version = 3.0) to the top of the fragment shaders to make sure they support ints. This may mean that the Intel driver fails though. What Intel hardware, drivers and OpenGL version support by the drivers do you have on your system? Robert. ___ osg-users mailing list osg-users@lists.openscenegraph.orgmailto:osg-users@lists.openscenegraph.org http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org ___ osg-users mailing list osg-users@lists.openscenegraph.org http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org
Re: [osg-users] Volume image crash on the latest version of Nvidia driver
Hi Clement, On 11 August 2015 at 08:02, clement@csiro.au wrote: I found opengl api https://www.opengl.org/sdk/docs/man/. I think most of systems should support it. On my search I came across suggestion of lack of support on OSX and GLES. I don't know how correct/up to date this is. Could you try replacing the if (isnan(num_iterations)) num_iterations = 2.0; With: if (num_iterations!=num_iterations) num_iterations = 2.0; I would try myself by my system that had the latest NVIdia drivers that show this bug has died due to a disk failure. I will need to buy and new disk and get everything back installed on it. This isn't something I can tackle today though. Robert. ___ osg-users mailing list osg-users@lists.openscenegraph.org http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org
Re: [osg-users] Volume image crash on the latest version of Nvidia driver
Hi Robert, Finally, I found out where causes the crash problem. The problem is not related to the type casting. See the source code below: float num_iterations = ceil(length((te-t0).xyz)/SampleDensityValue); The problem is the value of num_iterations is a NaN float value (not a number), so that why it makes an infinite running in the loop and then crashed. I think the problem is on the method length(). In some condition (different values of t0 and te) makes the method length() to return a NaN value. I added a line of code to check variable num_iterations. if (isnan(num_iterations)) num_iterations = 2.0; So far it works on both nvidia and intel cards. Regards, Clement From: osg-users [osg-users-boun...@lists.openscenegraph.org] on behalf of Robert Osfield [robert.osfi...@gmail.com] Sent: Monday, 10 August 2015 17:42 To: OpenSceneGraph Users Subject: Re: [osg-users] Volume image crash on the latest version of Nvidia driver Hi Clement, I have been looking GLSL for what might help us determine when to use float vs ints. The follow info might be useful: https://github.com/mattdesl/lwjgl-basics/wiki/GLSL-Versions Perhaps we should just add a #version 130 (OpenGL Version = 3.0) to the top of the fragment shaders to make sure they support ints. This may mean that the Intel driver fails though. What Intel hardware, drivers and OpenGL version support by the drivers do you have on your system? Robert. ___ osg-users mailing list osg-users@lists.openscenegraph.org http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org
Re: [osg-users] Volume image crash on the latest version of Nvidia driver
Hi Clement, I have been looking GLSL for what might help us determine when to use float vs ints. The follow info might be useful: https://github.com/mattdesl/lwjgl-basics/wiki/GLSL-Versions Perhaps we should just add a #version 130 (OpenGL Version = 3.0) to the top of the fragment shaders to make sure they support ints. This may mean that the Intel driver fails though. What Intel hardware, drivers and OpenGL version support by the drivers do you have on your system? Robert. ___ osg-users mailing list osg-users@lists.openscenegraph.org http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org
Re: [osg-users] Volume image crash on the latest version of Nvidia driver
On 10 August 2015 at 05:03, Robert Osfield robert.osfi...@gmail.com wrote: Hi Clement, I don't have an Intel system here to try. Could you experiment with the shaders to see if you can come up with a combination which works on both NVidia and Intel? For OSG-3.4 and svn/trunk one might be able to use #pragma(tic) shader composition to choose between float vs int for the loop iterator. Then the application could select between the two at runtime. OSG-3.2 branch we are probably stuck choosing between working on NVidia or Intel. Unless we can get the render string in GLSL. Robert. ___ osg-users mailing list osg-users@lists.openscenegraph.org http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org
Re: [osg-users] Volume image crash on the latest version of Nvidia driver
Hi Clement, I don't have an Intel system here to try. Could you experiment with the shaders to see if you can come up with a combination which works on both NVidia and Intel? Cheers, Robert. On 10 August 2015 at 02:39, clement@csiro.au wrote: Hi Robert, I have tested again on some machines with Intel graphic cards. The new change works on even both old and new nvidia drivers, but it does not work on Intel card. There is no image showed on display with using Intel card. You can reproduce the problem with the same test file on Intel graphic card. Does any method make both Intel and Nvidia drivers working properly? Thanks. *Regards,* *Clement* -- *From:* osg-users [osg-users-boun...@lists.openscenegraph.org] on behalf of Robert Osfield [robert.osfi...@gmail.com] *Sent:* Thursday, 6 August 2015 21:32 *To:* OpenSceneGraph Users *Subject:* Re: [osg-users] Volume image crash on the latest version of Nvidia driver Hi Clement, I have now ported the RayTracedTechnique shaders all across to use an int variable for the sampling loop and this fixes the hang when running with your model. I have checked this fix into OSG-svn/trunk, 3.4 branch and 3.2 branch and OpenSceneGraph-Data/trunk. This fix will required a OSG-3.2.3 release. Robert. ___ osg-users mailing list osg-users@lists.openscenegraph.org http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org ___ osg-users mailing list osg-users@lists.openscenegraph.org http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org
Re: [osg-users] Volume image crash on the latest version of Nvidia driver
Hi Robert, I have tested again on some machines with Intel graphic cards. The new change works on even both old and new nvidia drivers, but it does not work on Intel card. There is no image showed on display with using Intel card. You can reproduce the problem with the same test file on Intel graphic card. Does any method make both Intel and Nvidia drivers working properly? Thanks. Regards, Clement From: osg-users [osg-users-boun...@lists.openscenegraph.org] on behalf of Robert Osfield [robert.osfi...@gmail.com] Sent: Thursday, 6 August 2015 21:32 To: OpenSceneGraph Users Subject: Re: [osg-users] Volume image crash on the latest version of Nvidia driver Hi Clement, I have now ported the RayTracedTechnique shaders all across to use an int variable for the sampling loop and this fixes the hang when running with your model. I have checked this fix into OSG-svn/trunk, 3.4 branch and 3.2 branch and OpenSceneGraph-Data/trunk. This fix will required a OSG-3.2.3 release. Robert. ___ osg-users mailing list osg-users@lists.openscenegraph.org http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org
Re: [osg-users] Volume image crash on the latest version of Nvidia driver
Hi Robert, Thanks for your help. The solution fixed the problem. Thanks. Regards, Clement From: osg-users [osg-users-boun...@lists.openscenegraph.org] on behalf of Robert Osfield [robert.osfi...@gmail.com] Sent: Thursday, 6 August 2015 21:32 To: OpenSceneGraph Users Subject: Re: [osg-users] Volume image crash on the latest version of Nvidia driver Hi Clement, I have now ported the RayTracedTechnique shaders all across to use an int variable for the sampling loop and this fixes the hang when running with your model. I have checked this fix into OSG-svn/trunk, 3.4 branch and 3.2 branch and OpenSceneGraph-Data/trunk. This fix will required a OSG-3.2.3 release. Robert. ___ osg-users mailing list osg-users@lists.openscenegraph.org http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org
Re: [osg-users] Volume image crash on the latest version of Nvidia driver
Hi Clement, I have now ported the RayTracedTechnique shaders all across to use an int variable for the sampling loop and this fixes the hang when running with your model. I have checked this fix into OSG-svn/trunk, 3.4 branch and 3.2 branch and OpenSceneGraph-Data/trunk. This fix will required a OSG-3.2.3 release. Robert. ___ osg-users mailing list osg-users@lists.openscenegraph.org http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org
Re: [osg-users] Volume image crash on the latest version of Nvidia driver
Hi Clement and Julien, On 5 August 2015 at 02:12, clement@csiro.au wrote: Hi Julien, I am using osg stable release 3.2.1 64 bit windows version. Did you solve the problem so far? Can you create the problem with any of the standard OSG examples? If not could you create a small example that illustrates the problem. Trying out one of the OSG-3.4.0-rc as well as the OpenSceneGraph-Data in svn/trunk. I'll make a new OpenSceneGraph-Data for OSG-3.4.0. Robert. ___ osg-users mailing list osg-users@lists.openscenegraph.org http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org
Re: [osg-users] Volume image crash on the latest version of Nvidia driver
Hi Robert, The problem may be related to Nvidia driver. I tested on my laptop (Nvidia NVS 5200M) with driver 353.62 and deskop (Nvidia Quadro 2000) with driver 353.62. Both machines have the crash problem. I also tested the crash problem starting from the driver version 343. My application is using window MFC. There is no crash when the volume data is loaded. Once I started to rotate the image, it will crash the whole application. I also exported my volume data into dds format using (osg write image method). When I used osgviewer.exe to load the dds image, the crash problem also occurred if I rotated the image. I cannot run the osg volume data sample, but there is no crash problem with using others osg example data. I will try OSG-3.4.0-rc later. Regards, Clement From: osg-users [osg-users-boun...@lists.openscenegraph.org] on behalf of Robert Osfield [robert.osfi...@gmail.com] Sent: Wednesday, 5 August 2015 16:52 To: OpenSceneGraph Users Subject: Re: [osg-users] Volume image crash on the latest version of Nvidia driver Hi Clement and Julien, On 5 August 2015 at 02:12, clement@csiro.aumailto:clement@csiro.au wrote: Hi Julien, I am using osg stable release 3.2.1 64 bit windows version. Did you solve the problem so far? Can you create the problem with any of the standard OSG examples? If not could you create a small example that illustrates the problem. Trying out one of the OSG-3.4.0-rc as well as the OpenSceneGraph-Data in svn/trunk. I'll make a new OpenSceneGraph-Data for OSG-3.4.0. Robert. ___ osg-users mailing list osg-users@lists.openscenegraph.org http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org
Re: [osg-users] Volume image crash on the latest version of Nvidia driver
Hi Clement, Could you post the data and command line options with osgvolume that illustrate the problem so I can test here. Cheers, Robert. On 5 August 2015 at 13:36, clement@csiro.au wrote: Hi Robert, The problem may be related to Nvidia driver. I tested on my laptop (Nvidia NVS 5200M) with driver 353.62 and deskop (Nvidia Quadro 2000) with driver 353.62. Both machines have the crash problem. I also tested the crash problem starting from the driver version 343. My application is using window MFC. There is no crash when the volume data is loaded. Once I started to rotate the image, it will crash the whole application. I also exported my volume data into dds format using (osg write image method). When I used osgviewer.exe to load the dds image, the crash problem also occurred if I rotated the image. I cannot run the osg volume data sample, but there is no crash problem with using others osg example data. I will try OSG-3.4.0-rc later. *Regards,* *Clement* -- *From:* osg-users [osg-users-boun...@lists.openscenegraph.org] on behalf of Robert Osfield [robert.osfi...@gmail.com] *Sent:* Wednesday, 5 August 2015 16:52 *To:* OpenSceneGraph Users *Subject:* Re: [osg-users] Volume image crash on the latest version of Nvidia driver Hi Clement and Julien, On 5 August 2015 at 02:12, clement@csiro.au wrote: Hi Julien, I am using osg stable release 3.2.1 64 bit windows version. Did you solve the problem so far? Can you create the problem with any of the standard OSG examples? If not could you create a small example that illustrates the problem. Trying out one of the OSG-3.4.0-rc as well as the OpenSceneGraph-Data in svn/trunk. I'll make a new OpenSceneGraph-Data for OSG-3.4.0. Robert. ___ osg-users mailing list osg-users@lists.openscenegraph.org http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org ___ osg-users mailing list osg-users@lists.openscenegraph.org http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org
Re: [osg-users] Volume image crash on the latest version of Nvidia driver
Hi Clement, On 5 August 2015 at 17:06, clement@csiro.au wrote: Thanks for your help to test the problem. Please find the attachment of my test data. I tested with my laptop (Nvidia NVS 5200M with driver 353.62) and used osg 3.2.1 x64 windows version. The command is: osgviewer t1.osg Thanks for the model. I've run it on my Kubuntu system with Geforce GTX 760 + 346.59 drives and get a hang on my system when I rotate the view. This happens with OSG svn/trunk (equivalent to OSG-3.4). When I run osgvolume t1.dds I get the hang as well. Given the nature of the hang and how it locks out the whole desktop it all points to GPU/driver issue of some kind. If I run: osgvolume t1.dds --multi-pass I don't get any hang. The osgVolume::MultiPassTechnique is only part of OSG-3.4 so you won't have this option when using OSG3.2.1. By default osgvolume still defaults to the old RayTracedTechnique that you'll be familiar with, so I would suspect an issue with how the GPU is managing the RayTracedTechnique GLSL shaders. The thread that Julian Valentin linked discusses potentially the same issue, although the GLSL code Julian suggesting could be changed looks valid and the fix a workaround to otherwise valid code. I will do a review of the shaders to see I can spot a problem. Robert. ___ osg-users mailing list osg-users@lists.openscenegraph.org http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org
Re: [osg-users] Volume image crash on the latest version of Nvidia driver
Hi Clement, I have tried Julian's suggested approach of assign the num_iterations to an int and it works, simplify his original suggestion I get: Original shader line that results in hang: float num_iterations = ceil(length((te-t0).xyz)/SampleDensityValue); Revised shader line that works fine: int num_iterations = ceil(length((te-t0).xyz)/SampleDensityValue); Now, these both should result in the same value, as ceil rounds to an integer value, abeit within the scope of floating point math. Curiously the modern MultiPassTechnique shaders work fine too and they use an int. Originally I didn't use an int for the num_iterations as early GLSL implementations didn't fully support int's. I think the best solution would be to modify the shaders to all use int's for the num_iterations as a general coding approach using int's is sound. It shouldn't however be need a float should work in this context just as well as a int. What I suspect is that at certain values there is some kind of numerical bug on the GPU or perhaps driver that is resulting in an invalid float point value. Casting from float to int seems to nip this error in the bud so it doesn't propagate. It's end of day here so I'll have to make the changes and test them tomorrow and will update OpenSceneGraph-Data and OSG-3.2, OSG-3.4 an OSG-svn/trunk branches. Robert. Robert. On 5 August 2015 at 17:46, Robert Osfield robert.osfi...@gmail.com wrote: Hi Clement, On 5 August 2015 at 17:06, clement@csiro.au wrote: Thanks for your help to test the problem. Please find the attachment of my test data. I tested with my laptop (Nvidia NVS 5200M with driver 353.62) and used osg 3.2.1 x64 windows version. The command is: osgviewer t1.osg Thanks for the model. I've run it on my Kubuntu system with Geforce GTX 760 + 346.59 drives and get a hang on my system when I rotate the view. This happens with OSG svn/trunk (equivalent to OSG-3.4). When I run osgvolume t1.dds I get the hang as well. Given the nature of the hang and how it locks out the whole desktop it all points to GPU/driver issue of some kind. If I run: osgvolume t1.dds --multi-pass I don't get any hang. The osgVolume::MultiPassTechnique is only part of OSG-3.4 so you won't have this option when using OSG3.2.1. By default osgvolume still defaults to the old RayTracedTechnique that you'll be familiar with, so I would suspect an issue with how the GPU is managing the RayTracedTechnique GLSL shaders. The thread that Julian Valentin linked discusses potentially the same issue, although the GLSL code Julian suggesting could be changed looks valid and the fix a workaround to otherwise valid code. I will do a review of the shaders to see I can spot a problem. Robert. ___ osg-users mailing list osg-users@lists.openscenegraph.org http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org
[osg-users] Volume image crash on the latest version of Nvidia driver
Hi, I got a crash problem on volume rendering on Windows with using the latest Nvidia driver which is 353.62. Previously I am using version 341 without any problem. Have anyone got the same problem? Regards, Clement ___ osg-users mailing list osg-users@lists.openscenegraph.org http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org
Re: [osg-users] Volume image crash on the latest version of Nvidia driver
i had the same issue: http://forum.openscenegraph.org/viewtopic.php?t=14304 -- Read this topic online here: http://forum.openscenegraph.org/viewtopic.php?p=64619#64619 ___ osg-users mailing list osg-users@lists.openscenegraph.org http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org
Re: [osg-users] Volume image crash on the latest version of Nvidia driver
Hi Julien, I am using osg stable release 3.2.1 64 bit windows version. Did you solve the problem so far? Regards, Clement From: osg-users [osg-users-boun...@lists.openscenegraph.org] on behalf of Julien Valentin [julienvalenti...@gmail.com] Sent: Wednesday, 5 August 2015 09:33 To: osg-users@lists.openscenegraph.org Subject: Re: [osg-users] Volume image crash on the latest version of Nvidia driver i had the same issue: http://forum.openscenegraph.org/viewtopic.php?t=14304 -- Read this topic online here: http://forum.openscenegraph.org/viewtopic.php?p=64619#64619 ___ osg-users mailing list osg-users@lists.openscenegraph.org http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org ___ osg-users mailing list osg-users@lists.openscenegraph.org http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org