VirtualGL's purpose is to add hardware-accelerated 3D capabilities to a remote display environment. Trying to use it with Mesa is completely missing the point. Furthermore, there are technical reasons why using Mesa as a back end for VirtualGL doesn't work and won't ever work. VirtualGL supports only the nVidia and ATI proprietary drivers as a back end. We attempted to qualify the Intel DRI drivers for use with VirtualGL but found those drivers to be very unstable.
From your message below, you are getting a segfault when running glxinfo on the local display without vglrun. Thus, VirtualGL is not involved, and the segfault is not our fault. It is the fault of your display drivers and is thus out of context for this list. This article: http://www.virtualgl.org/Documentation/Mesa explains how to get Mesa running in TurboVNC. You do not use VirtualGL at all in that case. You do a custom build of Mesa, enabling its X11 driver, then you run your OpenGL application directly and configure LD_LIBRARY_PATH to use your custom build of Mesa instead of the system-supplied OpenGL implementation. On 11/11/13 7:05 AM, Fernando Alvarez wrote: > Hello, > > I am trying to get Centos 6.4 (amd64) + VirtualGL 2.3.3 + TurboVNC > working with software OpenGL rendering, in order to test some simple > OpenGL applications. > > I have deployed a simple PC-to-PC initial setup with CentOS 6.4, and I > have succesfully installed VirtualGL. When I run the sanity check with > the following command: > > /opt/VirtualGL/bin/glxinfo -display :0 -c > > I get the following output the 70% of the time: > > name of display: :0 > Segmentation fault > > The remaining 30% of the time, I get the expected output: > > # /opt/VirtualGL/bin/glxinfo -display :0 -c | grep render > direct rendering: Yes > OpenGL renderer string: Mesa DRI Intel(R) Sandybridge Desktop > GL_NV_conditional_render, GL_NV_depth_clamp, GL_NV_light_max_exponent, > > After that, I have installed the latest Mesa source (9.2.2) and > TurboVNC, and I got the VNC server running as expected. But when I > run ./vglrun glxgears on the VNC session, I got another "Segmentation > Fault" error, and I can see the glxgears window for less than half a > second. At the same time, the physical display on the PC gets black with > a garbaged still image of the three gears. Also, the turboVNC session > continues working. > > Could anyone give me any advice on this issue? Is there anything I can > do to get glxgears running with software rendering on CentOS? Should I > install 32-bit libraries? > > Thank you in advance, > > *Fernando* > > > ------------------------------------------------------------------------------ > November Webinars for C, C++, Fortran Developers > Accelerate application performance with scalable programming models. Explore > techniques for threading, error checking, porting, and tuning. Get the most > from the latest Intel processors and coprocessors. See abstracts and register > http://pubads.g.doubleclick.net/gampad/clk?id=60136231&iu=/4140/ostg.clktrk > > > > _______________________________________________ > VirtualGL-Users mailing list > VirtualGL-Users@lists.sourceforge.net > https://lists.sourceforge.net/lists/listinfo/virtualgl-users > ------------------------------------------------------------------------------ November Webinars for C, C++, Fortran Developers Accelerate application performance with scalable programming models. Explore techniques for threading, error checking, porting, and tuning. Get the most from the latest Intel processors and coprocessors. See abstracts and register http://pubads.g.doubleclick.net/gampad/clk?id=60136231&iu=/4140/ostg.clktrk _______________________________________________ VirtualGL-Users mailing list VirtualGL-Users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/virtualgl-users
