ChimeraX on Rocky 9

[martin.juhas@…]

Hi,

I am having troubles running chimeraX on my Rocky9 PC. The installation runs fine but upon starting the program, I see only black GUI window. Can you help me please?
The machine is accessed remotely through VNC protocol, if that is important.

I tried the latest stable (1.7.1) as well as RC version, same problem in both.

Best,

Martin

[Greg Couch]

This sounds like a bug in the graphics driver. Please send use the output from: "lspci | grep VGA" and "glxinfo | grep -E 'OpenGL (vendor|renderer|version)'". The glxinfo program in the glx-utils package, so "sudo dnf install glx-utils" if needed.

[martin.juhas@…]

Hi, requested info below. Your command returned nothing, so I grep only OpenGL info. I hope this is OK...

$ lspci | grep VGA
03:00.0 VGA compatible controller: ASPEED Technology, Inc. ASPEED Graphics Family (rev 41)
d8:00.0 VGA compatible controller: NVIDIA Corporation TU102 [GeForce RTX 2080 Ti] (rev a1)

$ glxinfo | grep -E 'OpenGL'
OpenGL vendor string: Mesa
OpenGL renderer string: llvmpipe (LLVM 17.0.6, 256 bits)
OpenGL core profile version string: 4.5 (Core Profile) Mesa 23.3.3
OpenGL core profile shading language version string: 4.50
OpenGL core profile context flags: (none)
OpenGL core profile profile mask: core profile
OpenGL core profile extensions:
OpenGL version string: 4.5 (Compatibility Profile) Mesa 23.3.3
OpenGL shading language version string: 4.50
OpenGL context flags: (none)
OpenGL profile mask: compatibility profile
OpenGL extensions:
OpenGL ES profile version string: OpenGL ES 3.2 Mesa 23.3.3
OpenGL ES profile shading language version string: OpenGL ES GLSL ES 3.20
OpenGL ES profile extensions:

Thank you for the help...

MJ

[Greg Couch]

Assuming you have your display connected to the NVidia graphics card, you need to install the proprietary NVidia driver for ChimeraX to work. The open source NVidia driver isn't good enough yet.

See ​https://forums.rockylinux.org/t/nvidia-drivers-on-rocky-linux/12366 for various ways it can be done.

[martin.juhas@…]

Thank you for the help.

I found the issue. I did not have the cuda module loaded that is why it was not working. Once I loaded the cuda module to the environment it works fine.

Now I need to figure out how to load the cuda module in a way that even when running chimerx from the Application (with the icon from the menu), i tis also loaded. Any advice?

Best,

MJ

[Greg Couch]

How are you loading the cuda module, and why isn't it always present?

I would have expect that you would have picked one of the methods in ​https://forums.rockylinux.org/t/nvidia-drivers-on-rocky-linux/12366 to install the NVidia driver. On an Alma Linux 9 system I have access to, it looks like they picked the rpmfusion option. And I see the following Nvidia packages installed:

$ rpm -qa | grep -i nvidia
kmod-nvidia-5.14.0-362.el9_3-545.29.06-1.el9.x86_64
xorg-x11-drv-nvidia-libs-550.78-1.el9.x86_64
xorg-x11-drv-nvidia-kmodsrc-550.78-1.el9.x86_64
xorg-x11-drv-nvidia-cuda-libs-550.78-1.el9.x86_64
nvidia-modprobe-550.78-1.el9.x86_64
nvidia-settings-550.78-1.el9.x86_64
xorg-x11-drv-nvidia-power-550.78-1.el9.x86_64
xorg-x11-drv-nvidia-550.78-1.el9.x86_64
kmod-nvidia-5.14.0-362.el9_3-550.78-1.el9.x86_64
akmod-nvidia-550.78-1.el9.x86_64

[martin.juhas@…]

Hi,

I installed the drivers based on the NVidia webpage instructions, which is the same as in the link you sent below (​https://forums.rockylinux.org/t/nvidia-drivers-on-rocky-linux/12366#using-nvidias-official-repos-10).

In my case, I see this:
$ rpm -qa | grep -i nvidia
nvidia-driver-cuda-libs-550.54.15-1.el9.x86_64
nvidia-libXNVCtrl-550.54.15-1.el9.x86_64
nvidia-driver-NVML-550.54.15-1.el9.x86_64
nvidia-driver-NvFBCOpenGL-550.54.15-1.el9.x86_64
nvidia-libXNVCtrl-devel-550.54.15-1.el9.x86_64
nvidia-persistenced-550.54.15-1.el9.x86_64
nvidia-driver-cuda-550.54.15-1.el9.x86_64
nvidia-driver-libs-550.54.15-1.el9.x86_64
nvidia-driver-devel-550.54.15-1.el9.x86_64
kmod-nvidia-latest-dkms-550.54.15-1.el9.x86_64
nvidia-kmod-common-550.54.15-1.el9.noarch
nvidia-driver-550.54.15-1.el9.x86_64
nvidia-modprobe-550.54.15-1.el9.x86_64
nvidia-settings-550.54.15-1.el9.x86_64
nvidia-xconfig-550.54.15-1.el9.x86_64
nvidia-fs-dkms-2.19.7-1.x86_64
nvidia-fs-2.19.7-1.x86_64
nvidia-gds-12-4-12.4.1-1.x86_64
nvidia-gds-12.4.1-1.x86_64
dnf-plugin-nvidia-2.0-2.el9.noarch

I am using lmod modules (module add cuda in this case) due to previous problems with cuda dependencies in Amber MD package. So the paths to cuda are not available in the environment unless module add cuda is called. I never noticed the issues before since GPU was recognized by the system and everything seemed to work fine. Only in chimerax, I needed to explicitly also add cuda for it to load.

I guess if cuda was not loaded, but persitently ad dtto the environment, as usual installation would, this would not have happened.

MJ

[Greg Couch]

ChimeraX doesn't need cuda, it needs OpenGL. The glxinfo output should have said Nvidia instead of Mesa. Not sure what's wrong in your setup. Make sure you have the libglvnd package installed.

[martin.juhas@…]

I see, I read about it in NVidia forums now. It would probably require a bit more configuration. Now it seems to work when run from command line somehow, that is enough. Thank you.

MJ

-----Original Message-----
From: ChimeraX <ChimeraX-bugs-admin@cgl.ucsf.edu> 
Sent: Tuesday, May 21, 2024 10:28 AM
To: Juhás Martin <martin.juhas@uhk.cz>; gregc@cgl.ucsf.edu
Cc: goddard@cgl.ucsf.edu
Subject: Re: [ChimeraX] #15247: ChimeraX on Rocky 9

#15247: ChimeraX on Rocky 9
-------------------------------------+------------------------
          Reporter:  martin.juhas@\u2026  |      Owner:  Greg Couch
              Type:  defect          |     Status:  closed
          Priority:  normal          |  Milestone:
         Component:  Graphics        |    Version:
        Resolution:  limitation      |   Keywords:
        Blocked By:                  |   Blocking:
Notify when closed:                  |   Platform:  all
           Project:  ChimeraX        |
-------------------------------------+------------------------
Comment (by Greg Couch):

 ChimeraX doesn't need cuda, it needs OpenGL.  The glxinfo output should  have said Nvidia instead of Mesa.  Not sure what's wrong in your setup.
 Make sure you have the libglvnd package installed.
--
Ticket URL: <https://www.rbvi.ucsf.edu/trac/ChimeraX/ticket/15247#comment:8>
ChimeraX <https://www.rbvi.ucsf.edu/chimerax/>
ChimeraX Issue Tracker