Here's a tailored for my HP ZBook Studio G3 that utilizes the Intel HD Graphics 530 for display and the NVIDIA Quadro M2000M for compute on Rocky Linux 8:
Install Necessary Drivers:
- Ensure that you have the appropriate drivers for both GPUs installed on your system.
-
Install the Intel OpenCL Runtime:
- You need to install the Intel OpenCL runtime to enable OpenCL support on the Intel GPU.
Add the repository and install the runtime:
sudo tee /etc/yum.repos.d/intel-opencl.repo <<EOF [intel-opencl] name=Intel(R) Graphics Compute Runtime for OpenCL(TM) baseurl=https://repositories.intel.com/graphics/rpm/rhel8 enabled=1 gpgcheck=1 gpgkey=https://repositories.intel.com/graphics/intel-graphics.key EOF sudo dnf install intel-opencl-clang intel-ocloc intel-opencl
-
Install the DKU:
- Download the appropriate DKU version from the Autodesk website. https://www.autodesk.com/support/technical/article/caas/sfdcarticles/sfdcarticles/CentOS-ISO-installer-and-Linux-Driver-Kernel-Utilities-for-Flame-Family.html
Download the correct DKU from the system requirements web page, for example: Flame System Requirements and extract the contents of the file, for example: tar -xvf <DKU_file_name>.tar. In the directory where the DKU was extracted, run the following to install the DKU: ./INSTALL_DKU. For usage, type INSTALL_DKU --help. If a Wacom tablet or storage devices were disconnected or turned off as part of the OS installation, reconnect and power them up. Restart the workstation with reboot.
- Create or Edit Xorg Configuration File: Create or edit the Xorg configuration file to configure both GPUs.
sudo nano /etc/X11/xorg.conf.d/10-intel.conf
Add the following content:
Section "Device"
Identifier "IntelGPU"
Driver "intel"
BusID "PCI:0:2:0"
Option "AccelMethod" "sna"
Option "TearFree" "true"
Option "DRI" "3"
EndSection
Section "Screen"
Identifier "ScreenIntel"
Device "IntelGPU"
Monitor "Monitor0"
DefaultDepth 24
SubSection "Display"
Depth 24
Modes "1920x1080"
EndSubSection
EndSection
Section "Monitor"
Identifier "Monitor0"
Option "DPMS"
EndSection
sudo nano /etc/X11/xorg.conf.d/20-nvidia.conf
Add the following content:
Section "Device"
Identifier "NvidiaGPU"
Driver "nvidia"
BusID "PCI:1:0:0"
Option "AllowEmptyInitialConfiguration"
Option "Coolbits" "28"
EndSection
Section "Screen"
Identifier "ScreenNvidia"
Device "NvidiaGPU"
EndSection
Section "Module"
Load "glx"
EndSection
Section "Files"
ModulePath "/usr/lib64/nvidia/xorg"
ModulePath "/usr/lib64/xorg/modules"
EndSection
sudo nano /etc/X11/xorg.conf.d/30-layout.conf
Add the following content:
Section "ServerLayout"
Identifier "layout"
Screen 0 "ScreenIntel"
Inactive "ScreenNvidia"
EndSection
-
Save and Close the Files:
- Save each file after editing by pressing
Ctrl+X
, thenY
, andEnter
.
- Save each file after editing by pressing
-
Reboot the System:
sudo reboot
-
Verify Configuration:
This setup should ensure that Autodesk Flame and other applications can utilize the Intel HD Graphics 530 for display and the NVIDIA Quadro M2000M for compute tasks on Rocky Linux 8.
Autodesk Flame:
- Ensure that Flame is configured to use the Intel GPU for specific tasks. This may involve modifying configuration files or setting environment variables as specified in the Flame documentation.
Autodesk Maya:
- Maya can use multiple GPUs for different tasks. Ensure that the Intel GPU is available and recognized by Maya.
-
Balancing Load Between GPUs:
- You can use the Intel GPU for less intensive tasks or as a secondary compute device. For example, you might configure rendering tasks to use the NVIDIA GPU while using the Intel GPU for other computations.
Environment Variable:
export GPU_DEVICE=Intel # This is just an example; actual configuration may vary
-
Reboot and Verify:
- Reboot your system to ensure that all changes take effect.
sudo reboot
- Use
clinfo
again to verify that both GPUs are available and correctly configured.
By following these steps, you should have the Intel HD Graphics 530 configured for display and the NVIDIA Quadro M2000M configured for compute tasks on Rocky Linux 8. This setup allows you to utilize both GPUs effectively for different purposes.