Nvidia Optimus is an optimization technology created by Nvidia which, depending on the resource load generated by client software applications, will transparently and seamlessly switch between two graphics adapters within a computer system in order to provide either maximum performance or minimum power draw from the system's graphics rendering hardware. From Bumblebee's FAQ: Bumblebee is a effort to make Nvidia Optimus enabled laptops work in GNU/Linux systems. Such feature involves two graphics cards with two different power consumption profiles plugged in a layered way sharing a single framebuffer.
#yum install -y libbsd-devel libbsd glibc-devel libX11-devel help2man autoconf git tar glib2 glib2-devel kernel-devel kernel-headers automake gcc gtk2-devel
#yum install VirtualGL
If you want to run 32 bit apps on your 64 bit system, then also install:
#yum install VirtualGL.i686
Install the Bumblebee yum repository
#yum -y --nogpgcheck install http://install.linux.ncsu.edu/pub/yum/itecs/public/bumblebee/fedora20/noarch/bumblebee-release-1.1-1.noarch.rpm
#yum -y install http://install.linux.ncsu.edu/pub/yum/itecs/public/bumblebee/fedora21/noarch/bumblebee-release-1.2-1.noarch.rpm
#yum -y install bbswitch bumblebee
Bumblebee with the NVIDIA proprietary drivers
The previous commands install bumbleblee with the nouveau drivers. If you want to install bumblebee with the NVIDIA proprietary drivers you must also follow these instructions:
Install the bumblebee-nonfree repository (the bumblebee-free repository is needed too):
#yum -y --nogpgcheck install http://install.linux.ncsu.edu/pub/yum/itecs/public/bumblebee-nonfree/fedora20/noarch/bumblebee-nonfree-release-1.1-1.noarch.rpm
A potentially more up-to-date driver may be installed using the bumblebee-nonfree-shortlived repo:
#yum -y --nogpgcheck install http://install.linux.ncsu.edu/pub/yum/itecs/public/bumblebee-nonfree-shortlived/fedora20/noarch/bumblebee-nonfree-shortlived-release-1.1-1.noarch.rpm
#yum -y install http://install.linux.ncsu.edu/pub/yum/itecs/public/bumblebee-nonfree/fedora21/noarch/bumblebee-nonfree-release-1.2-1.noarch.rpm
Now install bumblebee-nvidia
#yum -y install bumblebee-nvidia
"bumblebee-nvidia" is both a rpm package and a shell script. The "bumblebee-nvidia" package contains the shell "wrapper" script, init scripts to run it at bootup, and a SELinux security policy, and lastly a copy of the NVidia Linux binary driver "blob" downloaded from NVidia's UNIX drivers page. This is why the packages are so large compared to the bumblebee package.
Make your bumblebee works.
To verify you are now using the NVIDIA driver, run the glxgears demo (from mesa-demos):
optirun glxgears -info | grep "GL_VENDOR"
You should see some gears rotating. In the terminal you should see "GL_VENDOR = NVIDIA Corporation" or something similar.
For speed imrovements you can use primus. Primus is a shared library that provides OpenGL and GLX APIs and implements low-overhead local-only client-side OpenGL offloading via GLX forking, similar to VirtualGL. It intercepts GLX calls and redirects GL rendering to a secondary X display, presumably driven by a faster GPU. On swapping buffers, rendered contents are read back using a PBO and copied onto the drawable it was supposed to be rendered on in the first place. Currently primus requires the NVIDIA proprietary drivers.
#yum -y install primus
If you want to run 32 bit apps on your 64 bit system, then also install:
#yum -y install primus.i686
Open up a terminal and run:
PRIMUS_VERBOSE=2 optirun -b primus glxgears -info
If primus is working properly, you should see "primus: profiling:" messages every so often in the terminal you ran it from.
How to use Bumblebee
optirun [options] <application> [application-parameters]
If you installed primus, then you can use the following command to take advantage of it:
optirun -b primus <application> [application-parameters]
Start Windows applications:
optirun wine <windows application>.exe
optirun -b primus wine <windows application>.exe
By default, primus renders up to 60 fps. To render more frames per second:
vblank_mode=0 optirun -b primus <application> [application-parameters]
In practice this will probably only waste power, as most screens do not display more than 60 frames per second.
You can always run "man optirun" for more information.
Unable to use optirun with nouveau
In some cases, users on post Linux 3.4 kernels are unable to use bumblebee with nouveau. One less then optimal workaround is to not use optirun and instead request that the window be drawn using the primary or secondary graphics cards/drivers. This can be achieved using the following:
[user1@localhost ~]$ DRI_PRIME=0 glxgears -info | grep "GL_VENDOR" [user1@localhost ~]$ DRI_PRIME=1 glxgears -info | grep "GL_VENDOR"
These two commands should produce the demonstration gears and the driver vendor (often Intel and nouveau) in the terminal. For more information see (https://bugzilla.redhat.com/show_bug.cgi?id=964012).
Unable to load (NVIDIA) GPU driver error with optirun
Most of the information presented here was suggested by bumblebee maintainer Gary Gatling.
The verbose output of optirun provided by appending the switch -vv can give us an idea of the issue at hand.
[user1@localhost ~]$optirun -vv vlc [ 52.410461] [DEBUG]Reading file: /etc/bumblebee/bumblebee.conf [ 52.411039] [INFO]Configured driver: nvidia [ 52.411659] [DEBUG]optirun version 3.2.1 starting... [ 52.411716] [DEBUG]Active configuration: [ 52.411745] [DEBUG] bumblebeed config file: /etc/bumblebee/bumblebee.conf [ 52.411765] [DEBUG] X display: :8 [ 52.411785] [DEBUG] LD_LIBRARY_PATH: /usr/lib64/nvidia-bumblebee:/usr/lib/nvidia-bumblebee:/usr/lib64:/usr/lib [ 52.411805] [DEBUG] Socket path: /var/run/bumblebee.socket [ 52.411826] [DEBUG] Accel/display bridge: auto [ 52.411847] [DEBUG] VGL Compression: proxy [ 52.411868] [DEBUG] VGLrun extra options: [ 52.411886] [DEBUG] Primus LD Path: /usr/lib/primus:/usr/lib64/primus [ 52.411939] [DEBUG]Using auto-detected bridge virtualgl [ 52.434739] [INFO]Response: No - error: Could not load GPU driver [ 52.434758] [ERROR]Cannot access secondary GPU - error: Could not load GPU driver [ 52.434763] [DEBUG]Socket closed. [ 52.434781] [ERROR]Aborting because fallback start is disabled. [ 52.434787] [DEBUG]Killing all remaining processes.
When you upgrade from one version of nvidia binary to higher version of the module, sometimes the un-install phase is not complete and in that case this following suggestion should help.
1. First unload the nvidia module by logging in as root in a terminal. Terminal output of cat /proc/acpi/bbswitch should be OFF after this step.
systemctl restart bumblebeed
2. Uninstall the nvidia module
/usr/sbin/bumblebee-nvidia --debug --uninstall
3. modinfo nvidia probably gives you an output similar to this
[root@localhost ~]# modinfo nvidia filename: /lib/modules/3.11.8-300.fc20.x86_64/kernel/drivers/video/nvidia.ko modinfo: ERROR: could not get modinfo from 'nvidia': No such file or directory
4. Prepare to compile the nvidia driver at restart for the current kernel and reboot the system
[root@localhost ~]# touch /etc/sysconfig/nvidia/compile-nvidia-driver [root@localhost ~]# reboot
5. Post reboot should give the modinfo output something similar to this and you should be able to use optirun as intended.
@localhost ~]$ modinfo nvidia filename: /lib/modules/3.11.8-300.fc20.x86_64/kernel/drivers/video/nvidia.ko alias: char-major-195-* version: 331.20 supported: external license: NVIDIA alias: pci:v000010DEd00000E00sv*sd*bc04sc80i00* alias: pci:v000010DEd00000AA3sv*sd*bc0Bsc40i00* alias: pci:v000010DEd*sv*sd*bc03sc02i00* alias: pci:v000010DEd*sv*sd*bc03sc00i00* depends: drm,i2c-core vermagic: 3.11.8-300.fc20.x86_64 SMP mod_unload parm: NVreg_Mobile:int parm: NVreg_ResmanDebugLevel:int parm: NVreg_RmLogonRC:int parm: NVreg_ModifyDeviceFiles:int parm: NVreg_DeviceFileUID:int parm: NVreg_DeviceFileGID:int parm: NVreg_DeviceFileMode:int parm: NVreg_RemapLimit:int parm: NVreg_UpdateMemoryTypes:int parm: NVreg_InitializeSystemMemoryAllocations:int parm: NVreg_RMEdgeIntrCheck:int parm: NVreg_UsePageAttributeTable:int parm: NVreg_MapRegistersEarly:int parm: NVreg_RegisterForACPIEvents:int parm: NVreg_CheckPCIConfigSpace:int parm: NVreg_EnablePCIeGen3:int parm: NVreg_EnableMSI:int parm: NVreg_MemoryPoolSize:int parm: NVreg_RegistryDwords:charp parm: NVreg_RmMsg:charp. parm: NVreg_AssignGpus:charp
Make sure your nvidia module is unloaded when you do this steps. For some repeating the steps a couple of times was necessary.