primusrun(1) run an application on a discrete NVIDIA video card


primusrun command


Primus implements low-overhead local-only client-side OpenGL offloading via GLX forking.

It is currently intended to be used alongside Bumblebee and provides a drop-in replacement for optirun (i.e. "primusrun").


The following is a list of environment variables affecting primus library that may be relevant for end users:
Readback-display synchronization method (default: 0)
0: no sync, 1: synced, display previous frame, 2: synced, display latest frame
Verbosity level (default: 1)
0: only errors, 1: warnings, 2: profiling
The secondary Xorg server display number (default: :8)


primusrun glxgears -info
Runs the graphics demo supplied by mesa-utils to confirm whether the discrete card is being used for GL rendering.
PRIMUS_VERBOSE=2 primusrun glxgears
Runs the graphics demo supplied by mesa-utils with verbose output from primus.
vblank_mode=0 primusrun glxgears
Disable vblank synchronisation, typically used for benchmarking purposes.


Since compositing hurts performance, invoking primus when a compositing WM is active is not recommended. If you need to use primus with compositing and see flickering or bad performance, synchronizing primus' display thread with the application's rendering thread may help.

PRIMUS_SYNC=1 primusrun ...

This makes primus display the previously rendered frame. Alternatively, with PRIMUS_SYNC=2 primus will display the latest rendered frame, trading frame rate for reduced visual latency.


Primus was created by Alexander Monakov <[email protected]>.

This manual page was written by Vincent Cheng <[email protected]>, for the Debian project (and may be used by others).