Multiple instances of Viz Engine can be spawned along multiple graphic cards. The assignment of the GPU to be used is done automatically if the -u1 -u2 parameters are used. If in addition the -G1, -G2 parameters are provided, the GPU can be bound to the Viz Engine instance.

Some example use cases:

  • Two Program Outputs: Using two instances that can run two program channels on a Dual Channel system.

  • Viz Trio Onebox Installation: Using one instance to render a full SDI/IP Program channel on a strong GPU and a less powerful GPU to render a Viz Engine hosted by the Viz Trio client.

  • One Program Output and One Preview Output: Running a control application with one instance for preview and one instance for program output with video. Ideal for outside broadcast environments.

  • Multiple frame servers on one machine.

Note: The limitations to such systems (in general and not limited to Vizrt systems alone) are memory usage, access and a lack of power redundancy. Also keep in mind that the memory usage of each machine needs to be covered by the system (for example, two P6000 with 24 GB of texture memory require a machine with at least 48 GB).

A Dual GPU setup can also be used with Viz Engine version 5.3 or later to distribute the load between third party render engines (Unreal) and Viz Engine. For more details, please see Setting up a Multi-GPU Workflow in the Viz Artist User Guide.

See Also