The 15 tables in the nvidia grid documentation kind of shows how much of a mess it is. Different resolutions and maximum tenants depending on card and license

The 15 tables in the nvidia grid documentation kind of shows how much of a mess it is. Different resolutions and maximum tenants depending on card and license type, and you can't allocate resources homogeneously.