Geforce graphics cards can now allocate their computing power to a Windows virtual machine. An announcement that should especially interest players and game developers on Linux.
Nvidia announces that it is now possible to use the computing capacities of its GeForce graphics card in a Windows virtual machine. It is not uncommon, in fact, for a Linux user to install software on his computer to use the Microsoft OS, and thus to access certain applications.
Until now, it was impossible to take advantage of its graphics card within the virtual machine. This is what Nvidia has just changed. The functionality is also accessible for cards even quite old, since it is available for its Kepler architecture cards (which started with the GTX 680) and all those that follow.
Two types of Linux users will benefit in the first place: those who play on Linux via the virtual machine, and game developers who would like to test their code on both Linux and Windows.
For now, the process still looks quite restrictive. There can only be one GeForce per virtual machine, unless you have an Nvidia Tesla, Quadro or company RTX.
Also, if a graphics card is allocated to the virtual machine, it will not be able to run on the host OS, namely Linux. This means that if you want to have a transversal use of your graphics card on both OS, it is currently impossible. Someone who would like to take advantage of GeForce products on both OS will need to have two graphics cards, one for Linux, and one for the virtual machine.
Play on Linux
It is well known that being a fan of Linux while liking video games has sometimes led to some giving up. If the situation has improved significantly in recent years, there is still a significant proportion of games that cannot run on the penguin OS.
On ProtonDB, a site that lists Steam games that run on Linux through the Proton application, we see that 75% of the 1000 most popular games on the site are playable on Linux.
With this announcement, Nvidia is therefore unquestionably opening a new door for gaming and development on Linux. But with such severe constraints, not sure it will trigger a tidal wave of users.