What mechanisms (if any) modern GPUs implement to prevent different processes from accessing each other's memory (on graphics card)?

1 Answer 1



Source: https://docs.microsoft.com/en-us/windows-hardware/drivers/display/gpummu-model

That's the first result in a search engine while looking for "MMU GPU".

  • 2
    that does require the root process to set this up... (e.a. prevent process A from reading memory of process B). Most Kernels have the features for this onboard (Or memory protection would be non-existent). But it must also be supported in the Drivers of said video card (not so much an issue on windows, but on other OSes this is not a given yet...). Also consider that with things like pass-through you could yield control of this feature to the child OS, instead of the mother. making it possible to read the contents from the mother...
    – LvB
    Nov 23, 2020 at 10:47
  • @LvB Yes. Feel free to edit my answer or to post you own if you wish to improve upon mine.
    – A. Hersean
    Nov 23, 2020 at 14:08
  • 1
    I feel your answer is sufficient at a surface level. I added some more details in a comment for if someone wants to dive a little deeper. Your answer does not need to be edited imho because it is not wrong. (And I believe it’s up to the person who answered to judge if something should be added to it. Not me)
    – LvB
    Nov 23, 2020 at 14:26

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .