• 0 Posts
  • 11 Comments
Joined 3 years ago
cake
Cake day: June 19th, 2023

help-circle


  • I mean, can’t really tell while FEX has not been extensively tested, can we?

    Valve has a great track record with (their contributuons too and use of) Wine, the devs are extremely competent and have been on this for ages, and Valve had any freedom to not choose an ARM chip, as low-power AMD alternatives became viable.

    Can you go lower-power with ARM? Probably, but according to Valve it comes with a 10-20% performance hit for x86 Games at the moment. That means efficency takes a hit as well, whereas I would assume for 20% you are as efficient as an x86 chip at best.

    My guess is that Valve isn’t doing this to chase some efficiency, but for strategic reasons. They are growing independent from Microsoft, now they want to get independent from the oligopoly that is x86 as well. One thing we cannot forget in this discussion is that ARM is likely much cheaper to get, with much more vendors available, and custom designs being quite common.

    Maybe we will see custom ARM hardware with FEX acceleration from/for Valve and suddenly the overhead is almost gone.








  • There is a way to kinda make this work, this would be hardware based security. You could use a TPM to make reasonably sure the kernel is e.g. mainline / hardened / anything else acceptable. Hardware vendors (i.e. Intel, AMD etc.) would have to provide a service where they hash the kernel alongside their keys for the game devs to check against (probably not for free). You would absolutely have to use Secure Boot tho, and eventually keys may be leaked. Another possibility would be devs connecting directly to your TPM to make sure (afaik this is possible in principle, but not mean to be used that way).

    I think there are easier ways to prevent cheating tho, for example simply detecting suspicious activity on the server side, i.e. stats go way up, looking at data coming from clients other than yours.