The researchers published a list of mitigations they believe will address the vulnerabilities allowing both the FLOP and SLAP attacks. They said that Apple officials have indicated privately to them that they plan to release patches.
So this'll likely be mitigated soon, and while you're probably right about the performance hit (which will likely be minor), I don't think (most) Apple users need to be very worried about this.
I will die here
Kind of what you were sent there for :|
From the top of my head of the few numbers I've seen it ranges from no improvement at all to solid improvements depending on the game. I think Cyberpunk is one of the games that benefit from ntsync vs fsync quite substantially.
Vs. upstream Wine which doesn't have Fsync. Proton does and so the difference is way smaller (if any).
Apple was very late to add AV1 support to their ecosystem in general. As you state, support for hardware decoding was only added with the M3/A17 Pro chips in 2023. There's still no AV1 hardware encoder on any of Apple's chips.
I think they were waiting on H.266 and whether it succeeds for too long, they were/are big on H.265 (and all the other HEVC-related stuff like HEIC) so that'd make sense from that perspective.
Crawl a little further up Trump's arse, Nvidia!
Not buying them for my next upgrade.
So then it'll have native Linux Denuvo I guess..?
Denuvo is fully intact under Linux though. It's actually even worse than under Windows or at least it used to be, because switching between different Proton versions actually counts towards the 5 machines within 24 hours limit.
With Denuvo DRM!
GPU that's roughly on par with the Steam Deck.
...when comparing TFLOPs, and that's not comparable across architectures (by different companies as well!).
If we take similar-performing (in rasterization) Ampere and RDNA 2 cards (say a 3080 and 6800 XT), we can see the 3080 has 29.77 TFLOPs and the 6800 XT has 20.74 TFLOPs, an RDNA 2 FLOP is worth about 1.4x as much as an Ampere FLOP.
So extrapolating the 1.6 "RDNA 2 TFLOPs" of the Deck we get 2.24 "Ampere TFLOPs" and that'd make the Deck quite a bit faster than the Switch 2 in portable mode, but slower than the Switch 2 in docked mode.
This is obviously all just wild and silly speculation, but I doubt the Switch 2 will match the Deck in portable mode. Samsung 8nm would just eat too much power for this to realistically happen in a handheld form factor.
Pretty much every Secure Boot device trusts Microsoft by default, which is why I think it's pretty much useless (in its default state anyway).
Normally, list items have an active state when they are being tapped (example from Mlem):
Lemma doesn't seem to have any special state for an active list item. This can make it seem like one didn't actually touch the item, it feels kind of weird to use.
Speculative execution seems to be the source of a lot of security flaws in many different CPUs. CPU manufacturers seem to be so focused on winning the performance race that security aware architecture design takes the backseat.
Also, it's more and more clear that it's a bad idea that websites can just execute arbitrary code. The JS APIs are way too powerful and complex nowadays. Maybe websites and apps should've stayed separate concepts instead of merging into "web apps".
I also wonder if it'd be possible to design a CPU so vulnerabilities like these are fixable instead of just "mitigable". Similar to how you can reprogram an FPGA. I have no clue how chip design works though, but please feel free to reply if you know more about this.