When looking here:
https://en.wikipedia.org/wiki/Advanced_Vector_Extensions#CPUs_with_AVX
the most concerning sounds:

Not all CPUs from the listed families support AVX. Generally, CPUs with the commercial denomination Core i3/i5/i7/i9 support them, whereas Pentium and Celeron CPUs before Tiger Lake do not.

Tiger Lake is from 2020. So if this bug affects all CPUs without AVX, that would mean it affects very recent Pentium and Celeron CPUs. (And take into account that older CPU architectures still get sold, even when newer ones are already produced).

For Intel Core and AMD, it's much better, it would affect 10+ years old CPUs.

For now I'll vote as FE (maybe vote it as Blocker after read more about this.)

FinalFE +1

Posted some more stats in https://bugzilla.redhat.com/show_bug.cgi?id=2272758#c14 . On the Steam hardware survey currently 97.13% of systems have AVX, so 2.87% don't. Of course, you could say people running Steam are relatively more likely to have higher-end hardware.

This is at least FE for me, very close to blocker. I would love to fix this in the RC-1.13 we're going to have to run, but rebuilding gcc takes 21 hours :( so if we have to do that, we're slipping for sure.

FinalFE +1

FinalFE +1

I'd incline to be blocker-ish here. If we were to raise the x86 level and clearly state that AVX-less CPUs aren't supported, that would be fine with me (and others would be able to enjoy the small benefits of using x86_64-v3 or at least -v2, and we'll get back closer to the el). But this would be confusing random-crashing for the affected users.

This however is something that users gonna hit (are we sure it affects KDE only?), and is gonna stay on Lives for their lifespan.

FinalBlocker +1

Per Jacob, ignoring theoretical CPUs that don't exist yet: "So, that means misbehavior on all CPUs which do have AES but don't have AVX." ... "Looking at the GCC supported -march= CPUs which do have -maes but don't have -mavx it seems it is goldmont, goldmont-plus, tremont and lujiazui."

Affected Intel CPU models can be seen at https://en.wikipedia.org/wiki/Goldmont , https://en.wikipedia.org/wiki/Goldmont_Plus , https://en.wikipedia.org/wiki/Tremont_(microarchitecture) . lujiazui is a Zhaoxin microarchitecture used only in CPUs sold in China - https://en.wikichip.org/wiki/zhaoxin/microarchitectures/lujiazui . No idea how many of those there are, or how many folks using one might want to run Fedora.

Some of those were released as recently as 2019.

It probably doesn't just affect KDE, although that's a pretty visible thing to have egregiously affected. But since there's a fix pending, I could go for FE or even updates-available-sometime-real-soon if we can respin the KDE Live with that

Release F40 is no longer tracked by BlockerBugs, closing this ticket.