I’m not really defending it. I have a bad habit of pushing back against group think and dog piling.
On another forum I appear to be a Qualcomm white knight (in light of their recent benchmark snafu), simply because there’s a large majority of people there that are really critical of anything that could be perceived as challenging x86.
I getcha, but even if we're looking at this in a no money, only product way:
- MTL came out 9 months later than PHX, is far more expensive and offers the same broad perf, except Intel GPU drivers still suck at gaming
- RPL-R is not changing a thing to RPL except the wattage
- SPR being replaced by EMR is easily one of the best products that has come out in awhile, yes I think EMR is actually worth a damn even if it's not EPYC competitive
- ARL has yet to be confirmed perf wise and I doubt it'll compete with Zen 5
- All the advanced nodes, Intel 4 or 3 etc are not ramped up yet so prod is just meh
- Word is getting out that the govt may just buy Intel and that's a death sentence to me, govts are universally bad at producing things
I don't see a positive outside of EMR here. I don't have any expectations except a big sad stumble with SRF, because spamming e-cores feels like a commercial/license per CPU type of thing, not an actual good product.
Nothing about ARL sounds much better. At worst Skymont looks like it'll be a nice thing. And considering how they took Gracemont and somehow boosted the watts to twice what it should've been to get 10% extra perf, I'm doubtful that they'll do good with it.
It just really is that bad.