3) Does Quick Sync even matter if you've got something like an RX 5700 XT?
Yes it does.
I am not informed enough to offer a rundown of all the ways Mac employs the iGPU, but it's clear that the way a SW attack applies HW really matters to the total balance of system fitness.
I have no axe to grind regarding the details of Alder Lake and hackintosh, and my points aren't intended to cause disturbance of enthusiasm.
Rather, my intentions are to raise the conversation level to appreciate why designs are taking certain turns and just to express my own sense of 'enlightenment' at the big changes in Mac that have occurred in the last two years and how these make real Mac much better at what it is. It's a fascinating case study in architecture, and I have some background in matter, and have watched it unfold since back when Mac was young. I recall in Intel component engineering offices in 1985, next to a directors office there was an unused Apple Lisa as he raged that VAX/VMS had just thrown away hours of email he'd been writing on a VT100 but never saved. It was a big deal in late 80s when Motorola 68000-based Suns were deployed instead of DEC to do a custom scientific massively-parallel system component. The marketing dept used Macs even as PC was hitting stride. A decade later Intel was still not using its own silicon to design its silicon, and a vast and costly new 4-component system that had been intended to be a VAX killer ended completely failed to meet expectations as x86 ate the world. It's CPU was the i960 and it ended up not in advanced workstations, but as the heart of HP laser-printers, thanks to the GNU C compiler. A special version also got put in the Space Shuttle as a microcontroller. The director of Windows NT at Microsoft came from DEC and everyone noticed that WNT followed VMS just like HAL followed IBM. Along the way there was VHS vs Beta, CDs vs vinyl, cats and dogs living together! Then there was PowerPC Mac, the expunging of Jobs, the Megahertz Myth, the near death of Apply under Scully, NeXT, the return of Jobs, Apple G4 Personal Supercomputer with 1 GFLOPS, Final Cut Pro. Blah, blah. In 1999 Intel finished the fastest computer in the world which sustained 1 TFLOPS Linpack, and delivered to Sandia National Labs for 50 million. It consumed megawatts. Now trillion OPS is ordinary GPU. Around same time, internally, Intel component engineering fellows were on a presentation junket to inform mgmt that nothing was going much past 5GHz because of the exponential power density, where the powerpoint has a conclusion slide with a graph showing clockspeed versus density gains that ended at temperatures equal to surface of sun. Intel technology directors were holding up demo of 64 core dies. Again this was very late '90s. Then the day that will live in infamy: the 2006 4-core Xeon Mac Pro and the Core Duo Macbook Pro. It seemed that Intel had finally won over the entire world.
15 years later, I mention all this just to frame my curiousity and perspective on what incremental architectural improvements might mean for the future.
Apple has a good record at the long game. We should assume they're still thinking strategically with the M1, and that 1 means 2, etc.
All of the nuts and bolts are vastly more complicated than we prefer, or can imagine.
If you read up about how Apple stack applies AppleSi, you will quickly realize that they have a power/performance vision which Intel cannot touch, because it's not about feature-by-feature parity, it's about that progress now requires tuning the whole stack, and Intel can't do this.
Apple's influence lead Intel PC to become a little more architecturally coherent than it would otherwise be, but Apple probably couldn't influence Intel enough to suit its plans, and also suffered through some big gotchas on performance due to Intel architecture security defects. But the IA PC is a herd of cats. Left to their own terms, Intel and Microsoft can't evolve the PC beyond Office and games. Moreover, the architecture's woeful insecurity is an unavoidable manifest hazard to everything it's put in, which includes very high-risk applications. Mantaining productivity of Windows is a cost nightmare.
Apple's recent course change looks like one of those momentous points that leaves a whole way of thinking in its wake.
Again, I mention all this to support my view that the looming problem for hackintoshers is not 'what's the most compatible chipset', it's that there will soon be no clear way for hackintosh at all. It's an idea which begins and ends with Intel Macs. The end is nigh! However while the end is arriving, we can enjoy the soiree over what it feels like to live through end-times and count our blessings.
Carry on with the good work!
