- Joined
- Apr 12, 2021
- Messages
- 901
- Motherboard
- Asus z590 ROG Maximus XIII Hero
- CPU
- i9-11900K
- Graphics
- RX 6600 XT
- Mac
- Classic Mac
- Mobile Phone
This was promoted in GoogleNews today: A comparison of PCI3 vs PCI4 on RTX4090 gaming perf.
I don't care for these sorts of channels and the personality blah, blah I can take or leave, but the topic is of note.
tldr:
It foreshadows that PCI5 will make no diff to top-end graphics performance for gaming workloads. Accordingly, PC GPU innovation is driven by gaming market.
Shifting gears—
So the question gets begged:
How does AppleSi design harbor different ways of thinking about GPU (or storage or networking or other) acceleration, where next gen (say next year) includes performance features on radically different terms than PC.
My point for consideration is thus:
AppleSi opens doors to a lot more than just making a competitive PC for macOS. Apple can balance the system to very different priorities, where performance advantages might take new turns.
The world is overdue for a "network is the computer" (old Sun motto) innovation, which -could- be all about gaming, except games by definition must be centrally mediated (without this mediation it is war) and if design is centrally mediated then its purposes are also centrally mediated and that's robotic, not liberated.
The abject failure of Zuckerberg and his Meta to achieve anything other than a goofy encroachment on VR, given top-of-the-line goggles (especially considering Second Life has been at it for 20 years and no one cares) is a benchmark for failure. The next steps in personal computing seem not-obvious.
I cannot argue that Apple has any better grasp of the future—2024 cannot be like 1984—but do you just go your own way building your own architecture, soup through nuts, with no idea as to opportunities?
My sense of history is there's a tendency towards atrophy when one company owns applications for designs top-to-bottom: Bell and Xerox come to mind. For well understood applucations with enormous infrastructure, this is ok—say Boeing.
WRT to Intel PC, the coupling of physics to application is much more diffuse.
In one sense, Apple looks like an outlier.
But Apple WRT to Apple...!?
Interesting times.
My train of thought began with news that the latest PC bus advance doubles raw performance to offer a 5% end user performance increase... Sounds like something has gone wrong with the architecture.
I don't care for these sorts of channels and the personality blah, blah I can take or leave, but the topic is of note.
tldr:
It foreshadows that PCI5 will make no diff to top-end graphics performance for gaming workloads. Accordingly, PC GPU innovation is driven by gaming market.
Shifting gears—
So the question gets begged:
How does AppleSi design harbor different ways of thinking about GPU (or storage or networking or other) acceleration, where next gen (say next year) includes performance features on radically different terms than PC.
My point for consideration is thus:
AppleSi opens doors to a lot more than just making a competitive PC for macOS. Apple can balance the system to very different priorities, where performance advantages might take new turns.
The world is overdue for a "network is the computer" (old Sun motto) innovation, which -could- be all about gaming, except games by definition must be centrally mediated (without this mediation it is war) and if design is centrally mediated then its purposes are also centrally mediated and that's robotic, not liberated.
The abject failure of Zuckerberg and his Meta to achieve anything other than a goofy encroachment on VR, given top-of-the-line goggles (especially considering Second Life has been at it for 20 years and no one cares) is a benchmark for failure. The next steps in personal computing seem not-obvious.
I cannot argue that Apple has any better grasp of the future—2024 cannot be like 1984—but do you just go your own way building your own architecture, soup through nuts, with no idea as to opportunities?
My sense of history is there's a tendency towards atrophy when one company owns applications for designs top-to-bottom: Bell and Xerox come to mind. For well understood applucations with enormous infrastructure, this is ok—say Boeing.
WRT to Intel PC, the coupling of physics to application is much more diffuse.
In one sense, Apple looks like an outlier.
But Apple WRT to Apple...!?
Interesting times.
My train of thought began with news that the latest PC bus advance doubles raw performance to offer a 5% end user performance increase... Sounds like something has gone wrong with the architecture.
