The ARM-based M1 surely disrupted the CPU market. I have no idea what this means for the future of x86 windows/linux PCs. If Microsoft or some other manufacturer is able to produce an M1 equivalent for the PC (which is a matter of time because the incentive is there), what will happen to intel and AMD? Are we gonna live in a fragmented PC world? At least with Apple you know where things are headed because there is only one chip provider. But we're gonna be in a mess in the next few years if PC-ARM CPUs beco
We are at an inflection point, for sure. So far, it's been a simple choice. Power-budgeted? ARM. Performance? x86 (with some exceptions). We just got to a point where ARM & x86 desktop performance are on-par, where ARM is carried by two trends: 1. The performance improvement of x86 is ~30% year on year. ARM is a few hundred percentage points. 2. The power consumption of a properly designed ARM system is between a quarter and a full order of magnitude lower for the same performance.
Apple is going ARM native with its desktop OS and is providing a translation layer. Windows is going ARM native with its desktop OS and providing a translation layer.
x86 will run on inertia for a (compatibility and change viability "exponential tail long") while (think mainframe), but its death warrant is signed and its fate is sealed. x86 is dethroned.
The gaming world will take decade to figure out how to run if it wants a machine that can run its decades in the making steam library, a higher performing machine that can only run a subset or some hybrid option that can kinda do both. The server world will move much quicker as ARM versions of server OS's have long been out there, and there are fewer software vendors and a greater capability to just rebuild your business apps natively for ARM.
The desktop user who just needs a browser, an office suite and a Teams/Zoom/whatever client will jump overnight the moment the price is right. Whether you're gonna pay a premium for an apple product today or buy a cheaper ARM PC 2 years from now matters little in the grand things of things.
There is no trend, no force, no advantage and no black magic Intel/AMD can pull off that they haven't already, that will make any of these trends exceed what ARM can do, on any timescale - immediate, medium and long term.
The CPU seems to be at a cross-road (Score:2)
Re:The CPU seems to be at a cross-road (Score:2)
We are at an inflection point, for sure.
So far, it's been a simple choice. Power-budgeted? ARM. Performance? x86 (with some exceptions).
We just got to a point where ARM & x86 desktop performance are on-par, where ARM is carried by two trends:
1. The performance improvement of x86 is ~30% year on year. ARM is a few hundred percentage points.
2. The power consumption of a properly designed ARM system is between a quarter and a full order of magnitude lower for the same performance.
Apple is going ARM native with its desktop OS and is providing a translation layer.
Windows is going ARM native with its desktop OS and providing a translation layer.
x86 will run on inertia for a (compatibility and change viability "exponential tail long") while (think mainframe), but its death warrant is signed and its fate is sealed. x86 is dethroned.
The gaming world will take decade to figure out how to run if it wants a machine that can run its decades in the making steam library, a higher performing machine that can only run a subset or some hybrid option that can kinda do both. The server world will move much quicker as ARM versions of server OS's have long been out there, and there are fewer software vendors and a greater capability to just rebuild your business apps natively for ARM.
The desktop user who just needs a browser, an office suite and a Teams/Zoom/whatever client will jump overnight the moment the price is right. Whether you're gonna pay a premium for an apple product today or buy a cheaper ARM PC 2 years from now matters little in the grand things of things.
There is no trend, no force, no advantage and no black magic Intel/AMD can pull off that they haven't already, that will make any of these trends exceed what ARM can do, on any timescale - immediate, medium and long term.