I’d say we never actually had the x86 monoculture, we did in the desktop space for the most part (though Apple were rocking the 6800/68k/Power architectures for a large chunk of that too). But the server and workstation space was all sorts of weird and wonderful through all of that with Alpha, SPARC, mips, PowerPC hanging around until the 00s. The last SPARC processor was released less than a decade ago by Fujitsu.
Fujitsu’s current gen chips are based on ARM.
Apple have moved to ARM
More and more AWS compute workloads are getting deployed to ARM servers
Microsoft are getting closer to making windows on ARM actually useful for most people.
The entire mobile devices market is effectively all ARM.
I know RISC-V is coming along, you can buy Dev boards and it’s being deployed in some niche areas. But it needs to blow up like ARM has yesterday, I fear we’re rapidly approaching a situation that could be worse than the x86 one ever was.
Though I’m fairly confident it’s going to take a long time before Intel actually abandons x86
Microsoft are getting closer to making windows on ARM actually useful for most people.
I think it’s a bit early to claim this. Maybe with the Nvidia / Mediatek WoA devices next year. Snapdragon X Elite devices have little to offer over Lunar Lake / Strix Point devices other than compatibility problems. You need to have better performance (at least for ARM native apps), better battery life or price versus x86 devices for WoA to be useful for average consumers.
So every server I’ve ever racked has been x86, every laptop I’ve imaged or deployed was x86, only IOT and embedded have I had to work with a non x86 instruction set.
The x86 monoculture is all I’ve known and it hasn’t been a problem because it’s an open platform. I think the “RISC” with ARM is like what I said, locked down bootloaders and architecture. Asahi and Graphene are amazing projects in this space and they give me hope, but the ARM architecture won’t be the problem, vendor lock in will be.
I’d say we never actually had the x86 monoculture, we did in the desktop space for the most part (though Apple were rocking the 6800/68k/Power architectures for a large chunk of that too). But the server and workstation space was all sorts of weird and wonderful through all of that with Alpha, SPARC, mips, PowerPC hanging around until the 00s. The last SPARC processor was released less than a decade ago by Fujitsu.
Fujitsu’s current gen chips are based on ARM.
Apple have moved to ARM
More and more AWS compute workloads are getting deployed to ARM servers
Microsoft are getting closer to making windows on ARM actually useful for most people.
The entire mobile devices market is effectively all ARM.
I know RISC-V is coming along, you can buy Dev boards and it’s being deployed in some niche areas. But it needs to blow up like ARM has yesterday, I fear we’re rapidly approaching a situation that could be worse than the x86 one ever was.
Though I’m fairly confident it’s going to take a long time before Intel actually abandons x86
Edit: typo
I think it’s a bit early to claim this. Maybe with the Nvidia / Mediatek WoA devices next year. Snapdragon X Elite devices have little to offer over Lunar Lake / Strix Point devices other than compatibility problems. You need to have better performance (at least for ARM native apps), better battery life or price versus x86 devices for WoA to be useful for average consumers.
So every server I’ve ever racked has been x86, every laptop I’ve imaged or deployed was x86, only IOT and embedded have I had to work with a non x86 instruction set.
The x86 monoculture is all I’ve known and it hasn’t been a problem because it’s an open platform. I think the “RISC” with ARM is like what I said, locked down bootloaders and architecture. Asahi and Graphene are amazing projects in this space and they give me hope, but the ARM architecture won’t be the problem, vendor lock in will be.