What does /g/ think is the next step in computing? Will we see

64+ core arm chips

>>56281915
>next step
> 32-bit CPUs with 64-bit extensions bolted on with 128-bit extensions
Nigga you're describing SSE from 1999.

We're currently on 512bit bolted on extensions.

>>56281915
More efficient manufacturing processes, new materials, diminishing, incremental gains in performance to enhance the x86 architecture.

We're not ditching x86 any time soon, there is simply no reason to.

>>56281915
You mean will we see 16-bit CPUs with 32-bit extensions bolted on with 64-bit extensions bolted on with 128-bit extensions bolted on

daily reminder that itaniums were retarded shit and you should feel bad for being a nerd hipster

File: indigo_doorstop.jpg (39KB, 600x800px) Image search: [Google]

39KB, 600x800px

>>56281915
>Also, daily reminder that Intel actually tried to move on from x86 once, and AMD is the reason we're still on it.
Do yourself a favor and stop falling victim to the same mindless engineering masturbation that made the Itanic sink in the first place. It didn't die because AMD released a chip two years after the hype train derailed, it died because it was a fucking disaster; prohibitively expensive, poor native software support, slower than shit from an ass x86 execution, dreadful integer performance, disappointing compiler performance, horribly botched OEM migrations, lackluster systems. All the big SPECfp numbers in the world don't matter for shit in the face of that, especially in the consumer market that never bought PCs because they fast.

The Itanium and desktop RISC died for a reason, they may have been exciting, they may have been more "elegant", they may have introduced fascinating design paradigms that were ahead of their time and performed admirably for certain applications, but in the end, all that "elegance" does is hamper usability for little or no gain, because it's all a bunch of trivial bullshit that's hidden under several layers of abstraction for 99% of end users and developers alike anyway, while the 1% who care are already using different architectures, or already worked their way around those dreadful quirks decades ago.

>>56284746
They weren't totally retarded, while they absolutely failed to accomplish what Intel wanted, those big caches, blazing fast FP units and RAS features made them excellent chips for big iron database systems, supercomputers and mission-crit boxes.

>>56281915
stacked processors, just like stacked memory modules.

>>56281915
>Also, daily reminder that Intel actually tried to move on from x86 once, and AMD is the reason we're still on it.

The reason we're still on x86 is because the Itanium was literal shit.