ARM MacBook vs. Intel MacBook: A SIMD Benchmark

Re:ARM MacBook vs. Intel MacBook: A SIMD Benchmark by ithkuil on Sunday December 13, 2020 @04:11PM (#25410321)

Micro-ops, while require intermediate microarchitectural state, doesn't produce inconsistent architectural state when interrupted. The instruction either retired (produced effects) or didn't, from the POV of the ISA (and thus the kernel and user code)

[ Parent | on HN ]

Re:ARM MacBook vs. Intel MacBook: A SIMD Benchmark by addaon on Sunday December 13, 2020 @04:45PM (#25410622)

But I thought that the SVE extension allowed interrupt and restart of an instruction? The alternative seems like it would lead to potentially very high interrupt latency?

[ Parent | on HN ]

Re:ARM MacBook vs. Intel MacBook: A SIMD Benchmark by ithkuil on Sunday December 13, 2020 @04:55PM (#25410733)

I don't know about SVE, just some general principles/thought: you either postpone the serving of the interrupt until the current instruction is done or you throw away progress made in the current instruction.

My educated guess is that the latter is most frequently used as a technique.

Also traps are caused by the very instruction being executed so you cannot complete the current instruction in all circumstances before vectoring to the trap/interrupt handler.

[ Parent | on HN ]