The post with egregious errors was also put up on a Sunday afternoon. And while we're all acting conciliatory now, it's pretty remarkable how biased the post was, the author using some clearly erroneous numbers to prove their prior, baseless claim that the "M1 chip is far inferior" in some respects, when those respects were specifically SIMD. Then becoming strangely defensive when some people rightly pointed out that ARM64 has 128-bit NEON and a number of other advantages.
Far inferior becomes....actually superior in many cases, even at SIMD.
Let’s try to be charitable, shall we? Everyone makes mistakes sometimes, even leading experts in low-level algorithm optimization. Lemire was upfront about making a mistake, and not at all defensive about it; if you are reading it that way, it’s just you.
It is clearly the case that the M1 CPU/SoC has a significant performance advantage in typical branchy single-core code, but much less advantage if any for certain kinds of heavily optimized numerics. Beyond that high-level summary, it’s good to dive into the details, and spark discussions.
Everyone is just now getting their hands on these chips, learning how to work with them, and trying to figure out how to best optimize for them.
Far inferior becomes....actually superior in many cases, even at SIMD.