1. 11
  1.  

  2. 3

    Disappointing that the benchmark data is only shown for server-type CPUs, whereas when I think ARM I think of mobile (and now Apple desktops too.)

    In particular, it would be interesting to see how Apple’s CPUs compare — the M1 is alleged to have some major optimizations for atomic operations, since those tend to be bottlenecks in Swift and Obj-C refcounting.