How to hash, index, profile, multi-thread, and SIMD your way to incredible speeds.
  • Eager Eagle@lemmy.worldEnglish
    182·
    1 year ago

    Their actual speedup for switching languages was 8x. The rest was all about using better data structures and parallelism.

    • snaggen@programming.dev
      122·
      1 year ago

      But isn’t it kind of obvious that if you are able to do 180k times improvement, then the baseline is probably not very impressive to begin with. Still, that doesn’t take away that the optimizations were impressive, and that it was interesting to read about it.

      • Turun@feddit.de
        4·
        1 year ago

        I think your last sentence has one negation too much.

        If it was interesting to read about it, then the criticism did not take away that the optimizations were impressive.

        • snaggen@programming.dev
          41·
          1 year ago

          Fixed it… I come from a language culture were we like our negations :) Also, not native english speaker, so combine the two and you are in for a ride!

    • Sigmatics@lemmy.ca
      5·
      1 year ago

      Yeah, this one really had me scratching my head:

      ✓Note: there are lots of ways we could make the Python code faster, but the point of this post isn’t to compare highly-optimized Python to highly-optimized Rust. The point is to compare “standard-Jupyter-notebook” Python to highly-optimized Rust.

    • bluGill@kbin.social
      52·
      1 year ago

      With rust is the joke as if you couldn’t do it otherwise. Maybe c would be only 179,999x faster, or FORTRAN 180,001x, (numbers made up). Python could probably be made 60,000x faster as well.

  • uthredii@programming.dev
    7·
    1 year ago

    Would be interesting to see how fast polars (a dataframe library written in rust) would be as it can be used in python.

  • Tocano@lemmy.ml
    31·
    1 year ago

    Move infinitly faster with teleportation!!!

    Walking baseline