3 ms·

>>having to wait upwards of 10 seconds on toy projects<< meteor-contest Rust #2, 545 lines of code, 4.75s to build http://benchmarksgame.alioth.debian.org/u64

by igouy 11y ago

>>having to wait upwards of 10 seconds on toy projects<< meteor-contest Rust #2, 545 lines of code, 4.75s to build http://benchmarksgame.alioth.debian.org/u64q/program.php?test=meteor&lang=rust#log http://benchmarksgame.alioth.debian.org/u64q/program.php?tes... I do see what you mean.

kibwen 11y ago
Grossly misleading. You're measuring the build time of an optimized build, which is not what developers are building during routine development. Optimized builds deliberately trade compilation speed for superior performance.
- igouy 11y ago
  fwiw $ time /usr/local/src/rust-1.3.0-x86_64-unknown-linux-gnu/rustc/bin/rustc meteor.rs -o meteor.rust-2.rust_run real 0m2.042s user 0m1.940s sys 0m0.096s
  - kibwen 11y ago
    And most of that is constant time overhead of linking and trans, rather than linearly scaling overhead. :P
    - igouy 11y ago
      fwiw thread-ring Rust, 38 lines of code $ time /usr/local/src/rust-1.3.0-x86_64-unknown-linux-gnu/rustc/bin/rustc threadring.rs -o threadring.rust_run real 0m1.466s user 0m1.340s sys 0m0.112s
      - kibwen 11y ago
        Yes, thank you for proving my point while I'm away from my computer. :)
        igouy 11y ago
        I don't have an-axe-to-grind :-)
- Jweb_Guru 11y ago
  It's not misleading, because Rust focuses so little on -O0 performance that it's often not plausible to use a Rust build with that optimization level. We can't have it both ways.