utterstep / advent-2019

Rusty Advent of Code solutions – 2019

advent-2019

Rusty Advent of Code 2019 solutions

Current running time

utterstep@utterstep-nix:~/my/advent-2019$ head /proc/cpuinfo
processor       : 0
vendor_id       : GenuineIntel
cpu family      : 6
model           : 158
model name      : Intel(R) Core(TM) i7-8750H CPU @ 2.20GHz
stepping        : 10
microcode       : 0xca
cpu MHz         : 4018.411
cache size      : 9216 KB
physical id     : 0

utterstep@utterstep-nix:~/my/advent-2019$ taskset -c 3 hyperfine --warmup 20 ./target/release/run-all
Benchmark #1: ./target/release/run-all
  Time (mean ± σ):     121.3 ms ±   0.5 ms    [User: 106.6 ms, System: 14.6 ms]
  Range (min … max):   120.4 ms … 122.8 ms    24 runs

Currently day-12 (2nd part) is taking ~50% of all running time (~68ms, down from ~400ms initially).

Growth points

• day-3: check, why precomputed sums looks slower, than brute force approach
• day-7: get rid of permutohedron crate
• day-13: more effective bot strategy (current time ~9ms because there are lot of redundant moves in the end)
• day-14: compute SolutionPrecalc once for both parts
• day-17: current solution looks like tailored to the specific class of cases. Rewrite in more generic way OR prove that it's generic
• day-19: determine error sign — currently I'm not totaly sure that constants are good for every possible input (though they are quite conservative)