compile with ghc -dynamic -O2 solver.hs
| input | score | time |
|---|---|---|
| a_example.txt | 17 | <1s |
| b_read_on.txt | 5,822,900 | <1s |
| c_incunabula.txt | 5,688,777 | 49s |
| d_tough_choices.txt | 5,028,725 | 7min |
| e_so_many_books.txt | 5,044,872 | 20s |
| f_libraries_of_the_world.txt | 5,308,034 | 7s |
| total | 26,883,325 | 8min16 |
Same idea behind all these solvers.
- Python solver: https://github.com/PicoJr/2020-hashcode
- Rust solver: https://github.com/PicoJr/2020-hashcode-rust
best in bold
| input | Haskell | time | Python | time | Rust | time |
|---|---|---|---|---|---|---|
| a_example.txt | 17 | <1s | 17 | <1s | 17 | <1s |
| b_read_on.txt | 5,822,900 | <1s | 5,724,900 | 1.5s | 5,821,000 | <1s |
| c_incunabula.txt | 5,688,777 | 49s | 5,685,165 | 47s | 5,683,014 | 4s |
| d_tough_choices.txt | 5,028,725 | 7min | 5,028,465 | 12min | 5,028,725 | 46s |
| e_so_many_books.txt | 5,044,872 | 20s | 4,992,207 | 18s | 4,525,774 | 2s |
| f_libraries_of_the_world.txt | 5,308,034 | 7s | 5,027,259 | 2s | 5,183,128 | <1s |
| total | 26,883,325 | 8min16 | 26,458,013 | 13min | 26,241,658 | 52s |
Performances:
- I am fairly inexperienced with Haskell (1st time writing Haskell)
- Python was run using "vanilla" python (not
cpython), it could probably be faster - Rust was run using
cargo --release, it did not use hardware specific optimizations
Scores:
- Haskell and Rust both use integral division whereas Python use float division
- Rust uses an unstable sort by key