Advent of Code 2017

All my Advent of Code repos:

• AoC 2015 in Nim, Python
• AoC 2016 in Python, Clojure (+ visualizations)
• AoC 2017 in Nim, OCaml, Python (this repo)
• AoC 2018 in Nim, Python, Racket
• AoC 2019 in OCaml, Python
• AoC 2020 in Nim, one liner-y Python, Racket
• AoC 2021 in Python, Racket
• AoC 2022 in Python, Clojure
• AoC 2023 in Clojure
• AoC 2024 in Clojure (Clerk notebooks)

 

Solutions

My aim is to provide clean and readable, yet idiomatic, solutions in all three languages. If you have any comment/suggestion/advice, please let me know!

Originally, I've solved all tasks in Nim and Python as they were released. OCaml solutions were added in November 2019 as a preparation for AoC 2019 (these were my first steps in OCaml).

Task	Nim solution	OCaml solution	Python solution	Note
Day 1: Inverse Captcha	day01.nim	day01.ml	day01.py	Taking advantage of Python's negative indices.
Day 2: Corruption Checksum	day02.nim	day02.ml	day02.py
Day 3: Spiral Memory	day03.nim	day03.ml	day03.py	Building a spiral with table/map/dict in all three versions; using iterators in Nim and Python.
Day 4: High-Entropy Passphrases	day04.nim	day04.ml	day04.py
Day 5: A Maze of Twisty Trampolines, All Alike	day05.nim	day05.ml	day05.py	Used try-except in Python for some nice speed improvement.
Day 6: Memory Reallocation	day06.nim	day06.ml	day06.py	Python doesn't have OrderedSet (had to use OrderedDict).
Day 7: Recursive Circus	day07.nim	day07.ml	day07.py	Python's Counter.most_common() is quite helpful/useful here.
Day 8: I Heard You Like Registers	day08.nim	day08.ml	day08.py
Day 9: Stream Processing	day09.nim	day09.ml	day09.py
Day 10: Knot Hash	day10.nim	day10.ml	day10.py	Changed solutions to be reusable for Day 14. Python version uses deque with pop, rotate, and insert. Nim version is a more 'traditional' one.
Day 11: Hex Ed	day11.nim	day11.ml	day11.py	Python version uses cube coordinates, Nim and OCaml versions use axial coordinates.
Day 12: Digital Plumber	day12.nim	day12.ml	day12.py	BFS in Python, DFS in Nim and OCaml.
Day 13: Packet Scanners	day13.nim	day13.ml	day13.py	All three versions precalculate possible values of delay using Chinese remainder theorem to gain a significant speedup.
Day 14: Disk Defragmentation	day14.nim	day14.ml	day14.py
Day 15: Dueling Generators	day15.nim	day15.ml	day15.py	Python: generator generator generating generator's values. In Nim, using bit masking gives great speed boost.
Day 16: Permutation Promenade	day16.nim	day16.ml	day16.py
Day 17: Spinlock	day17.nim	day17.ml	day17.py	Brute force in Python, using deque.rotate. The expected version in Nim, optimized.
Day 18: Duet	day18.nim	day18.ml	day18.py
Day 19: A Series of Tubes	day19.nim	day19.ml	day19.py	All three solutions use complex numbers, which are great for the rotations in 2D plane.
Day 20: Particle Swarm	day20.nim	day20.ml	day20.py
Day 21: Fractal Art	day21.nim	day21.ml	day21.py	Unoptimized solution in OCaml. Nim and Python solutions are optimized for the second part. Python version uses numpy and expands the grid (3 steps at once), Nim version counts the number of times each pattern is present after 18 iterations.
Day 22: Sporifica Virus	day22.nim	day22.ml	day22.py	OCaml: sum types and pattern matching is the name of the game. Python version uses a dict and a complex plane, Nim version uses an array (faster than a table) of a regular 2D plane with enum for the rotating directions.
Day 23: Coprocessor Conflagration	day23.nim	day23.ml	day23.py
Day 24: Electromagnetic Moat	day24.nim	day24.ml	day24.py	BFS in Python. A recursive search in Nim and OCaml, optimized.
Day 25: The Halting Problem	day25.nim	day25.ml	day25.py	Python version uses (default)dict. Nim version uses arrays, which are much faster than tables.
Total time:	0.49 sec	1.01 sec*	16.1 sec*	* OCaml: unoptimized day21.ml. Python: without the brute-forced day17.py, and day15.py was run in pypy3. For the detailed run times, see below.

 

Run times

• Nim version: 1.6.6
• OCaml version: 4.13.1
• Python version: 3.8.2
• CPU: AMD Ryzen 3700x @ 3.6 GHz (Linux 5.15)

Times are in milliseconds, the reported results are the average of 20 runs.

day	nim	ocaml	python
01	0.3	0.8	13.3
02	0.5	0.8	13.2
03	0.4	0.7	13.6
04	2.3	2.4	15.7
05	69.0	63.8	2732.1
06	2.4	3.6	32.6
07	2.4	3.9	20.9
08	1.2	1.5	14.5
09	0.4	1.9	17.2
10	0.5	1.4	24.7
11	0.6	1.3	19.5
12	1.8	2.3	16.7
13	1.4	17.4	21.5
14	8.8	32.3	566.2
15	195.4	383.2	2083*
16	6.8	71.0	143.5
17	0.9	4.8	-*
18	0.7	4.6	88.8
19	0.8	1.2	21.4
20	9.2	18.2	1622.5
21	0.5	145.0	197.1
22	50.4	139.3	3555.9
23	0.5	1.0	15.1
24	10.8	52.1	1242.2
25	48.8	52.3	3649.1

OCaml Day21 was not optimized.
Python Day15 was run with pypy3, Python Day17 was brute-forced.