Friday, 2017-12-01

Advent of Code 2017

Project website: Advent of Code 2017.

Previous years: 2015, 2016.

I use Perl for all the solutions.

Most assume the input data is in a file called input.txt in the same directory as the file. Some require extra modules from CPAN.

A note on scoring

I score my problems to mark where I’ve finished a solution myself or given up and looked for hints. A score of 2 means I solved both the daily problems myself, a score of 1 means I looked up a hint for one of the problems, and a zero score means I didn’t solve any of the problems myself.

The times are hours and minutes from the time of the puzzle release (00:00 EST) until I posted an announcement on Twitter.

My goals for this year (in descending order of priority):

  • match or beat last year’s score of 49/50.
  • solve all problems within 24 hours of release
  • be among the first 750 on each day’s leaderboard

Final score

  • Problems solved by myself: 47/50. Not really happy with this. I got caught up in the hunt for keeping my place on the local leaderboard, instead of taking the time to actually solve the problems.
  • 3 parts not solved within 24 hours.
  • Average finishing place: 1,801 for part 1, 2,151 for part 2.

Day 1 - Day 2 - Day 3 - Day 4 - Day 5 - Day 6 - Day 7 - Day 8 - Day 9 - Day 10 - Day 11 - Day 12 - Day 13 - Day 14 - Day 15 - Day 16 - Day 17 - Day 18 - Day 19 - Day 20 - Day 21 - Day 22 - Day 23 - Day 24 - Day 25

Day 1 — Inverse Captcha

Day 1 - complete solution

A nice little warmup problem. Eric pulls his usual trick of making you assume something in part 1 that you have to unassume in part 2.

Score: 2

Time: 47m

Leaderboard placement: 783 / 748

Day 2 — Corruption Checksum

Day 2 - complete solution

My suspicion that there’s going to be a lot of functional list-munging in this year’s edition is strengthening.

Score: 2

Time: 2h40m (it’s a Saturday, so I slept in!)

Leaderboard placement: 2,676 / 2,318

Day 3 — Spiral Memory

My solution for day 3: part 1, part 2

A nice headfake again. Of course you’re not going to brute-force the spiral, keeping track of the coordinates for each element, then taking the Manhattan distance of the element you want when you reach it!

Instead you use a solution from Project Euler to segment the search space and find the solution nice and quick.

Then you get to part 2 and discover you need to generate the spiral after all…

Score: 2

Time: 8h23m (first Advent Sunday, shopping for Christmas)

Leaderboard: 2,578 / 3,249

Day 4 — High-Entropy Passphrases

Day 4 - complete solution

A stark contrast to yesterday’s problem. Perl’s text processing features usually make parsing input a breeze but in this case it was almost too easy. I guess I have to be careful for what I wish for in the future.

Score: 2

Time: 37m

Leaderboard: 1,544 / 1,314

Day 5 — A Maze of Twisty Trampolines, All Alike

Day 5 - complete solution

Ah, the list of instructions! An AoC classic.

Wise to previous problems I decided to incorporate test input from the get-go, and to include a debug output. This was a nice addition for part 2, as I could just run the test input and instantly confirm my changes worked.

Runtime for part 2 is around 17.5s, which, as the answer is around 25M steps leads to a “MIPS” value of 1.4.

Score: 2

Time: 41m

Leaderboard: 1,865 / 1,712

Day 6 — Memory Reallocation

Day 6 - complete solution

I liked this puzzle, even if it was a bit on the easy side.

An interesting algorithm for detecting cycles was pointed out in the daily solutions thread: Floyd’s Tortoise and Hare. I don’t think it will give better performance for the limited data set of the puzzle, but part of the fun of AoC is learning new things.


If I get some time in the future I’d like to rework this problem using an iterator for generate the states and using the Floyd algorithm.

Score: 2

Time: 51m

Leaderboard: 1,553 / 1,433

Day 7 — Recursive Circus

Day 7 - complete solution

I was sure my lack of CS knowledge would bite me, but after 10 hours I’m not seeing any comments like “actually this is just algorithm X” on the subreddit daily thread. I guess most people solved this like I did.

A fun problem, all in all.

Score: 2

Time: 5h22m - solved part one before leaving for work, but part two was delayed by meetings and a reinstalled work computer.

Leaderboard: 1,518 / 2,406

Day 8 — I Heard You Like Registers

Day 8 - complete solution

As this is not my first register rodeo, I ran a quick check to make sure the input values were what I expected (comparison operators, whether a field was an integer or not). After that it was simply a matter of encoding the rules.

I could have used Perl’s eval to evaluate the expressions, but I’m not 100% comfortable with it so I just used a giant if/else statement instead.

I chose to write a specific function to read values from the register hash, which made inserting a statement to capture part 2 very simple.


Implement using dispatch table?

Score: 2

Time: 1h15m

Leaderboard: 1,667 / 1,639

Day 9 — Stream Processing

Day 9 - complete solution

This was a fun one!

After perusing other solutions for day 8 I decided to implement a dispatch table for handling the characters. I’m glad I did because it made adding further checks quite easy.

Part 2 threw me for a loop. I got the wrong answer for the number of garbage characters, and trying to find where my assumptions were wrong in the giant puzzle input felt impossible. However, there are a number of example inputs, and after running my code on them I found where I was mistaken…

Score: 2

Time: 5h35m (I’m not gonna get up at 6AM on a Saturday…)

Leaderboard: 2,604 / 2,850

Day 10 — Knot Hash

My solution for day 10: part 1, part 2

Difficulty is ramping up.

The biggest issue I had with this was the @#%&*! array manipulation in part 1.

Score: 2

Time: 5h09m

Leaderboard: 2,149 / 1,867

Day 11 — Hex Ed

Day 11 - complete solution

A fun problem. I didn’t know anything about hex grids but like everyone and their mom I googled and found this excellent site that explains them well.

Score: 2

Time: 1h02m

Leaderboard: 1,110 / 1,007

Day 12 — Digital Plumber

Day 12 - complete solution

I could probably have solved this in a more Perlish way but here I at least understand what’s going on.

Score: 2

Time: 2h47m

Leaderboard: 2,108 / 2,038

Day 13 — Packet Scanners

My solution for day 13: part 1, part 2

This was the first problem that has a significant runtime (without optimization). Even after some work I have a ~10m runtime for part 2.


Optimize for faster runtime.

Score: 2

Time: 13h23m

Leaderboard: 2,462 / 4,224

Day 14 — Disk Defragmentation

Day 14 - complete solution

A nice mix of difficulty and entertainment. Unreasonably proud of independently finding the four-way flood fill algorithm.

Score: 2

Time: 4h51m

Leaderboard: 1,096 / 1,636

Day 15 — Dueling Generators

Day 15 - complete solution

Brute force, baby!

I used an iterator for part 2, then wrangled both solutions into one using that technique.

The generators are well known.

Score: 2

Time: 3h06m

Leaderboard: 1,130 / 1,929

Day 16 — Permutation Promenade

Day 16 - complete solution

After the usual struggles with Perl’s lists I got this working.

I decided to go with a dispatch table from the start, which made part 2 easier to implement.

Score: 2

Time: 4h34m

Leaderboard: 1,682 / 1,370

Day 17 — Spinlock

My solution for day 17: part 1, part 2

Solving part 1 gave some data for how to solve part 2. It took some fiddling though!

Score: 2

Time: 8h52m

Leaderboard: 2,067 / 2,990

Day 18 — Duet

My solution for day 18: part 1, part 2

So far the most entertaining problem. Bravo!

Score: 2

Time: 4h33m

Leaderboard: 1,015 / 1,057

Day 19 — A Series of Tubes

Day 19 - complete solution

This was supposed to be a “breather” problem, instead it stymied me something fierce.

Score: 2

Time: 8h36m

Leaderboard: 2,744 / 2,688

Day 20 — Particle Swarm

My solution for day 20: part 1, part 2

Gratified to finally be among the first 1,000 to solve this problem, especially considering how relatively easy it was.

Update Day 20 - alternative part 1 - closed form

Score: 2

Time: 2h29m

Leaderboard: 913 / 984

Day 21 — Fractal Art

Day 21 - complete solution

What a fiddly problem!

Score: 2

Time: > 24h

Leaderboard: 2,426 / 3,035

Day 22 — Sporifica Virus

My solution for day 22: part 1, part 2

This puzzle is based on Langton’s ant which I’ve grappled with for Project Euler. I chose not to re-use any code though.

Runtime for part 2 is 31s which is just barely acceptable

Score: 2

Time: 2h47m

Leaderboard: 1,094 / 1,022

Day 23 — Coprocessor Conflagration

My solution for day 23: part 1, part 2

I had to give up on part 2, it wasn’t interesting enough to power through. Credt for part 2 in source.

Score: 1

Time: 11h31m

Leaderboard: 1,287 / 1,826

Day 24 — Electromagnetic Moat

Day 24 - complete solution

I just couldn’t get this to work so had to bail. Credit in source!

Score: 0

Time: >24h

Leaderboard: 3,057 / 3,012

Day 25 — The Halting Problem

Day 25 - complete solution

My solution runs in 30s which can surely be improved.

Score: 2 (really 1)

Time: 3h31m for part 1

Leaderboard: 1,197 / 2,340