Advent of Code - Day 19

Learning Questions & Discussions Coding Challenges
,
4

I got stuck but after reading your replies I tried to also optimize the loop and I arrived to the same conclusion as @sasajuric: part 2 is basically computing the sum of all divisors for a given number, with some busy wait thrown in for good measure, haha.

However, I have optimized it slightly differently. If I trace the results I get this:

{{1, 0, 0, 0, 0, 0}, :addi, 1, 16, 1, {1, 16, 0, 0, 0, 0}}
{{1, 17, 0, 0, 0, 0}, :addi, 3, 2, 3, {1, 17, 0, 2, 0, 0}}
{{1, 18, 0, 2, 0, 0}, :mulr, 3, 3, 3, {1, 18, 0, 4, 0, 0}}
{{1, 19, 0, 4, 0, 0}, :mulr, 1, 3, 3, {1, 19, 0, 76, 0, 0}}
{{1, 20, 0, 76, 0, 0}, :muli, 3, 11, 3, {1, 20, 0, 836, 0, 0}}
{{1, 21, 0, 836, 0, 0}, :addi, 4, 3, 4, {1, 21, 0, 836, 3, 0}}
{{1, 22, 0, 836, 3, 0}, :mulr, 4, 1, 4, {1, 22, 0, 836, 66, 0}}
{{1, 23, 0, 836, 66, 0}, :addi, 4, 18, 4, {1, 23, 0, 836, 84, 0}}
{{1, 24, 0, 836, 84, 0}, :addr, 3, 4, 3, {1, 24, 0, 920, 84, 0}}
{{1, 25, 0, 920, 84, 0}, :addr, 1, 0, 1, {1, 26, 0, 920, 84, 0}}
{{1, 27, 0, 920, 84, 0}, :setr, 1, 4, 4, {1, 27, 0, 920, 27, 0}}
{{1, 28, 0, 920, 27, 0}, :mulr, 4, 1, 4, {1, 28, 0, 920, 756, 0}}
{{1, 29, 0, 920, 756, 0}, :addr, 1, 4, 4, {1, 29, 0, 920, 785, 0}}
{{1, 30, 0, 920, 785, 0}, :mulr, 1, 4, 4, {1, 30, 0, 920, 23550, 0}}
{{1, 31, 0, 920, 23550, 0}, :muli, 4, 14, 4, {1, 31, 0, 920, 329700, 0}}
{{1, 32, 0, 920, 329700, 0}, :mulr, 4, 1, 4, {1, 32, 0, 920, 10550400, 0}}
{{1, 33, 0, 920, 10550400, 0}, :addr, 3, 4, 3, {1, 33, 0, 10551320, 10550400, 0}}
{{1, 34, 0, 10551320, 10550400, 0}, :seti, 0, 0, 0, {0, 34, 0, 10551320, 10550400, 0}}
{{0, 35, 0, 10551320, 10550400, 0}, :seti, 0, 1, 1, {0, 0, 0, 10551320, 10550400, 0}}
{{0, 1, 0, 10551320, 10550400, 0}, :seti, 1, 4, 5, {0, 1, 0, 10551320, 10550400, 1}}
{{0, 2, 0, 10551320, 10550400, 1}, :seti, 1, 4, 2, {0, 2, 1, 10551320, 10550400, 1}}
{{0, 3, 1, 10551320, 10550400, 1}, :mulr, 5, 2, 4, {0, 3, 1, 10551320, 1, 1}}

My instruction point is in the 2nd register and I noticed that when it has value 1, it is when the loop starts, so I just need to get the number at position 4th in the registry and compute the sum of its divisors. So I added this instruction:

  # Specialized instruction for part 2.
  # The problem is basically computing the sum of all numbers in register 3.
  defp compute(_ip, _instructions, {_, 1, 0, number, _, _}) do
    Enum.sum(for i <- 1..number, rem(number, i) == 0, do: i)
  end

Both parts finish in 0.6 seconds. :slight_smile: This is by far the puzzle that took me the longest.

2 Likes