Risto Viitanen

Jep, just avasin ton pääsivun ja huomasin 😅

Lumilapio-piirros norppalivessä

Lumilapio-piirros norppalivessä

Ruksi norppalivessä

Lumilapio-piirros norppalivessä

#norppalive­ssä näköjään tapahtuu tänää jotain, näyttäis 5min välein tulevan joku kuva ruutuun­­­

Norppaliven kivikko, missä yksi kivi on merkattu "Norppa (88%)"

Miten joka #norppalive­ssä on tommonen norpan muotonen kivi?!

Picture of a cat cleaning himself with a label "norppa 0.93"

Bonus Norb

First test run of this newly trained YOLOv11n-model detecting norppas (or saimaa ringed seals) seems to be working pretty well

#norppalive

Taulu, jossa on norppa kivellä ja norpan ympärillä punainen laatikko, jonka päällä lukee "norppa 0.87"

Uutta norppabottia kouluttaessa, kun tutkii vanhoja hakemistoja, löytyy välillä aarteita

#AdventOfCode day 16 done and it was pretty quick with the help of pathfinding-crate, which felt a bit like cheating

For some reason the last test fails, but I still got the correct solution

github.com/Dilaz/advent...

Day 15 of #adventofcode done.

Starting to take a bit too long to solve these puzzles so I might have to give up soon 😅

github.com/Dilaz/advent...

Day 14 of #AdventOfCode was pretty.. different. I think I got a bit lucky and guessed how the part 2 works, but it's not super fast solution, takes about 75ms to run

github.com/Dilaz/advent...

Oh, totally forgot about Rayon again; day 12 part 2 is now 7.5ms

Had a few busy days, but here are solutions for day 12 & 13 of #adventofcode

Day 12 part 2 is my slowest solution, it takes around 17ms to run but I think day 13 part 2 is one of the faster ones (51µs) if you don't count parsing the data (47µs)

github.com/Dilaz/advent...
github.com/Dilaz/advent...

I have to admit that copilot did the most of the "manual" work 😅

I've completed "Plutonian Pebbles" - Day 11 - Advent of Code 2024 #AdventOfCode adventofcode.com/2024/day/11

I tried "cached" crate, but it messed up my benchmarks, so I made my own cache.

github.com/Dilaz/advent...

Hours watched: 18000, distinct days visited: 347

Another busy year, apparently I've been watching Twitch about 52 hours per day on average!

#twitchrecap

Quickselect - Wikipedia

If you want a fun exercise with day 5 part 2, you can actually use quick select instead and just find out what is the middle number without sorting the whole list

en.wikipedia.org/wiki/Quickse...

Day 10 of #adventofcode was easier than 9. Actually due to my lazy solution, part 2 is faster than part 1. The only difference is that part 1 needs to get the unique routes and part 2 just counts them all (261µs vs. 187µs)

github.com/Dilaz/advent...

Finally got #adventofcode day 9 part 2 working, I have no idea why it was so hard for my brain.

Also changed things a bit in the project and I'll try to fix up the rest of the days to the same format to have a separate lib, main & benchmarks for each day

github.com/Dilaz/advent...

Back to microseconds in #adventofcode day 8
Spent way too many hours debugging too low result on part 1 and the issue was, that I needed '0'..='9' instead '0'..'9', which left out 9.
You get a warning for that by adding:
`"rust-analyzer.check.command": "clippy"`
to VSCode

github.com/Dilaz/advent...

Turns out my silly little optimization which caused so much trouble only skips 16/850 lines, which makes pretty much no difference

Pretty happy with #adventofcode day 7 solution. Had some issues getting the correct answer due to a premature optimization, which I didn't think through:
Who would have known that x + 1 is larger than x * 1?

But now part 1 runs in 1.8ms and part 2 in 5.8ms

github.com/Dilaz/advent...

#adventofcode day 6 done! Managed to optimize part 2 from 2.5s runtime to 125ms with the help of Rayon lib & other stuff, but I'm sure there's still something I could do to make it faster

github.com/Dilaz/advent...

Day 5 of #adventofcode done and it's at least less hacky than they day4 solution

github.com/Dilaz/advent...

Day 4 of #adventofcode is my laziest solution so far since I really dislike these kind of tasks, but at least it works and it is ⚡Blazingly fast⚡

github.com/Dilaz/advent...

Made a lazy regex solution for #adventofcode day 3 and then a faster/more optimized one with nom.

It's not like the regex solution was slow or anything, but the run time dropped from 0.004s to 0.001s

github.com/Dilaz/advent...
github.com/Dilaz/advent...

Never mind, it's 7x faster when I added Rayon optimization to the higher level loop

#adventofcode day 2 done and almost got to use the Rayon lib, but it just got slower with that 😅

github.com/Dilaz/advent...

Day 1 done in Advent of Code, let's see if I can manage to solve the first 10 puzzles this year

github.com/Dilaz/advent...

Jos oisit Dyykkis, sul ois kasa ilmasia poppareita

Tarkotit varmaan Näässtadia tai Meri-Mansee

I guess I'll try to move all my bots (@norppalivebot.bsky.social and others) to Bluesky and in the future, some new bots I've been thinking about doing