I'm diving into Advent of Code 2025 using PowerShell (version 7.5.4 on macOS), after previously working with Ruby. Sometimes, I hit walls when the math gets tricky or the computations become too complex for my usual brute-force methods. I recently started reworking 2015's challenges in PowerShell and got stuck on day 4, where I need to find a string from the puzzle input and an integer that generates an MD5 hash starting with '00000' (five zeros). I'm experimenting with parallel execution here—trying to run 10 threads, each handling 10,000 numbers (for example, 300,000-309,999, then 310,000-319,999, etc.). Is this approach effective? Also, what's the difference between using Start-ThreadJob and ForEach-Object -Parallel? Lastly, does the ThrottleLimit setting relate to the number of CPU cores? Any insights would be greatly appreciated!
3 Answers
Yes, the ThrottleLimit does relate to the number of cores, but indirectly. It controls the number of runspaces, which correspond to threads, and those threads get managed by .NET and the OS. If you set it too high, you might end up queuing due to a bottleneck with the cores. For CPU-intensive tasks, it's usually best to set it to one less than your total core count, so you keep one core available for managing processes.
For that particular puzzle, I think brute force is the way to go. If you mess around with the input to the MD5, you could end up undermining the whole point of the hash function. Writing the output to a file and calculating the hash can slow you down due to file I/O overhead. Plus, using tools like md5sum could be bottlenecked by antivirus scans. Have you considered using C# for the MD5 calculations? It can be seamlessly integrated into your PowerShell script to speed things up. Here's a quick script you can use: ... Just tweak your conditions a bit to optimize for part two.
You can tackle all Advent of Code puzzles without resorting to multithreading or any specific language features. In general, most solutions can be found in under a minute! Personally, I just tackled the problem in PowerShell and got the first part done in less than 2 minutes, so you should be good with that.
That's right! In scenarios with slow I/O, like network requests, you could afford a higher throttle limit. But when dealing with CPU-heavy computations, keeping it a bit lower will help in not overwhelming the system.