avinassh/hashevals

HashEvals is a testing tool that checks whether hash functions are actually good at their job. A hash function takes any input (a name, a file, a number) and produces a fixed-size number as output. The catch is that good hash functions need to spread outputs evenly and randomly, so small changes in the input produce completely different outputs, and no input produces the same output as another (ideally). This tool measures exactly how well various hash functions meet those standards. The tool runs three main tests on each hash function. First, it checks the "avalanche effect" by flipping individual bits in the input one at a time and counting how many output bits change, a good hash should flip roughly half the output bits for each input bit flipped. Second, it tests for collisions when hashing integers with only the last few bytes filled in, mimicking what happens when you use a hash table with integer keys. Third, it runs statistical distribution checks to spot whether some output values appear much more often than others, which would cause problems in bucketed data structures or load balancers. You'd use this if you're building something that relies heavily on hashing, a database, a cache, a distributed system, and you want to know which hash function actually performs best for your use case. The results table shows different hash functions ranked by metrics like how evenly they distribute outputs (lower bias is better) and how fast they run. The README includes results broken down by input size (tiny, short, and long chunks), since some hash functions perform differently depending on what they're hashing. The code is intentionally minimal and written in Rust, so it runs fast and the results are deterministic, you get the same numbers every time you run it with the same settings. If you want to test a new hash function, you just need to add it to the list and the tool automatically runs all three test suites against it.