Proxy & ShutUp
ShutUp
Hey, I’ve been tinkering with a new hashing algorithm that’s supposedly faster than SHA‑256. Think you can spot any weak spots?
Proxy
Looks good on the surface, but speed often comes at a cost. Make sure it has a strong avalanche effect—tiny input changes should ripple through the output. Also check collision resistance and see if the round functions mix the bits enough; weak mixing is a common backdoor. If those hold up, you might have something solid.
ShutUp
Okay, running avalanche tests, crunching collision stats, and tightening the round mixers. Will ping you with the numbers.
Proxy
Sure, just hit me with the stats. I’ll crunch them and see if there’s any pattern worth watching.
ShutUp
Avalanche: 49.6 % bits flip on average per 1-bit input change, 99.9 % of 32-bit blocks mix in every round, collision search cost roughly 2^128. No obvious patterns in test data. Ready for your crunch.
Proxy
49.6% is just shy of the 50% ideal, so it’s borderline. The 99.9% mixing is solid, and 2^128 collision cost is acceptable. Still, run a few more edge‑case tests and watch for any subtle patterns that might surface under different key sizes. If all holds, it’s probably fine.
ShutUp
Got it. Running additional edge‑case tests on 128‑bit and 256‑bit keys now, will send the expanded stats once I’m done.