chickenstrips

4 ms·

1. 1.▲
   An implementation of the differential attack on IOTA's Curl hashing function(github.com)

   1 points by chickenstrips 8y ago|0 comments
2. 2.▲
   A full implementation of the attack on IOTA's hash function(github.com)

   3 points by chickenstrips 8y ago|0 comments
3. 3.▲
   by chickenstrips 8y ago

   You're correct. The correspondence just seems to be due to some bound on the prime gap for arithmetic progressions (or perhaps something even more trivial that I'm missing). A cursory search suggests it might be a consequence of I
4. 4.▲
   by chickenstrips 8y ago

   For fixed n, a block (x,y) contains the numbers n/2 y^2-n/2 y + yz + x for all 0 <= z < n.
5. 5.▲
   by chickenstrips 8y ago

   It also doesn't reduce the time to O(1). Each of the ranges is of size O(2^(N/2)) for an N bit prime, so it's really not useful at all.
6. 6.▲
   by chickenstrips 8y ago

   No, this won't help with finding primes. As they noted, the pattern for ranges of size k only holds for k lines. So to find a prime of length N (on the order of 2^N), we need to have k=O(2^(N/2)). However, this yields a guarantee