CubeHash

CubeHash is a cryptographic hash function submitted to the NIST hash function competition by Daniel J. Bernstein. CubeHash has a 128 byte state, uses wide pipe construction, and is ARX based. Message blocks are XORed into the initial bits of a 128-byte state, which then goes through an r-round bijective transformation between blocks. The initial NIST proposal ("Cubehash8/1") required about 200 cycles per byte. After clarifications from NIST, the author changed the proposal to Cubehash16/32, which "is approximately 16 times faster than CubeHash8/1, easily catching up to both SHA-256 and SHA-512 on the reference platform" while still maintaining a "comfortable security margin".

CubeHash advanced to the second round of the competition, but was not chosen as one of the 5 finalists. Bernstein has since tuned the parameters further.

This description refers to the latest specification, and not the NIST submission.

CubeHash has 5 parameters, a certain instance is denoted by CubeHashi+r/b+f-h.

In the original NIST submission, i and f was fixed to 10r. The obsolete notation CubeHashr/b-h indicates i and f being implicitly 10r.

The internal state is defined as a five-dimensional array of words (four-byte integers), 0-1 in both dimensions. The words are referred to with their coordinates [00000] to [11111]. The words are treated as little-endian.

The internal state is initialized by setting the first three words ([00000], [00001], [00010]) to h/8, b, and r respectively, all other words to zero. The state is then run through i rounds, and the initialization stage is complete. The state is now the Initialization Vector (IV). The IV can be saved and reused for a given combination of h, b, r.

The message is padded and split to b-byte blocks. The padding appends a 1 bit, followed by as many 0 bits as necessary to make a complete block.

Each block is inputed to by XORing into the first b bytes of the state, and then performing r rounds of transformation.

...
Wikipedia