bruteforces skein hashes for almamater thingo
C
Switch branches/tags
Nothing to show
Pull request Compare This branch is 19 commits ahead, 31 commits behind cemeyer:master.
Fetching latest commit…
Cannot retrieve the latest commit at this time.
Permalink
Failed to load latest commit information.
.gitignore
Makefile
README
SHA3api_ref.c
SHA3api_ref.h
block_bench.c
brg_endian.h
brg_types.h
main.c
screen-nice-wrapper.sh
skein.c
skein.h
skein_block.c
skein_block_x64.s
skein_debug.c
skein_debug.h
skein_iv.h
skein_port.h
submit.py

README

Brute-forcer for http://almamater.xkcd.com/

To run:
	make
	./screen-nice-wrapper.sh
	(executing main directly will not automatically submit solutions)

The problem is basically hashing large swaths of key space to find
near-collisions with a given bitstring. The better the 'near-ness', the higher
your rank. It is computed by taking the bitwise distance from the hash of the
input string to the given bitstring.

An ideal solver would be an implementation of the inverse skein function. Hah.

See main.c.

Skein block performance:

With GCC -O3 on this Core i7 machine, the C implementation yields:
    Trial 0 Skeins/sec: 667672.35
    Trial 1 Skeins/sec: 677591.17
    Trial 2 Skeins/sec: 693625.83
    Trial 3 Skeins/sec: 703415.01
    Trial 4 Skeins/sec: 660963.22

With the assembly implementation:
    Trial 0 Skeins/sec: 933069.85
    Trial 1 Skeins/sec: 1338020.03
    Trial 2 Skeins/sec: 996060.88
    Trial 3 Skeins/sec: 1022662.78
    Trial 4 Skeins/sec: 1141148.95

Unfortunately, the hashes produced by the assembly implementation seem
incorrect. =(