LZX pre/post-processing improvements
authorEric Biggers <ebiggers3@gmail.com>
Thu, 25 Dec 2014 01:58:04 +0000 (19:58 -0600)
committerEric Biggers <ebiggers3@gmail.com>
Thu, 25 Dec 2014 02:15:13 +0000 (20:15 -0600)
commit3872f9e0d30f6439b2a08e091fbd3df042841b6a
tree967d91b257bf7001446c7e223791b9eb6e1cc95f
parent41065b5e5a884313d438797435a922b20dac9bce
LZX pre/post-processing improvements

- much faster inner loop for non-vectorized version
- use either SSE2 or AVX-2 for vectorized version
- faster inner loop for vectorized version
- support vectorized version on buffers that are not properly aligned
src/lzx-common.c