moepgf finite field arithmetic library
======================================

The moepgf finite field arithmetic library offers high-performance operations
over GF(2), GF(4), GF(16), and GF(256) using SSE/AVX/NEON SIMD extensions on
x86/ARM platforms.

There is basic support for MIPS and RISC-V with more optimizations planned.


Installation
------------

To install moepgf you can use the following commands:

	autoreconf -fi
	./configure
	make
	make install

Make sure to have installed autoconf, automake, libtool etc.

See ./configure --help for details on the available parameters. For example you
can use --prefix=DIR to specify the installation prefix.

To check the results before installation you can issue:

	make check

If you use a git checkout you must run the following commands first:

	apt-get install automake autoconf autoconf-archive autotools-dev libtool
	autoreconf -fi


Compilers and Options
---------------------

The library requires a version of the GNU C compiler supporting AVX2 instrinsics
(x86) or NEON instrinsics (ARM). Otherwise, compilation will fail.


Documentation
-------------

All functions users should interface with are defined and documented in
includedir/moepgf/moepgf.h.


Knwon bugs
----------
- The function moepgf_init() currently accepts only GF_SELFTEST and
  GF_ALGORITHM_BEST as algorithms, i.e., either the most trivial implementation
  or the best implementation available according to the supported feature set is
  returned.  In order to get, for instance, the SSE2 implementation on a processor
  supporting AVX2, you have to use gf_get_algorithms() which returns a list with
  all algorithms.
  This should not be of concern in any practical cases since the all suboptimal
  implementations are only useful for benchmarks.


Where to get the latest
-----------------------

moepgf is available at moep80211.net.