Default Windows build fails due to the new DEBUG_OPT.
When not passing this option, build fails with the following issues:

Cannot add commands to previously defined target ..
Cannot add commands to previously defined target DEBUG_OPT.
Cannot add commands to previously defined target =.

Setting default value for DEBUG_OPT earlier fixes the issue.

Add new compilation flag DEBUG_OPT to modify the optimization
level when debugging (DEBUG=y).

Example: make DEBUG=y DEBUG_OPT=-O3

- poly-1305 hash algorithm added
- fixed GMAC-128/192/256 hash algorithms names
- added chacha20-poly1305 to standard test combinations

- generic x86_64 implementation added
- poly1305 fixed vector tests added

Updated test applications to use the new enums
when iterating through all cipher and hash algorithms

The 16x8-bit parallel lookup functions have two parameters:
a pointer to the LUT and __m128i containing the indices.
On Linux, this __m128i is passed through xmm0, but on Windows,
this is stored in the stack and a pointer to this address
is passed through a GP register.
Therefore, this memory must be read into xmm0, so it is compatible
with the Linux implementation.

This commit fixes issue #59.

Optimize the handling of the last block, less than 16 bytes,
by using mask registers, instead of branching depending on the
number of bytes left.
Also, include the rotation and XOR with the last keystream block.

Separate ZUC initialization from the authentication function,
so it can be called from job submit/flush function.
This way, only the state for new buffers will be initialized.

This commit also simplifies the flush logic in EEA3 and fixes
an issue when using bsf instruction to get the first bit set.

Add AVX512 versions of Round64B and Remainder functions
which are used for ZUC-EIA3.

The digest was being calculated skipping the first 16 bytes
of the keystream for each single buffer, instead of using the whole keystream.

Length to hash in ZUC-EIA3 is passed in bits and not in bytes,
so the length from the job size/s must be multiplied by 8.

Deprecated --auto-detect option. By default, the application
will now test all interfaces supported by the processor.