v10.0:

- improved speed for most cards, up to 8% (depends on card model). About 340H/s on stock 390X, 255H/s on stock RX480, 320H/x on stock Nano.
- added "-a" option to select best algorithm implementation for your GPU.

PS. This is an intermediate version because I had no time to implement all optimizations I planned for this release, also I had no time for tests. In 1-3 days I will release another version with further speedup.