New major mathematical optimization giving boost of up to 14% on my n…

…Vidia GTX 1070. Number of multiplications during point addition phase has been reduced from 3 to 2. In addition to this a visible progress counter has been added during initialization.

Add debug output where important kernels are timed in milliseconds. S…

…plit main memory area for points into two separate areas for X and Y coordinates respectively to improve memory access performance. Switch from Montgomery multiplication to improved interleaved Barret reduction eliminating two modular multiplications per point when moving out of montgomery form every iteration. Add a ton of explanations to the program.

Update README to better reflect current performance

Fix contract address generation broken by comment

Major rework on Keccak implementation which together with some other …

…minor changes gives up to 4x performance boost on nVidia cards

Rework initialization so devices can be initialized in parallell.

Rework initialization so devices can be initialized in parallell.

Major overhaul. Users now have more control over memory usage on the …

…device. Overall a minor speed increase.

Add benchmark speed for a GTX 1070 to README and sort table.

Improve printed GPU index and fallback to automatic local worksize on…

… error.