r5

• Add AVX512 optimization.
• Add 32 bit float support for radius=2.

r4.1

• Add Meson build system.

r4

• Add faster SIMD code for radius 2.

r3

• Add opt parameter and code path for SSE2 and AVX2.
• Change the supported bit depth to 8, 10, 12, 14 and 16.

r2

• Use C++ casting operator instead of C-style type casting.
• Code cleanup.

r1

Initial commit