0.28.0

This version is thanks to Prof Osamu Watanabe.
The performance of encoding and decoding, using AVX2, has improved significantly, as shown in the table below.

What's Changed

• AVX2 block coder optimizations (encoder + decoder) by @osamu620 in #276

Workload (4K, reversible 5/3 unless noted)	Encode	Decode
8-bit, 64×64 code-blocks	1.91×	1.19×
8-bit, 512×8 code-blocks	1.94×	1.17×
8-bit, lossy (qstep 0.005)	1.49×	1.15×
16-bit, 64×64 code-blocks	1.80×	1.29×

New Contributors

• @osamu620 made their first contribution in #276

Full Changelog: 0.27.4...0.28.0