Concurrent Chunk Management Engine (Fabric)

This version is uploaded automatically by GitHub Actions.

Changelog:

• e6af606a fix: use strict floating point for DoublePerlinNoiseSampler - ishland
• fe15c49e new: add toggle for native accel - ishland
• f601de7b new: add accuracy testing for DoublePerlinNoiseSampler - ishland
• cae187bf new: add accuracy testing for InterpolatedNoiseSampler - ishland
• e54650c3 fix: use strict floating point for InterpolatedNoiseSampler - ishland
• c202f109 refactor: rename InterpolatedNoiseBenchmark - ishland
• 339ae4e8 new: EndIslandsAccuracy - ishland
• daa46624 change: disable nativeaccel by default - ishland
• 2f32bee8 new: BiomeAccessAccuracy - ishland
• 7445671d perf: use int32 math rather than int8 - ishland
• 3fef2b47 perf: improve memory layout of InterpolatedNoiseSampler - ishland
• 7bfd314b fix: avoid loading native library when not needed - ishland
• 926f6e8a change: add logging before native call - ishland
• 18314d59 change: make AVX-512 usage opt-in - ishland
• 3095c814 fix: benchmark - ishland
• 9f983de4 docs: update README.md [ci skip] - ishland
• 961d22d7 Merge branch 'feature/nativeaccel-v2' into ver/1.21.1 - ishland
• 7d555732 build: get native builds quieter - ishland
• c5e9678e perf: get slightly more from parallel scheduling - ishland
• c3ea7b8c fix: nativeaccel not properly enabling itself - ishland