mirror of
https://github.com/godotengine/godot.git
synced 2025-12-07 22:00:10 +00:00
75ce42d463
This is mostly a maintenance update that brings the compressor inline with the recently published Khronos Data Format Specification 1.4 release which clarified some ambiguity in the specification. This update also gives minor codec optimizations, bug fixes, and image quality improvements. The biggest improvement for Godot is that builds using MSVC cl.exe will now correctly default to the SSE2-optimized backend rather than the reference C backend. This makes compression more than 3 times faster. Builds using other compilers (GCC, LLVM/Clang) were not impacted by the underlying issue, and see no performance uplift. |
||
|---|---|---|
| .. | ||
| astcenc.h | ||
| astcenc_averages_and_directions.cpp | ||
| astcenc_block_sizes.cpp | ||
| astcenc_color_quantize.cpp | ||
| astcenc_color_unquantize.cpp | ||
| astcenc_compress_symbolic.cpp | ||
| astcenc_compute_variance.cpp | ||
| astcenc_decompress_symbolic.cpp | ||
| astcenc_diagnostic_trace.cpp | ||
| astcenc_diagnostic_trace.h | ||
| astcenc_entry.cpp | ||
| astcenc_find_best_partitioning.cpp | ||
| astcenc_ideal_endpoints_and_weights.cpp | ||
| astcenc_image.cpp | ||
| astcenc_integer_sequence.cpp | ||
| astcenc_internal.h | ||
| astcenc_internal_entry.h | ||
| astcenc_mathlib.cpp | ||
| astcenc_mathlib.h | ||
| astcenc_mathlib_softfloat.cpp | ||
| astcenc_partition_tables.cpp | ||
| astcenc_percentile_tables.cpp | ||
| astcenc_pick_best_endpoint_format.cpp | ||
| astcenc_quantization.cpp | ||
| astcenc_symbolic_physical.cpp | ||
| astcenc_vecmathlib.h | ||
| astcenc_vecmathlib_avx2_8.h | ||
| astcenc_vecmathlib_common_4.h | ||
| astcenc_vecmathlib_neon_4.h | ||
| astcenc_vecmathlib_none_4.h | ||
| astcenc_vecmathlib_sse_4.h | ||
| astcenc_vecmathlib_sve_8.h | ||
| astcenc_weight_align.cpp | ||
| astcenc_weight_quant_xfer_tables.cpp | ||
| LICENSE.txt | ||