John Platts
John Platts
The SlideMask1Up, SlideMask1Down, SlideMaskUpLanes, and SlideMaskDownLanes operations were added as it is more efficient to slide a mask up or down by doing a left or right shift of `m.raw`...
There are some dynamic dispatch scenarios that require compiling the same C++ source files more than once (but with different C++ flags for some of the compilation phases), such as...
The upcoming C++23 standard adds support for the std::float16_t and std::bfloat16_t types, and the upcoming GCC 13 release will have support for the std::float16_t and std::bfloat16_t types with the ```-std=c++23```...
Added `.option arch, +v` to inline assembly in RVV DetectTargets to fix compilation errors with Clang on RISC-V if the `-march=rv64gcv1p0` or `-march=rv32gcv1p0` option is not specified.
Added the StreamLoad op as SSE4/AVX2/AVX3/PPC have non-temporal aligned load instructions for vectors that are 16 bytes or larger and as SVE has non-temporal load instructions for all vector sizes.
Added HWY_ASSERT checks after AllocateAligned that were previously missing in tests to ensure allocation was successful.
Updated CMakeLists.txt to build libjxl in C++17 mode if supported by the C++ compiler. For optimal dynamic dispatch with Google Highway, Clang 5 or later, GCC 7 or later, or...
Resolves issue #2183 Updated PromoteInRangeTo/DemoteInRangeTo/ConvertInRangeTo/NearestIntInRange ops that are wrappers around SSE2/AVX2/AVX3 float to integer intrinsics on x86 to use inline assembly when compiling with GCC to work around possible undefined...
Added support for dynamic dispatch on macOS/iOS/iPadOS on AArch64 as some Apple Silicon CPU's have support for the ARM BF16 extension.
Fixed compiler error with ICX as HWY_COMPILER_CLANG and HWY_COMPILER_ICX are both non-zero when compiling with icx since icx is a clang-based compiler.