Up to 10x faster strings for C, C++, Python, Rust, and Swift, leveraging SWAR and SIMD on Arm Neon and x86 AVX2 & AVX-512-capable chips to accelerate search, sort, edit distances, alignment scores, etc 🦖
When building with Clang it seems that there is some inconsistcy between different builds.
On WSL (Intel i5) and in the pipelines, it looks like the some of the compiler flags are not properly being set.
When I build using Clang 10.0.0 on wsl, all of the test programs were built with C++17, and none of them with AVX. (All of the built test programs were identical).
It is strange because if I build it with Clang 17.0.6 using termux on android, everything builds correctly (albeit with a few minor tweaks to get it to build), with the right compiler flags.
So I don't know if its a CMake issue or a Clang version issue (or both).
When building with Clang it seems that there is some inconsistcy between different builds.
On WSL (Intel i5) and in the pipelines, it looks like the some of the compiler flags are not properly being set.
When I build using Clang 10.0.0 on wsl, all of the test programs were built with C++17, and none of them with AVX. (All of the built test programs were identical).
It is strange because if I build it with Clang 17.0.6 using termux on android, everything builds correctly (albeit with a few minor tweaks to get it to build), with the right compiler flags.
So I don't know if its a CMake issue or a Clang version issue (or both).