From e9bf72af62fcbeb7602a533b90e67f78014c992f Mon Sep 17 00:00:00 2001 From: Kristoffer Carlsson Date: Mon, 4 Mar 2024 16:34:00 +0100 Subject: [PATCH] add deprecation notice (#107) --- README.md | 2 ++ 1 file changed, 2 insertions(+) diff --git a/README.md b/README.md index 0c41081..7c5e5f7 100644 --- a/README.md +++ b/README.md @@ -8,6 +8,8 @@ --- +# NOTE: Looking for new maintainers, otherwise deprecated in Julia 1.11 + This is a library providing basic SIMD support in Julia. VectorizationBase exists in large part to serve the needs of [LoopVectorization.jl](https://github.com/JuliaSIMD/LoopVectorization.jl)'s code gen, prioritizing this over a stable user-facing API. Thus, you may wish to consider [SIMD.jl](https://github.com/eschnett/SIMD.jl) as an alternative when writing explicit SIMD code in Julia. That said, the `Vec` and `VecUnroll` types are meant to "just work" as much as possible when passed to user-defined functions, so it should be reasonably stable in practice. Other parts of the code -- e.g, loading and storing vectors as well as the `stridedpointer` function -- will hopefully converge reasonably soon, and have support for various `AbstractArray` types propogated through the ecosystem by taking advantage of [ArrayInterface.jl](https://github.com/SciML/ArrayInterface.jl), so that VectorizationBase can begin to offer a stable, ergonomic, and well supported API fairly soon. It additionally provides some information on the host computer it is running on, which can be used to automate target-specific optimizations. Currently, x86_64 support is best on that front, but I'm looking to improve the quality of information provided for other architectures.