diff --git a/docs/source/backend/reduction_codegen.md b/docs/source/backend/reduction_codegen.md index 360c692568ada1147b99ef7c829ab6cdc9af02fa..f08fa980a9050cacd7bbccad505cadde350dcd9c 100644 --- a/docs/source/backend/reduction_codegen.md +++ b/docs/source/backend/reduction_codegen.md @@ -114,9 +114,3 @@ You can find more detailed information about warp size alignment in {ref}`gpu_co ps.inspect(kernel_gpu_opt) ``` -:::{admonition} Developers To Do: - -- Support for HIP platforms -- Support vectorization using NEON intrinsics -::: - diff --git a/docs/source/user_manual/reductions.md b/docs/source/user_manual/reductions.md index 6af0e8580f215f8b205e3250e65ec482870b4523..0b0c04279e34f0919d69beff83ec184f5c0cef61 100644 --- a/docs/source/user_manual/reductions.md +++ b/docs/source/user_manual/reductions.md @@ -122,8 +122,3 @@ Since our reduction result is a single scalar value, it is sufficient to set up For GPU platforms, the concepts remain the same but the fields and the write-back pointer now require device memory, i.e. instances of {any}`cupy.ndarray`. - -:::{admonition} Developers To Do: - -- Support for higher-order data types for reductions, e.g. vector/matrix reductions -:::