abrown commented on Issue #2898:
There are other considerations to this:
- perhaps the arm32 and aarch64 backends could be simplified now that they don't have to lower the scalar versions of these operations
- we could alternately implement the operations for x64 but I don't particularly see the need at the moment: the lowering is not optimal and these operations were introduced to implement the Wasm SIMD specification.
bjorn3 commented on Issue #2898:
They could be useful to lower the saturating arithmetic intrinsics of rust to.
Last updated: Nov 22 2024 at 16:03 UTC