Float16 (half-precision floating-point)

Question

Float16 (half-precision floating-point)

stephentyrone opened this issue 6 years ago · 3 comments

Implement the IEEE-754 Float16 type, conforming to BinaryFloatingPoint. Use x86 and arm half<->float conversions, and armv8.2 half-precision arithmetic when available.

Answer 1 · 2019-12-12T03:19:33.000Z

Would having this type in swift-numerics vs stdlib hinder its use as a SIMD type?

Answer 2 · 2019-12-12T03:20:40.000Z

For reference:

swift evolution discussion
https://forums.swift.org/t/add-float16/19370

an implementation in progress (PR)
swiftlang/swift#21738

Answer 3 · 2020-08-03T02:04:11.000Z

Can this be closed now?