I currently have a project where I'm using SIMD64<Double>
, but have noticed that I need more precision than Double
provides. Is there a canonical way to get more precision, which is also SIMD-compatiable?
Maybe you can use Float80
? But you'll need to conform it to SIMDScalar
first.
Swift's standard library doesn't have any arbitrary precision type, but there are some 3rd party ones (e.g. attaswift/BigInt). You'll need to implement SIMDScalar
conformance for those types yourself too.
Ah too bad. I was afraid I'd have to do the SIMD-conformance myself