ajtribick/twofloat

A double-double arithmetic library for Rust

twofloat's Issues

Implement traits from "simba" to allow usage in nalgebra

Implementing ComplexField and RealField from the simba crate will allow usage in nalgebra (as Vector3<TwoFloat> etc).

This opens up usage in high-precision simulations, something I need for an application I am building.

support total ordering

Rust stabilized in 1.62 the total_cmp method that follows the totalOrder predicate from the IEEE754 spec (v2008, 5.10).

The same method could be added, that calculates the total order for the the high bits first and then the lower bits.

If bumping the MSRV to 1.62 is a problem, it'd be also very easy to reimplement it from scratch.

Improve `powf` implementation

The current implementation is done in terms of ln and exp. This isn't ideal, should figure out how to do it properly.

missing a const constructor

I was missing several simple constants (0, 1, -1...) and I realized it would be very easy to make a const constructor to allow defining custom constants.

I'll send a PR.

Re-enable MinGW

Should be possible to work around the FMA issues with libm

no_fma mode?

Might be useful to enable a mode that doesn't rely on FMA for multiplications. This might be necessary to address #3

Switch to GitHub Actions

Change from Travis to GitHub Actions

no_std mode?

Currently blocked by rust-lang/libm#242

Improve display

Current implementation of display is very basic, more options would be useful!

Use quad precision for function evaluation

Explore using quad-precision to improve the mathematical functions.

add `num-traits` feature

It would be very nice to have optionally implemented num-traits like ToPrimitive, FromPrimitive, Num, Float, FloatCore and Bounded, in a similar way as what the half crate does.

Add conversions to/from 128-bit integers

Add conversions for i128/u128.