Closed Piedone closed 6 years ago
Standard IEEE double support is costly. Something else? Maybe https://github.com/Lombiq/Hastlayer-SDK/issues/21.
We could even use this library: https://github.com/CodesInChaos/SoftFloat/ It's mostly transformable directly.
Going with #21. SoftFloat is very old, abandoned and lacks an implementation for division. We could add some other soft float implementation in the future to compare posits to.
Standard IEEE double support is costly. Something else? Maybe https://github.com/Lombiq/Hastlayer-SDK/issues/21.
We could even use this library: https://github.com/CodesInChaos/SoftFloat/ It's mostly transformable directly.