I have long wanted to see a simplified "core" subset of the IEEE 754 floating point standard. Can we make it happen?
I blogged about my thoughts here: https://www.bitsnbites.eu/ieee-754-suggestion-a-core-subset/ …
#ieee754 #floatingpoint
-
-
I think that -0 and +/- infinity have value for propagating information. E.g. after a series of computations like -42/INF=-0 ... 7/-0=-INF, the final result can safely be used in a less/greater than comparison & branch. Plus -0, INF & NaN are cheap in hw.
Thanks. Twitter will use this to make your timeline better. UndoUndo
-
-
-
Maybe we can implement a subset and pick options which optimize for speed and power/area.
Thanks. Twitter will use this to make your timeline better. UndoUndo
-
Loading seems to be taking a while.
Twitter may be over capacity or experiencing a momentary hiccup. Try again or visit Twitter Status for more information.