I don't know much about floats yet with Rust, beyond my old patch making them print like C's floats. Perhaps you could dig into the Float and FloatMath traits (and f64, f32) and further document what we're doing so far: http://doc.rust-lang.org/std/num/trait.Float.html (All these are still marked "experimental", by the way.)
I'm sure particularly standards-compatible (or wildly experimental) implementations would be welcome! Floating point math is *very* important for some applications, though probably with conflicting goals: speed and accuracy. When rendering a rocket exploding, speed is king, but when trying to prevent a real rocket from exploding... :D Kevin On Tue, Sep 30, 2014 at 9:56 AM, dragon lala lalo <[email protected]> wrote: > Floating Point 's Issue > > I was in a little research related to the loss of significance of the > Floating Point. > > http://en.wikipedia.org/wiki/Loss_of_significance > > After, i looked for other options to the floating point, here is one. > > "The Residue Logarithmic Number System: Theory and Implementation" > Mark G. Arnold - Lehigh University > > "Circuit Design Residue Logarithmic Number System (RLNS) Using the One-Hot > system" > > Now, because, you, people, are in the area of compilers, I have 2 doubts: > > 1) Is there some place where accuracy play a very very very important role? > > 2) Does this issue have a real importance in the development of a Compiler? > > ...or simply it's more like an if-it-works-don't-change-it thing? > > :D > > _______________________________________________ > Rust-dev mailing list > [email protected] > https://mail.mozilla.org/listinfo/rust-dev > _______________________________________________ Rust-dev mailing list [email protected] https://mail.mozilla.org/listinfo/rust-dev
