[racket-dev] flonum vs. inexact-real

From: Matthew Flatt (mflatt at cs.utah.edu)
Date: Sun Oct 3 07:43:52 EDT 2010

At Sun, 3 Oct 2010 11:17:44 +0100, Noel Welsh wrote:
> On Sat, Oct 2, 2010 at 7:48 PM, Matthew Flatt <mflatt at cs.utah.edu> wrote:
> > With the current memory manager, I don't think there's any potential space
> > gain from using 32-bit floats instead of 64-bit floats. Is there any
> > other reason to use 32-bit floating point?
> 
> In theory one can get better performance with floats. For example, you
> can pack more floats than doubles into the SIMD registers on Intel
> chips.

So far, I'm leaning toward splitting "flonum" from "inexact real" and
maybe even enabling 32-bit float support by default.

Sam and Vincent: Any thoughts on how easy or difficult the change would
be for Typed Racket (and its optimizer)?



Posted on the dev mailing list.