Re: Number of significand bits in long double?

From: Steve Kargl <sgk_at_troutmask.apl.washington.edu>
Date: Fri, 5 Aug 2005 08:58:38 -0700
On Fri, Aug 05, 2005 at 05:49:16PM +1000, Peter Jeremy wrote:
> Long double on various FreeBSD architectures is:
> Alpha:	53 bits (no hardware long double)
> ARM:    53 bits (not sure if ARM supports anything else)
> amd64:	64 bits (can be restricted to 24 or 53 bits)
> i386:	64 bits (can be restricted to 24 or 53 bits)
> iA64:   64 bits
> PPC:    53 bits (though I believe the h/w supports 106 or 112 bits)
> SPARC:  113 bits

Thanks for the info.  The code I've written should work on
all of the above with the exception of sparc.  I don't have
access to that hardware, so I won't be writing code for sparc.

> IMHO, it would be nice to run the i386 in native precision but that
> opens up a can of worms (since expressions will wind up being
> evaluated in different precisions depending on whether the compiler
> needs to spill registers onto the stack and whether temporary
> variables are registers or stack).

I'm aware of these worms.

> You probably need to have a chat to bde_at_

I have a whole mailbox full of bde emails concerning
the polynomial approximations.

-- 
Steve
Received on Fri Aug 05 2005 - 13:58:55 UTC

This archive was generated by hypermail 2.4.0 : Wed May 19 2021 - 11:38:40 UTC