Subject: Re: More ELF stuff
To: Stephen Borrill <netbsd@brio.precedence.co.uk>
From: Richard Earnshaw <rearnsha@arm.com>
List: port-arm26
Date: 02/15/2001 11:00:11
> On Wed, 14 Feb 2001, Richard Earnshaw wrote:
> > Even with existing processors, only the ARM7500FE has been 
> > shipped with an FPA in significant numbers, and on netbsd we still don't 
> > make a use of it.
> >
> > So, to the question.  If I knock up a hack to the compiler and assembler 
> > so that for NetBSD/ARM/ELF they change to a pure little-endian floating 
> > point format, would there be any objections?
> 
> I guess that depends on whether there is ever likely to be support for the
> FPA in the ARM7500FE. If the answer is no, then do it. However, I would
> expect a speed up for X11 with an FPA and I'm quite interested in using
> things like NCs as X terminals. They are not overendowed with processing
> power and so any gain would be useful.
> 

Well, is there?  It would only happen if someone volunteered to do it.

IIRC most of the float code in X is single-precision.  It might be 
possible to knock up a version of the compiler that could just use the FPA 
for single-precision operations and revert to libcalls for doubles.  This 
would then make X operations much more efficient while still allowing us 
to switch to the new word endian format.

R.