This is the mail archive of the
gcc-patches@gcc.gnu.org
mailing list for the GCC project.
Re: [PATCH, powerpc] Improve integer to floating point conversions on powerpc
- From: Michael Meissner <meissner at linux dot vnet dot ibm dot com>
- To: Nathan Froyd <froydnj at codesourcery dot com>
- Cc: Michael Meissner <meissner at linux dot vnet dot ibm dot com>, gcc-patches at gcc dot gnu dot org, dje dot gcc at gmail dot com
- Date: Tue, 24 Aug 2010 15:45:22 -0400
- Subject: Re: [PATCH, powerpc] Improve integer to floating point conversions on powerpc
- References: <20100824145857.GA6425@hungry-tiger.westford.ibm.com> <20100824190345.GL25394@codesourcery.com>
On Tue, Aug 24, 2010 at 12:03:45PM -0700, Nathan Froyd wrote:
> On Tue, Aug 24, 2010 at 10:58:57AM -0400, Michael Meissner wrote:
> > I built an powerpc-eabispe compiler, and verified that the compiler with the
> > patches generates exactly the same code as the compiler without the patches for
> > -mcpu=e500mc and -mcpu=e500mc64.
>
> This is kind of an odd combination, since the e500mc and the e500mc64
> use "classic" floating-point, i.e. the same model as POWER* chips. The
> more interesting combination would be a powerpc-eabispe compiler with:
>
> - -mcpu=8540 -mfloat-gprs=single (hard single, soft double)
> - -mcpu=8548 -mfloat-gprs=double (hard single and double)
>
> since that will exercise the E500 code paths a bit more. You may also
> want to check your new FP testcases with such a toolchain. (Though I
> think they will still pass, due to the explicit -mcpu.)
Except for the .ident strings, they produce the same code for both sets of
options. Thanks for pointing me it the right direction of options to use.
--
Michael Meissner, IBM
5 Technology Place Drive, M/S 2757, Westford, MA 01886-3141, USA
meissner@linux.vnet.ibm.com