This is the mail archive of the gcc-help@gcc.gnu.org mailing list for the GCC project.

Index Nav:	[Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav:	[Date Prev] [Date Next]	[Thread Prev] [Thread Next]
Other format:	[Raw text]

Re: PowerPC long double

From: Ira Hill <ijh3 at ufl dot edu>
To: Ian Lance Taylor <iant at google dot com>
Cc: gcc-help at gcc dot gnu dot org
Date: Sat, 08 Sep 2012 04:40:00 -0400
Subject: Re: PowerPC long double
References: <504A462C.3090900@ufl.edu> <CAKOQZ8xBbx_J8xWWcm6z1LoX12uSvUKju3rGaXWyqg--doZ5Fw@mail.gmail.com>

On 09/07/2012 03:55 PM, Ian Lance Taylor wrote:

On Fri, Sep 7, 2012 at 12:08 PM, Ira Hill<ijh3@ufl.edu> wrote:
Now I also can enable the "-mlong-double-128" flag which appears to give me
quad precision, but that produces binaries with undefined symbols on my
target (_q_mul, _q_add, etc..) and will not run.
I'm not certain, but I think you need to use the -mno-xl-compat option.

Ian

Thanks for the reply! Unfortunately, -mno-xl-compat still produces the _q_mul symbols. "powerpc-wrs-vxworks-objdump -t PPC603gnu/testGCC.o | grep _q" 00000000 *UND* 00000000 _q_flt 00000000 *UND* 00000000 _q_fge 00000000 *UND* 00000000 _q_fgt 00000000 *UND* 00000000 _q_fle 00000000 *UND* 00000000 _q_div 00000000 *UND* 00000000 _q_mul

I've found another thread where a user was able to express long double as two doubles (__LBL_MANT_DIG__106) on PowerPC-linux: http://gcc.gnu.org/ml/gcc-help/2011-05/msg00200.html. I'm not sure why I'm not able to do the same (is it VxWorks specific?). Ultimately, I need higher precision for one of our robotic algorithms.

Thanks,
Ira

References:
- PowerPC long double
  - From: Ira Hill
- Re: PowerPC long double
  - From: Ian Lance Taylor

Index Nav:	[Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav:	[Date Prev] [Date Next]	[Thread Prev] [Thread Next]