cgal/Interval_arithmetic/include/CGAL/FPU.h

429 lines
13 KiB
C++

// Copyright (c) 1998-2004 Utrecht University (The Netherlands),
// ETH Zurich (Switzerland), Freie Universitaet Berlin (Germany),
// INRIA Sophia-Antipolis (France), Martin-Luther-University Halle-Wittenberg
// (Germany), Max-Planck-Institute Saarbruecken (Germany), RISC Linz (Austria),
// and Tel-Aviv University (Israel). All rights reserved.
//
// This file is part of CGAL (www.cgal.org); you can redistribute it and/or
// modify it under the terms of the GNU Lesser General Public License as
// published by the Free Software Foundation; version 2.1 of the License.
// See the file LICENSE.LGPL distributed with CGAL.
//
// Licensees holding a valid commercial license may use this file in
// accordance with the commercial license agreement provided with the software.
//
// This file is provided AS IS with NO WARRANTY OF ANY KIND, INCLUDING THE
// WARRANTY OF DESIGN, MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE.
//
// $URL$
// $Id$
//
//
// Author(s) : Sylvain Pion
#ifndef CGAL_FPU_H
#define CGAL_FPU_H
// This file specifies some platform dependant functions, regarding the FPU
// directed rounding modes. There is only support for double precision.
//
// It also contains the definition of the Protect_FPU_rounding<> class,
// which helps to protect blocks of code needing a particular rounding mode.
#if defined __alpha__ && defined __linux__
extern "C" {
# include <fenv.h>
}
#elif defined __linux__ && !defined __PGI && !defined __ia64__ \
&& !defined __x86_64__
# include <fpu_control.h>
#elif defined __SUNPRO_CC || (defined __KCC && defined __sun)
# include <ieeefp.h>
#elif defined __osf || defined __osf__
# ifdef __GNUG__
// GCC seems to remove (fixincludes) read_rnd/write_rnd...
# include "/usr/include/float.h"
# else
# include <float.h>
# endif
#elif defined __BORLANDC__
# include <float.h>
#elif defined __sgi
# include <sys/fpu.h>
#elif defined _MSC_VER || defined __sparc__ || \
(defined __i386__ && !defined __PGI)
// Nothing to include.
#else
// By default we use the ISO C99 version.
# include <fenv.h>
#endif
// GCC 3.0.0 has some bugs, which can be worked around, but it's
// not worth maintaining them anymore.
#if defined __GNUG__ && !defined __INTEL_COMPILER && \
(__GNUG__ == 3 && __GNUC_MINOR__ == 0 && __GNUC_PATCHLEVEL__ == 0)
# error GCC 3.0.0 is buggy, use at your own risk.
#endif
// Some useful constants
#if defined CGAL_CFG_NO_LIMITS
# if defined CGAL_CFG_DENORMALS_COMPILE_BUG
// For compilers crashing when dealing with denormalized values.
// So we have to generate it at run time instead.
# define CGAL_IA_MIN_DOUBLE (CGAL::CGALi::minimin)
# else
# define CGAL_IA_MIN_DOUBLE (5e-324)
# endif
# define CGAL_IA_MAX_DOUBLE (1.7976931348623157081e+308)
#else
# include <limits>
# define CGAL_IA_MIN_DOUBLE std::numeric_limits<double>::denorm_min()
# define CGAL_IA_MAX_DOUBLE std::numeric_limits<double>::max()
#endif
// Pure and safe SSE2 mode (g++ -mfpmath=sse && (-msse2 || -march=pentium4))
// can be detected by :
// TODO : see what Intel and VC++ have to say about this.
#if defined __FLT_EVAL_METHOD__ && defined __SSE2_MATH__ && \
(__FLT_EVAL_METHOD__ == 0 || __FLT_EVAL_METHOD__ == 1)
# define CGAL_SAFE_SSE2
# include <xmmintrin.h>
#endif
// We do not handle -mfpmath=387,sse yet.
#if defined __SSE2_MATH__ && \
! (__FLT_EVAL_METHOD__ == 0 || __FLT_EVAL_METHOD__ == 1)
# warning Unsafe SSE2 mode : not supported yet.
#endif
CGAL_BEGIN_NAMESPACE
namespace CGALi {
#ifdef CGAL_CFG_DENORMALS_COMPILE_BUG
extern double minimin;
#endif
#ifdef __INTEL_COMPILER
const double infinity = std::numeric_limits<double>::infinity();
#else
const double infinity = HUGE_VAL;
#endif
} // namespace CGALi
// Inline function to stop compiler optimization.
inline double IA_force_to_double(double x)
{
#if defined __GNUG__ && !defined __INTEL_COMPILER
// Intel does not emulate GCC perfectly...
asm("" : "=m"(x) : "m"(x));
// asm("" : "+m"(x) );
return x;
#else
volatile double e = x;
return e;
#endif
}
// The x86 processor keeps too wide exponents (15bits) in registers, even in
// double precision mode ! It's a problem when the intervals overflow or
// underflow. To work around that, we store the variable to memory when
// needed, using the macro below.
// Another possible workaround would be to use intervals of "long doubles"
// directly, but I think it would be much slower.
#if !defined (CGAL_IA_NO_X86_OVER_UNDER_FLOW_PROTECT) && \
((defined __i386__ && !defined CGAL_SAFE_SSE2) \
|| defined _MSC_VER || defined __BORLANDC__)
# define CGAL_IA_FORCE_TO_DOUBLE(x) CGAL::IA_force_to_double(x)
#else
# define CGAL_IA_FORCE_TO_DOUBLE(x) (x)
#endif
// We sometimes need to stop constant propagation,
// because operations are done with a wrong rounding mode at compile time.
// G++ also uses __builtin_constant_p().
#ifndef CGAL_IA_DONT_STOP_CONSTANT_PROPAGATION
# if defined __GNUG__ && __GNUG__ < 3
// Note : GCC 3 doesn't guarantee __builtin_constant_p to return false
// when he will not do cprop :(.
# define CGAL_IA_STOP_CPROP(x) \
(__builtin_constant_p (x) ? CGAL::IA_force_to_double(x) : (x) )
# define CGAL_IA_STOP_CPROP2(x,y) \
(__builtin_constant_p (y) ? CGAL_IA_STOP_CPROP(x) : (x) )
# else
# define CGAL_IA_STOP_CPROP(x) CGAL::IA_force_to_double(x)
# define CGAL_IA_STOP_CPROP2(x,y) CGAL::IA_force_to_double(x)
# endif
#else
# define CGAL_IA_STOP_CPROP(x) (x)
# define CGAL_IA_STOP_CPROP2(x,y) (x)
#endif
// std::sqrt(double) on VC++ and CygWin is buggy when not optimizing.
#ifdef _MSC_VER
inline double IA_bug_sqrt(double d)
{
_asm
{
fld d
fsqrt
fstp d
}
return d;
}
# define CGAL_BUG_SQRT(d) CGAL::IA_bug_sqrt(d)
#elif defined __SSE2_MATH__
// For SSE2, we need to call __builtin_sqrt() instead of libc's sqrt().
# define CGAL_BUG_SQRT(d) __builtin_sqrt(d)
#elif defined __CYGWIN__
inline double IA_bug_sqrt(double d)
{
double r;
asm volatile ("fsqrt" : "=t"(r) : "0"(d));
return r;
}
# define CGAL_BUG_SQRT(d) CGAL::IA_bug_sqrt(d)
#else
# define CGAL_BUG_SQRT(d) CGAL_CLIB_STD::sqrt(d)
#endif
// Here are the operator macros that make use of the above.
// With GCC, we can do slightly better : test with __builtin_constant_p()
// that both arguments are constant before stopping one of them.
// Use inline functions instead ?
#define CGAL_IA_ADD(a,b) CGAL_IA_FORCE_TO_DOUBLE((a)+CGAL_IA_STOP_CPROP2(b,a))
#define CGAL_IA_SUB(a,b) CGAL_IA_FORCE_TO_DOUBLE((a)-CGAL_IA_STOP_CPROP2(b,a))
#define CGAL_IA_MUL(a,b) CGAL_IA_FORCE_TO_DOUBLE((a)*CGAL_IA_STOP_CPROP2(b,a))
#define CGAL_IA_DIV(a,b) CGAL_IA_FORCE_TO_DOUBLE((a)/CGAL_IA_STOP_CPROP2(b,a))
#define CGAL_IA_SQUARE(a) CGAL_IA_MUL(a,a)
#define CGAL_IA_SQRT(a) \
CGAL_IA_FORCE_TO_DOUBLE(CGAL_BUG_SQRT(CGAL_IA_STOP_CPROP(a)))
#if defined __i386__ && !defined __PGI
# if defined CGAL_SAFE_SSE2
#define CGAL_IA_SETFPCW(CW) _MM_SET_ROUNDING_MODE(CW)
#define CGAL_IA_GETFPCW(CW) CW = _MM_GET_ROUNDING_MODE()
typedef unsigned int FPU_CW_t;
#define CGAL_FE_TONEAREST _MM_ROUND_NEAREST
#define CGAL_FE_TOWARDZERO _MM_ROUND_TOWARD_ZERO
#define CGAL_FE_UPWARD _MM_ROUND_UP
#define CGAL_FE_DOWNWARD _MM_ROUND_DOWN
# else
// The GNU libc version (cf powerpc) is nicer, but doesn't work on libc 5 :(
// This one also works with CygWin.
// Note that the ISO C99 version is not enough because of the extended
// mantissa issue on x86 (required by Fixed_precision_nt, modular computations
// in the future, but not IA right now).
#define CGAL_IA_SETFPCW(CW) asm volatile ("fldcw %0" : :"m" (CW))
#define CGAL_IA_GETFPCW(CW) asm volatile ("fnstcw %0" : "=m" (CW))
typedef unsigned short FPU_CW_t;
#define CGAL_FE_TONEAREST (0x000 | 0x127f)
#define CGAL_FE_TOWARDZERO (0xc00 | 0x127f)
#define CGAL_FE_UPWARD (0x800 | 0x127f)
#define CGAL_FE_DOWNWARD (0x400 | 0x127f)
# endif
#elif defined __powerpc__
#define CGAL_IA_SETFPCW(CW) _FPU_SETCW(CW)
#define CGAL_IA_GETFPCW(CW) _FPU_GETCW(CW)
typedef fpu_control_t FPU_CW_t;
#define CGAL_FE_TONEAREST (_FPU_RC_NEAREST | _FPU_DEFAULT)
#define CGAL_FE_TOWARDZERO (_FPU_RC_ZERO | _FPU_DEFAULT)
#define CGAL_FE_UPWARD (_FPU_RC_UP | _FPU_DEFAULT)
#define CGAL_FE_DOWNWARD (_FPU_RC_DOWN | _FPU_DEFAULT)
#elif defined __SUNPRO_CC || (defined __KCC && defined __sun)
#define CGAL_IA_SETFPCW(CW) fpsetround(fp_rnd(CW))
#define CGAL_IA_GETFPCW(CW) CW = fpgetround()
typedef unsigned int FPU_CW_t;
#define CGAL_FE_TONEAREST FP_RN
#define CGAL_FE_TOWARDZERO FP_RZ
#define CGAL_FE_UPWARD FP_RP
#define CGAL_FE_DOWNWARD FP_RM
#elif defined __sparc__
#define CGAL_IA_SETFPCW(CW) asm volatile ("ld %0,%%fsr" : :"m" (CW))
#define CGAL_IA_GETFPCW(CW) asm volatile ("st %%fsr,%0" : "=m" (CW))
typedef unsigned int FPU_CW_t;
#define CGAL_FE_TONEAREST (0x0 | 0x20000000 | 0x1f)
#define CGAL_FE_TOWARDZERO (0x40000000 | 0x20000000 | 0x1f)
#define CGAL_FE_UPWARD (0x80000000 | 0x20000000 | 0x1f)
#define CGAL_FE_DOWNWARD (0xc0000000 | 0x20000000 | 0x1f)
#elif defined __sgi
typedef unsigned int FPU_CW_t;
inline FPU_CW_t sgi_get_fpu_cw()
{
fpc_csr csr;
csr.fc_word = get_fpc_csr();
return csr.fc_struct.rounding_mode;
}
inline void sgi_set_fpu_cw(FPU_CW_t cw)
{
fpc_csr csr;
csr.fc_word = get_fpc_csr();
csr.fc_struct.rounding_mode = cw;
csr.fc_struct.flush = 0; // By default, denormals are flushed to zero !
set_fpc_csr(csr.fc_word);
}
#define CGAL_IA_SETFPCW(CW) sgi_set_fpu_cw(CW)
#define CGAL_IA_GETFPCW(CW) CW = sgi_get_fpu_cw()
#define CGAL_FE_TONEAREST ROUND_TO_NEAREST
#define CGAL_FE_TOWARDZERO ROUND_TO_ZERO
#define CGAL_FE_UPWARD ROUND_TO_PLUS_INFINITY
#define CGAL_FE_DOWNWARD ROUND_TO_MINUS_INFINITY
#elif defined __mips__ // && !defined __sgi
#define CGAL_IA_SETFPCW(CW) asm volatile ("ctc1 %0,$31" : :"r" (CW))
#define CGAL_IA_GETFPCW(CW) asm volatile ("cfc1 %0,$31" : "=r" (CW))
typedef unsigned int FPU_CW_t;
#define CGAL_FE_TONEAREST (0x0)
#define CGAL_FE_TOWARDZERO (0x1)
#define CGAL_FE_UPWARD (0x2)
#define CGAL_FE_DOWNWARD (0x3)
#elif defined __osf || defined __osf__ // Not yet supported.
#define CGAL_IA_SETFPCW(CW) write_rnd(CW)
#define CGAL_IA_GETFPCW(CW) CW = read_rnd()
typedef unsigned int FPU_CW_t;
#define CGAL_FE_TONEAREST FP_RND_RN
#define CGAL_FE_TOWARDZERO FP_RND_RZ
#define CGAL_FE_UPWARD FP_RND_RP
#define CGAL_FE_DOWNWARD FP_RND_RM
#elif defined __alpha__ // preliminary support.
#define CGAL_IA_SETFPCW(CW) (__ieee_set_fp_control(CW))
#define CGAL_IA_GETFPCW(CW) (CW = __ieee_get_fp_control())
typedef unsigned long FPU_CW_t;
#define CGAL_FE_TONEAREST FE_TONEAREST
#define CGAL_FE_TOWARDZERO FE_TOWARDZERO
#define CGAL_FE_UPWARD FE_UPWARD
#define CGAL_FE_DOWNWARD FE_DOWNWARD
#elif defined _MSC_VER
// Found in http://msdn.microsoft.com/library/sdkdoc/directx/imover_7410.htm :
#define CGAL_IA_SETFPCW(CW) __asm fldcw (CW)
#define CGAL_IA_GETFPCW(CW) __asm fstcw (CW)
typedef unsigned short FPU_CW_t;
#define CGAL_FE_TONEAREST (0x0 | 0x127f)
#define CGAL_FE_TOWARDZERO (0xC00 | 0x127f)
#define CGAL_FE_UPWARD (0x800 | 0x127f)
#define CGAL_FE_DOWNWARD (0x400 | 0x127f)
#elif defined __BORLANDC__
#define CGAL_IA_SETFPCW(CW) _control87(CW,~0)
#define CGAL_IA_GETFPCW(CW) CW = _control87(0,0)
typedef unsigned short FPU_CW_t;
#define CGAL_FE_TONEAREST (0x0 | 0x127f)
#define CGAL_FE_TOWARDZERO (0xC00 | 0x127f)
#define CGAL_FE_UPWARD (0x800 | 0x127f)
#define CGAL_FE_DOWNWARD (0x400 | 0x127f)
#else
// This is a version following the ISO C99 standard, which aims at portability.
// The drawbacks are speed on one hand, and also, on x86, it doesn't fix the
// extended mantissa issue (this is not a problem for IA, but it is one for
// Fixed_precision_nt, and some future modular computations as well).
#define CGAL_IA_SETFPCW(CW) fesetround(CW)
#define CGAL_IA_GETFPCW(CW) CW = fegetround()
typedef int FPU_CW_t;
#define CGAL_FE_TONEAREST FE_TONEAREST
#define CGAL_FE_TOWARDZERO FE_TOWARDZERO
#define CGAL_FE_UPWARD FE_UPWARD
#define CGAL_FE_DOWNWARD FE_DOWNWARD
#endif
// User interface:
inline
FPU_CW_t
FPU_get_cw (void)
{
FPU_CW_t cw;
CGAL_IA_GETFPCW(cw);
return cw;
}
inline
void
FPU_set_cw (FPU_CW_t cw)
{
CGAL_IA_SETFPCW(cw);
}
inline
FPU_CW_t
FPU_get_and_set_cw (FPU_CW_t cw)
{
FPU_CW_t old = FPU_get_cw();
FPU_set_cw(cw);
return old;
}
// The following is meant to truncate the mantissa of x86 FPUs to 53 bits.
// It is used by the Fixed_precision_nt.
inline void force_ieee_double_precision()
{
#if defined __i386__ || defined _MSC_VER || defined __BORLANDC__
FPU_set_cw(CGAL_FE_TONEAREST);
#endif
}
// A class whose constructor sets the FPU mode to +inf, saves a backup of it,
// and whose destructor resets it back to the saved state.
template <bool Protected = true> struct Protect_FPU_rounding;
template <>
struct Protect_FPU_rounding<true>
{
Protect_FPU_rounding(FPU_CW_t r = CGAL_FE_UPWARD)
: backup( FPU_get_and_set_cw(r) ) {}
~Protect_FPU_rounding()
{
FPU_set_cw(backup);
}
private:
FPU_CW_t backup;
};
template <>
struct Protect_FPU_rounding<false>
{
Protect_FPU_rounding()
{
CGAL_expensive_assertion(FPU_get_cw() == CGAL_FE_UPWARD);
}
Protect_FPU_rounding(FPU_CW_t /*= CGAL_FE_UPWARD*/)
{
//CGAL_expensive_assertion(FPU_get_cw() == CGAL_FE_UPWARD);
}
~Protect_FPU_rounding() {}
// just to shut up a warning, but it has a performance issue with GCC 2.95,
// so I should disable it for the moment. Use __attribute__((unused)) ?
};
CGAL_END_NAMESPACE
#endif // CGAL_FPU_H