On Wed, 27 Oct 2021, Jakub Jelinek wrote:
> On Wed, Oct 27, 2021 at 03:20:29PM +0200, Richard Biener via Gcc-patches
> wrote:
> > The following honors -frounding-math when converting a FP constant
> > to another FP type.
> >
> > Bootstrapped and tested on x86_64-unknown-linux-gnu, OK?
> >
> > I wonder what a good way to test this in a portable way, the bugreport
> > unfortunately didn't contain something executable and I don't see
> > much -frounding-math test coverage to copy from.
>
> E.g. following tests call fesetround, use fenv effective target etc.:
> torture/fp-int-convert-float128-timode-3.c: fesetround (FE_TOWARDZERO);
> torture/fp-int-convert-timode-2.c: fesetround (FE_DOWNWARD);
> torture/fp-int-convert-timode-3.c: fesetround (FE_UPWARD);
> torture/fp-int-convert-timode-4.c: fesetround (FE_TOWARDZERO);
>
> And the test can just hardcode one or more common float/double etc.
> configurations, checked using
> __{FLT,DBL}_{DIG,MANT_DIG,RADIX,MIN_EXP,MAX_EXP}__ etc. macros.
> Say just test double to float conversions of some specific values assuming
> float is IEEE754 single precicion and double is IEEE754 double precision
> in all the 4 rounding modes.
So something like the following below? Note I have to fix
simplify_const_unary_operation to not perform the invalid constant
folding with (not worrying about the exact conversion case - I doubt
any of the constant folding is really relevant on RTL these days,
maybe we should simply punt for all unary float-float ops when either
mode has sign dependent rounding modes)
diff --git a/gcc/simplify-rtx.c b/gcc/simplify-rtx.c
index bbbd6b74942..9522a31570e 100644
--- a/gcc/simplify-rtx.c
+++ b/gcc/simplify-rtx.c
@@ -2068,6 +2073,9 @@ simplify_const_unary_operation (enum rtx_code code,
machine_mode mode,
and the operand is a signaling NaN. */
if (HONOR_SNANS (mode) && REAL_VALUE_ISSIGNALING_NAN (d))
return NULL_RTX;
+ /* Or if flag_rounding_math is on. */
+ if (HONOR_SIGN_DEPENDENT_ROUNDING (mode))
+ return NULL_RTX;
d = real_value_truncate (mode, d);
break;
case FLOAT_EXTEND:
/* PR57245 */
/* { dg-do run } */
/* { dg-require-effective-target fenv } */
/* { dg-additional-options "-frounding-math" } */
#include <fenv.h>
#include <stdlib.h>
int
main ()
{
#if __DBL_MANT_DIG__ == 53 && __FLT_MANT_DIG__ == 24
fesetround (FE_UPWARD);
float f = 1.3;
if (f != 0x1.4ccccep+0f)
__builtin_abort ();
fesetround (FE_TONEAREST);
/* Use different actual values so the bogus CSE we perform does not
break things. */
f = 1.33;
if (f != 0x1.547ae2p+0f)
abort ();
fesetround (FE_DOWNWARD);
f = 1.333;
if (f != 0x1.553f7cp+0f)
abort ();
fesetround (FE_TOWARDZERO);
f = 1.3333;
if (f != 0x1.555326p+0f)
abort ();
#endif
return 0;
}
> > 2021-10-27 Richard Biener <rguent...@suse.de>
> >
> > PR middle-end/57245
> > * fold-const.c (fold_convert_const_real_from_real): Honor
> > -frounding-math if the conversion is not exact.
> > ---
> > gcc/fold-const.c | 6 ++++++
> > 1 file changed, 6 insertions(+)
> >
> > diff --git a/gcc/fold-const.c b/gcc/fold-const.c
> > index ff23f12f33c..c7aebf9cc7e 100644
> > --- a/gcc/fold-const.c
> > +++ b/gcc/fold-const.c
> > @@ -2139,6 +2139,12 @@ fold_convert_const_real_from_real (tree type,
> > const_tree arg1)
> > && REAL_VALUE_ISSIGNALING_NAN (TREE_REAL_CST (arg1)))
> > return NULL_TREE;
> >
> > + /* With flag_rounding_math we shuld respect the current rounding mode
>
> s/shuld/should/
>
> > + unless the conversion is exact. */
> > + if (HONOR_SIGN_DEPENDENT_ROUNDING (arg1)
> > + && !exact_real_truncate (TYPE_MODE (type), &TREE_REAL_CST (arg1)))
> > + return NULL_TREE;
> > +
> > real_convert (&value, TYPE_MODE (type), &TREE_REAL_CST (arg1));
> > t = build_real (type, value);
>
> Jakub
>
>
--
Richard Biener <rguent...@suse.de>
SUSE Software Solutions Germany GmbH, Maxfeldstrasse 5, 90409 Nuernberg,
Germany; GF: Felix Imendörffer; HRB 36809 (AG Nuernberg)
