From: Adhemerval Zanella <adhemerval.zanella@linaro.org>
To: libc-alpha@sourceware.org
Subject: Re: [PATCH 11/13] Use GCC builtins for trunc functions if desired.
Date: Thu, 05 Dec 2019 20:40:00 -0000 [thread overview]
Message-ID: <3397a1bc-aeaf-b7c9-08f6-2335deaf06c3@linaro.org> (raw)
In-Reply-To: <1575297977-2589-12-git-send-email-stli@linux.ibm.com>
On 02/12/2019 11:46, Stefan Liebler wrote:
> This patch is using the corresponding GCC builtin for truncf, trunc,
> truncl and truncf128 if the USE_FUNCTION_BUILTIN macros are defined to one
> in math-use-builtins.h.
>
> This is the case for s390 if build with at least --march=z196 --mzarch.
> Otherwise the generic implementation is used.
LGTM, thanks.
Reviewed-by: Adhemerval Zanella <adhemerval.zanella@linaro.org>
> ---
> sysdeps/generic/math-use-builtins.h | 5 +++++
> sysdeps/ieee754/dbl-64/s_trunc.c | 6 +++++-
> sysdeps/ieee754/float128/float128_private.h | 3 +++
> sysdeps/ieee754/flt-32/s_truncf.c | 6 +++++-
> sysdeps/ieee754/ldbl-128/s_truncl.c | 6 +++++-
> sysdeps/s390/fpu/math-use-builtins.h | 6 ++++++
> 6 files changed, 29 insertions(+), 3 deletions(-)
>
> diff --git a/sysdeps/generic/math-use-builtins.h b/sysdeps/generic/math-use-builtins.h
> index 076ec661b0..ab379f45ba 100644
> --- a/sysdeps/generic/math-use-builtins.h
> +++ b/sysdeps/generic/math-use-builtins.h
> @@ -41,4 +41,9 @@
> #define USE_CEILL_BUILTIN 0
> #define USE_CEILF128_BUILTIN 0
>
> +#define USE_TRUNC_BUILTIN 0
> +#define USE_TRUNCF_BUILTIN 0
> +#define USE_TRUNCL_BUILTIN 0
> +#define USE_TRUNCF128_BUILTIN 0
> +
> #endif /* math-use-builtins.h */
Ok.
> diff --git a/sysdeps/ieee754/dbl-64/s_trunc.c b/sysdeps/ieee754/dbl-64/s_trunc.c
> index 0de68ce298..6189ba4a31 100644
> --- a/sysdeps/ieee754/dbl-64/s_trunc.c
> +++ b/sysdeps/ieee754/dbl-64/s_trunc.c
> @@ -22,11 +22,14 @@
>
> #include <math_private.h>
> #include <libm-alias-double.h>
> -
> +#include <math-use-builtins.h>
>
> double
> __trunc (double x)
> {
> +#if USE_TRUNC_BUILTIN
> + return __builtin_trunc (x);
> +#else
> int64_t i0, j0;
> int64_t sx;
>
> @@ -49,6 +52,7 @@ __trunc (double x)
> }
>
> return x;
> +#endif /* USE_TRUNC_BUILTIN */
> }
> #ifndef __trunc
> libm_alias_double (__trunc, trunc)
Ok.
> diff --git a/sysdeps/ieee754/float128/float128_private.h b/sysdeps/ieee754/float128/float128_private.h
> index 01881b574a..9cd3a63f11 100644
> --- a/sysdeps/ieee754/float128/float128_private.h
> +++ b/sysdeps/ieee754/float128/float128_private.h
> @@ -148,6 +148,8 @@
> #define USE_FLOORL_BUILTIN USE_FLOORF128_BUILTIN
> #undef USE_CEILL_BUILTIN
> #define USE_CEILL_BUILTIN USE_CEILF128_BUILTIN
> +#undef USE_TRUNCL_BUILTIN
> +#define USE_TRUNCL_BUILTIN USE_TRUNCF128_BUILTIN
>
> /* IEEE function renames. */
> #define __ieee754_acoshl __ieee754_acoshf128
> @@ -355,6 +357,7 @@
> #define __builtin_rintl __builtin_rintf128
> #define __builtin_floorl __builtin_floorf128
> #define __builtin_ceill __builtin_ceilf128
> +#define __builtin_truncl __builtin_truncf128
>
> /* Get the constant suffix from bits/floatn-compat.h. */
> #define L(x) __f128 (x)
Ok.
> diff --git a/sysdeps/ieee754/flt-32/s_truncf.c b/sysdeps/ieee754/flt-32/s_truncf.c
> index e587706b5b..de6dc24a3f 100644
> --- a/sysdeps/ieee754/flt-32/s_truncf.c
> +++ b/sysdeps/ieee754/flt-32/s_truncf.c
> @@ -22,11 +22,14 @@
>
> #include <math_private.h>
> #include <libm-alias-float.h>
> -
> +#include <math-use-builtins.h>
>
> float
> __truncf (float x)
> {
> +#if USE_TRUNCF_BUILTIN
> + return __builtin_truncf (x);
> +#else
> int32_t i0, j0;
> int sx;
>
> @@ -49,6 +52,7 @@ __truncf (float x)
> }
>
> return x;
> +#endif /* USE_TRUNCF_BUILTIN */
> }
> #ifndef __truncf
> libm_alias_float (__trunc, trunc)
Ok.
> diff --git a/sysdeps/ieee754/ldbl-128/s_truncl.c b/sysdeps/ieee754/ldbl-128/s_truncl.c
> index de4dd34d59..ca62e2f970 100644
> --- a/sysdeps/ieee754/ldbl-128/s_truncl.c
> +++ b/sysdeps/ieee754/ldbl-128/s_truncl.c
> @@ -23,11 +23,14 @@
>
> #include <math_private.h>
> #include <libm-alias-ldouble.h>
> -
> +#include <math-use-builtins.h>
>
> _Float128
> __truncl (_Float128 x)
> {
> +#if USE_TRUNCL_BUILTIN
> + return __builtin_truncl (x);
> +#else
> int32_t j0;
> uint64_t i0, i1, sx;
>
> @@ -54,5 +57,6 @@ __truncl (_Float128 x)
> }
>
> return x;
> +#endif /* USE_TRUNCL_BUILTIN */
> }
> libm_alias_ldouble (__trunc, trunc)
Ok.
> diff --git a/sysdeps/s390/fpu/math-use-builtins.h b/sysdeps/s390/fpu/math-use-builtins.h
> index dc364edaa8..1d88a2c40f 100644
> --- a/sysdeps/s390/fpu/math-use-builtins.h
> +++ b/sysdeps/s390/fpu/math-use-builtins.h
> @@ -42,16 +42,22 @@
> # define USE_CEILF_BUILTIN 1
> # define USE_CEILL_BUILTIN 1
>
> +# define USE_TRUNC_BUILTIN 1
> +# define USE_TRUNCF_BUILTIN 1
> +# define USE_TRUNCL_BUILTIN 1
> +
> # if __GNUC_PREREQ (8, 1)
> # define USE_NEARBYINTF128_BUILTIN 1
> # define USE_RINTF128_BUILTIN 1
> # define USE_FLOORF128_BUILTIN 1
> # define USE_CEILF128_BUILTIN 1
> +# define USE_TRUNCF128_BUILTIN 1
> # else
> # define USE_NEARBYINTF128_BUILTIN 0
> # define USE_RINTF128_BUILTIN 0
> # define USE_FLOORF128_BUILTIN 0
> # define USE_CEILF128_BUILTIN 0
> +# define USE_TRUNCF128_BUILTIN 0
> # endif
>
> #else
>
Ok.
next prev parent reply other threads:[~2019-12-05 20:40 UTC|newest]
Thread overview: 35+ messages / expand[flat|nested] mbox.gz Atom feed top
2019-12-02 14:46 [PATCH 00/13] Use GCC builtins for some math " Stefan Liebler
2019-12-02 14:46 ` [PATCH 09/13] Use GCC builtins for floor " Stefan Liebler
2019-12-05 20:40 ` Adhemerval Zanella
2019-12-02 14:46 ` [PATCH 08/13] Use GCC builtins for rint " Stefan Liebler
2019-12-05 20:40 ` Adhemerval Zanella
2019-12-02 14:46 ` [PATCH 01/13] Always use wordsize-64 version of s_nearbyint.c Stefan Liebler
2019-12-05 20:40 ` Adhemerval Zanella
2019-12-02 14:46 ` [PATCH 07/13] Use GCC builtins for nearbyint functions if desired Stefan Liebler
2019-12-05 20:40 ` Adhemerval Zanella
2019-12-02 14:46 ` [PATCH 05/13] Always use wordsize-64 version of s_trunc.c Stefan Liebler
2019-12-05 20:40 ` Adhemerval Zanella
2019-12-02 14:46 ` [PATCH 03/13] Always use wordsize-64 version of s_floor.c Stefan Liebler
2019-12-05 20:40 ` Adhemerval Zanella
2019-12-02 14:46 ` [PATCH 02/13] Always use wordsize-64 version of s_rint.c Stefan Liebler
2019-12-05 20:40 ` Adhemerval Zanella
2019-12-02 15:10 ` [PATCH 11/13] Use GCC builtins for trunc functions if desired Stefan Liebler
2019-12-05 20:40 ` Adhemerval Zanella [this message]
2019-12-02 15:15 ` [PATCH 04/13] Always use wordsize-64 version of s_ceil.c Stefan Liebler
2019-12-05 20:40 ` Adhemerval Zanella
2019-12-02 15:18 ` [PATCH 10/13] Use GCC builtins for ceil functions if desired Stefan Liebler
2019-12-05 20:40 ` Adhemerval Zanella
2019-12-02 15:46 ` [PATCH 06/13] Always use wordsize-64 version of s_round.c Stefan Liebler
2019-12-05 20:40 ` Adhemerval Zanella
2019-12-02 15:50 ` [PATCH 12/13] Use GCC builtins for round functions if desired Stefan Liebler
2019-12-05 20:41 ` Adhemerval Zanella
2019-12-02 15:52 ` [PATCH 13/13] Use GCC builtins for copysign " Stefan Liebler
2019-12-02 21:00 ` Joseph Myers
2019-12-03 8:27 ` Stefan Liebler
2019-12-03 16:51 ` Joseph Myers
2019-12-04 13:15 ` Stefan Liebler
2019-12-04 13:20 ` Joseph Myers
2019-12-04 16:34 ` Stefan Liebler
2019-12-04 20:43 ` Joseph Myers
2019-12-05 15:40 ` Stefan Liebler
2019-12-09 12:58 ` [PATCH 00/13] Use GCC builtins for some math " Stefan Liebler
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=3397a1bc-aeaf-b7c9-08f6-2335deaf06c3@linaro.org \
--to=adhemerval.zanella@linaro.org \
--cc=libc-alpha@sourceware.org \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for read-only IMAP folder(s) and NNTP newsgroup(s).