On Mon, Aug 09, 2021 at 03:23:50PM -0500, Paul A. Clarke wrote:
> Suppress exceptions (when specified), by saving, manipulating, and
> restoring the FPSCR. Similarly, save, set, and restore the floating-point
> rounding mode when required.
>
> No attempt is made to optimize writing the FPSCR (by checking if the new
> value would be the same), other than using lighter weight instructions
> when possible.
There are __builtin_set_fpscr_rn and friends, please use those, those
are optimised for any platform.
> * config/rs6000/smmintrin.h (_mm_ceil_pd, _mm_ceil_ps, _mm_ceil_sd,
> _mm_ceil_ss, _mm_floor_pd, _mm_floor_ps, _mm_floor_sd, _mm_floor_ss):
> Convert from function to macro.
Please explain why you regress this (not in the changelog of course).
> +/* Rounding mode macros. */
> +#define _MM_FROUND_TO_NEAREST_INT 0x00
> +#define _MM_FROUND_TO_ZERO 0x01
> +#define _MM_FROUND_TO_POS_INF 0x02
> +#define _MM_FROUND_TO_NEG_INF 0x03
> +#define _MM_FROUND_CUR_DIRECTION 0x04
You can just write "0" .. "4", heh.
> +
> +#define _MM_FROUND_NINT \
> + (_MM_FROUND_TO_NEAREST_INT | _MM_FROUND_RAISE_EXC)
> +#define _MM_FROUND_FLOOR \
> + (_MM_FROUND_TO_NEG_INF | _MM_FROUND_RAISE_EXC)
> +#define _MM_FROUND_CEIL \
> + (_MM_FROUND_TO_POS_INF | _MM_FROUND_RAISE_EXC)
> +#define _MM_FROUND_TRUNC \
> + (_MM_FROUND_TO_ZERO | _MM_FROUND_RAISE_EXC)
> +#define _MM_FROUND_RINT \
> + (_MM_FROUND_CUR_DIRECTION | _MM_FROUND_RAISE_EXC)
> +#define _MM_FROUND_NEARBYINT \
> + (_MM_FROUND_CUR_DIRECTION | _MM_FROUND_NO_EXC)
All these macro definitions will comfortably fit on one line.
> +__inline __m128d
> +__attribute__ ((__gnu_inline__, __always_inline__, __artificial__))
> +_mm_round_pd (__m128d __A, int __rounding)
> +{
Non-static inline is not what you want, esp. with gnu-inline? Or, what
is the goal, and why can you not do it with modern inline?
> + __v2df __r;
> + union {
> + double __fr;
> + long long __fpscr;
> + } __save, __tmp;
> +
> + if (__rounding & _MM_FROUND_NO_EXC)
> + {
Wrong indent. This code is very hard to read because of that.
If you figure that gee, it would be a nice if we had a builtin for
mffsce, then please make one? :-)
> + case _MM_FROUND_TO_NEAREST_INT:
> + __tmp.__fr = __builtin_mffsl ();
> + __attribute__((fallthrough));
Space before (.
> + case _MM_FROUND_TO_NEAREST_INT |_MM_FROUND_NO_EXC:
Space after |.
Please fix these things and resend.
Segher
