[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: [Qemu-devel] [PATCH v1 12/14] hostfloat: support float32/64 square r
From: |
Alex Bennée |
Subject: |
Re: [Qemu-devel] [PATCH v1 12/14] hostfloat: support float32/64 square root |
Date: |
Thu, 22 Mar 2018 01:29:23 +0000 |
User-agent: |
mu4e 1.1.0; emacs 26.0.91 |
Emilio G. Cota <address@hidden> writes:
> Performance results for fp-bench run under aarch64-linux-user
> on an Intel(R) Core(TM) i7-4790K CPU @ 4.00GHz host:
>
> - before:
> sqrt-single: 13.23 MFlops
> sqrt-double: 13.24 MFlops
>
> - after:
> sqrt-single: 15.02 MFlops
> sqrt-double: 15.07 MFlops
>
> Note that sqrt in soft-ft is relatively fast, which means
> that fp-bench is not very sensitive to changes to sqrt's
> emulation speed.
Weird, I thought we had slowed it down quite a bit in the re-factor as
we eschewed the estimate step for an easier to read but slower iterative
process. That's why I chose sqrt for my hostfp hack experiment.
>
> Signed-off-by: Emilio G. Cota <address@hidden>
> ---
> include/fpu/hostfloat.h | 2 ++
> include/fpu/softfloat.h | 4 ++--
> fpu/hostfloat.c | 20 ++++++++++++++++++++
> fpu/softfloat.c | 6 ++++--
> 4 files changed, 28 insertions(+), 4 deletions(-)
>
> diff --git a/include/fpu/hostfloat.h b/include/fpu/hostfloat.h
> index c006576..b1e0689 100644
> --- a/include/fpu/hostfloat.h
> +++ b/include/fpu/hostfloat.h
> @@ -16,11 +16,13 @@ float32 float32_sub(float32 a, float32 b, float_status
> *status);
> float32 float32_mul(float32 a, float32 b, float_status *status);
> float32 float32_div(float32 a, float32 b, float_status *status);
> float32 float32_muladd(float32 a, float32 b, float32 c, int f, float_status
> *s);
> +float32 float32_sqrt(float32 a, float_status *status);
>
> float64 float64_add(float64 a, float64 b, float_status *status);
> float64 float64_sub(float64 a, float64 b, float_status *status);
> float64 float64_mul(float64 a, float64 b, float_status *status);
> float64 float64_div(float64 a, float64 b, float_status *status);
> float64 float64_muladd(float64 a, float64 b, float64 c, int f, float_status
> *s);
> +float64 float64_sqrt(float64 a, float_status *status);
>
> #endif /* HOSTFLOAT_H */
> diff --git a/include/fpu/softfloat.h b/include/fpu/softfloat.h
> index 866bd3b..8d5a50a 100644
> --- a/include/fpu/softfloat.h
> +++ b/include/fpu/softfloat.h
> @@ -348,7 +348,7 @@ float32 soft_float32_mul(float32, float32, float_status
> *status);
> float32 soft_float32_div(float32, float32, float_status *status);
> float32 float32_rem(float32, float32, float_status *status);
> float32 soft_float32_muladd(float32, float32, float32, int, float_status *s);
> -float32 float32_sqrt(float32, float_status *status);
> +float32 soft_float32_sqrt(float32, float_status *status);
> float32 float32_exp2(float32, float_status *status);
> float32 float32_log2(float32, float_status *status);
> int float32_eq(float32, float32, float_status *status);
> @@ -488,7 +488,7 @@ float64 soft_float64_mul(float64, float64, float_status
> *status);
> float64 soft_float64_div(float64, float64, float_status *status);
> float64 float64_rem(float64, float64, float_status *status);
> float64 soft_float64_muladd(float64, float64, float64, int, float_status *s);
> -float64 float64_sqrt(float64, float_status *status);
> +float64 soft_float64_sqrt(float64, float_status *status);
> float64 float64_log2(float64, float_status *status);
> int float64_eq(float64, float64, float_status *status);
> int float64_le(float64, float64, float_status *status);
> diff --git a/fpu/hostfloat.c b/fpu/hostfloat.c
> index a56b70a..974bd57 100644
> --- a/fpu/hostfloat.c
> +++ b/fpu/hostfloat.c
> @@ -270,3 +270,23 @@ GEN_FPU_DIV(float64_div, float64, double, fabs, DBL_MIN)
> GEN_FPU_FMA(float32_muladd, float32, float, fmaf, fabsf, FLT_MIN)
> GEN_FPU_FMA(float64_muladd, float64, double, fma, fabs, DBL_MIN)
> #undef GEN_FPU_FMA
> +
> +#define GEN_FPU_SQRT(name, soft_t, host_t, host_sqrt_func) \
> + soft_t name(soft_t a, float_status *s) \
> + { \
> + soft_t ## _input_flush1(&a, s); \
> + if (likely((soft_t ## _is_normal(a) || soft_t ## _is_zero(a)) && \
> + !soft_t ## _is_neg(a) && \
> + s->float_exception_flags & float_flag_inexact && \
> + s->float_rounding_mode == float_round_nearest_even)) { \
> + host_t ha = soft_t ## _to_ ## host_t(a); \
> + host_t hr = host_sqrt_func(ha); \
> + \
> + return host_t ## _to_ ## soft_t(hr); \
> + } \
> + return soft_ ## soft_t ## _sqrt(a, s); \
> + }
> +
> +GEN_FPU_SQRT(float32_sqrt, float32, float, sqrtf)
> +GEN_FPU_SQRT(float64_sqrt, float64, double, sqrt)
> +#undef GEN_FPU_SQRT
> diff --git a/fpu/softfloat.c b/fpu/softfloat.c
> index da81ec9..096b658 100644
> --- a/fpu/softfloat.c
> +++ b/fpu/softfloat.c
> @@ -1980,14 +1980,16 @@ float16 __attribute__((flatten)) float16_sqrt(float16
> a, float_status *status)
> return float16_round_pack_canonical(pr, status);
> }
>
> -float32 __attribute__((flatten)) float32_sqrt(float32 a, float_status
> *status)
> +float32 __attribute__((flatten))
> +soft_float32_sqrt(float32 a, float_status *status)
> {
> FloatParts pa = float32_unpack_canonical(a, status);
> FloatParts pr = sqrt_float(pa, status, &float32_params);
> return float32_round_pack_canonical(pr, status);
> }
>
> -float64 __attribute__((flatten)) float64_sqrt(float64 a, float_status
> *status)
> +float64 __attribute__((flatten))
> +soft_float64_sqrt(float64 a, float_status *status)
> {
> FloatParts pa = float64_unpack_canonical(a, status);
> FloatParts pr = sqrt_float(pa, status, &float64_params);
--
Alex Bennée
- Re: [Qemu-devel] [PATCH v1 08/14] hostfloat: support float32/64 addition and subtraction, (continued)
[Qemu-devel] [PATCH v1 09/14] hostfloat: support float32/64 multiplication, Emilio G. Cota, 2018/03/21
[Qemu-devel] [PATCH v1 01/14] tests: add fp-bench, a collection of simple floating-point microbenchmarks, Emilio G. Cota, 2018/03/21
[Qemu-devel] [PATCH v1 11/14] hostfloat: support float32/64 fused multiply-add, Emilio G. Cota, 2018/03/21
[Qemu-devel] [PATCH v1 10/14] hostfloat: support float32/64 division, Emilio G. Cota, 2018/03/21
[Qemu-devel] [PATCH v1 12/14] hostfloat: support float32/64 square root, Emilio G. Cota, 2018/03/21
- Re: [Qemu-devel] [PATCH v1 12/14] hostfloat: support float32/64 square root,
Alex Bennée <=
[Qemu-devel] [PATCH v1 05/14] softfloat: add float32_is_normal and float64_is_normal, Emilio G. Cota, 2018/03/21
[Qemu-devel] [PATCH v1 13/14] hostfloat: support float32/64 comparison, Emilio G. Cota, 2018/03/21
[Qemu-devel] [PATCH v1 14/14] hostfloat: support float32_to_float64, Emilio G. Cota, 2018/03/21
[Qemu-devel] [PATCH v1 02/14] tests: add fp-test, a floating point test suite, Emilio G. Cota, 2018/03/21