[PATCH] benchtests: Add more benchtests for rounding functions.

Adhemerval Zanella Netto adhemerval.zanella@linaro.org
Fri Jan 19 12:33:10 GMT 2024



On 17/01/24 05:17, Junxian Zhu wrote:
> From: Junxian Zhu <zhujunxian@oss.cipunited.com>
> 
> This patch adds more benchtests for rounding functions.
> The double inputs are copied from trunc-inputs, the float inputs are copied from truncf-inputs. and the rintf is copied from rint-inputs.

LGTM, thanks.

Reviewed-by: Adhemerval Zanella  <adhemerval.zanella@linaro.org>

> ---
>  benchtests/Makefile          | 11 +++++++++++
>  benchtests/ceil-inputs       | 22 ++++++++++++++++++++++
>  benchtests/ceilf-inputs      | 21 +++++++++++++++++++++
>  benchtests/floor-inputs      | 22 ++++++++++++++++++++++
>  benchtests/floorf-inputs     | 21 +++++++++++++++++++++
>  benchtests/llrint-inputs     | 22 ++++++++++++++++++++++
>  benchtests/llrintf-inputs    | 21 +++++++++++++++++++++
>  benchtests/lrint-inputs      | 22 ++++++++++++++++++++++
>  benchtests/lrintf-inputs     | 21 +++++++++++++++++++++
>  benchtests/nearbyint-inputs  | 22 ++++++++++++++++++++++
>  benchtests/nearbyintf-inputs | 21 +++++++++++++++++++++
>  benchtests/rintf-inputs      |  7 +++++++
>  12 files changed, 233 insertions(+)
>  create mode 100644 benchtests/ceil-inputs
>  create mode 100644 benchtests/ceilf-inputs
>  create mode 100644 benchtests/floor-inputs
>  create mode 100644 benchtests/floorf-inputs
>  create mode 100644 benchtests/llrint-inputs
>  create mode 100644 benchtests/llrintf-inputs
>  create mode 100644 benchtests/lrint-inputs
>  create mode 100644 benchtests/lrintf-inputs
>  create mode 100644 benchtests/nearbyint-inputs
>  create mode 100644 benchtests/nearbyintf-inputs
>  create mode 100644 benchtests/rintf-inputs
> 
> diff --git a/benchtests/Makefile b/benchtests/Makefile
> index 4fb9bae023..c3e281f076 100644
> --- a/benchtests/Makefile
> +++ b/benchtests/Makefile
> @@ -32,6 +32,8 @@ bench-math := \
>    atan2 \
>    atanh \
>    cbrt \
> +  ceil \
> +  ceilf \
>    cos \
>    cosf \
>    cosh \
> @@ -44,6 +46,8 @@ bench-math := \
>    exp2f \
>    expf \
>    expm1 \
> +  floor \
> +  floorf \
>    fmax \
>    fmaxf \
>    fmin \
> @@ -68,10 +72,17 @@ bench-math := \
>    logb \
>    logbf \
>    logf \
> +  llrint \
> +  llrintf \
> +  lrint \
> +  lrintf \
>    modf \
> +  nearbyint \
> +  nearbyintf \
>    pow \
>    powf \
>    rint \
> +  rintf \
>    roundeven \
>    roundevenf \
>    sin \
> diff --git a/benchtests/ceil-inputs b/benchtests/ceil-inputs
> new file mode 100644
> index 0000000000..49ff407a6a
> --- /dev/null
> +++ b/benchtests/ceil-inputs
> @@ -0,0 +1,22 @@
> +## args: double
> +## ret: double
> +## includes: math.h
> +0.0
> +-0.0
> +0.001
> +-0.001
> +0.5
> +-0.5
> +0.999
> +-0.999
> +1.0
> +-1.0
> +1.001
> +-1.001
> +123.5
> +-123.5
> +12345.1
> +-1000000.1
> +1e15
> +-1e30
> +1e200
> diff --git a/benchtests/ceilf-inputs b/benchtests/ceilf-inputs
> new file mode 100644
> index 0000000000..c37c5dacba
> --- /dev/null
> +++ b/benchtests/ceilf-inputs
> @@ -0,0 +1,21 @@
> +## args: float
> +## ret: float
> +## includes: math.h
> +0.0f
> +-0.0f
> +0.001f
> +-0.001f
> +0.5f
> +-0.5f
> +0.999f
> +-0.999f
> +1.0f
> +-1.0f
> +1.001f
> +-1.001f
> +123.5f
> +-123.5f
> +12345.1f
> +-1000000.5f
> +1e15f
> +-1e30f
> diff --git a/benchtests/floor-inputs b/benchtests/floor-inputs
> new file mode 100644
> index 0000000000..49ff407a6a
> --- /dev/null
> +++ b/benchtests/floor-inputs
> @@ -0,0 +1,22 @@
> +## args: double
> +## ret: double
> +## includes: math.h
> +0.0
> +-0.0
> +0.001
> +-0.001
> +0.5
> +-0.5
> +0.999
> +-0.999
> +1.0
> +-1.0
> +1.001
> +-1.001
> +123.5
> +-123.5
> +12345.1
> +-1000000.1
> +1e15
> +-1e30
> +1e200
> diff --git a/benchtests/floorf-inputs b/benchtests/floorf-inputs
> new file mode 100644
> index 0000000000..c37c5dacba
> --- /dev/null
> +++ b/benchtests/floorf-inputs
> @@ -0,0 +1,21 @@
> +## args: float
> +## ret: float
> +## includes: math.h
> +0.0f
> +-0.0f
> +0.001f
> +-0.001f
> +0.5f
> +-0.5f
> +0.999f
> +-0.999f
> +1.0f
> +-1.0f
> +1.001f
> +-1.001f
> +123.5f
> +-123.5f
> +12345.1f
> +-1000000.5f
> +1e15f
> +-1e30f
> diff --git a/benchtests/llrint-inputs b/benchtests/llrint-inputs
> new file mode 100644
> index 0000000000..49ff407a6a
> --- /dev/null
> +++ b/benchtests/llrint-inputs
> @@ -0,0 +1,22 @@
> +## args: double
> +## ret: double
> +## includes: math.h
> +0.0
> +-0.0
> +0.001
> +-0.001
> +0.5
> +-0.5
> +0.999
> +-0.999
> +1.0
> +-1.0
> +1.001
> +-1.001
> +123.5
> +-123.5
> +12345.1
> +-1000000.1
> +1e15
> +-1e30
> +1e200
> diff --git a/benchtests/llrintf-inputs b/benchtests/llrintf-inputs
> new file mode 100644
> index 0000000000..c37c5dacba
> --- /dev/null
> +++ b/benchtests/llrintf-inputs
> @@ -0,0 +1,21 @@
> +## args: float
> +## ret: float
> +## includes: math.h
> +0.0f
> +-0.0f
> +0.001f
> +-0.001f
> +0.5f
> +-0.5f
> +0.999f
> +-0.999f
> +1.0f
> +-1.0f
> +1.001f
> +-1.001f
> +123.5f
> +-123.5f
> +12345.1f
> +-1000000.5f
> +1e15f
> +-1e30f
> diff --git a/benchtests/lrint-inputs b/benchtests/lrint-inputs
> new file mode 100644
> index 0000000000..49ff407a6a
> --- /dev/null
> +++ b/benchtests/lrint-inputs
> @@ -0,0 +1,22 @@
> +## args: double
> +## ret: double
> +## includes: math.h
> +0.0
> +-0.0
> +0.001
> +-0.001
> +0.5
> +-0.5
> +0.999
> +-0.999
> +1.0
> +-1.0
> +1.001
> +-1.001
> +123.5
> +-123.5
> +12345.1
> +-1000000.1
> +1e15
> +-1e30
> +1e200
> diff --git a/benchtests/lrintf-inputs b/benchtests/lrintf-inputs
> new file mode 100644
> index 0000000000..c37c5dacba
> --- /dev/null
> +++ b/benchtests/lrintf-inputs
> @@ -0,0 +1,21 @@
> +## args: float
> +## ret: float
> +## includes: math.h
> +0.0f
> +-0.0f
> +0.001f
> +-0.001f
> +0.5f
> +-0.5f
> +0.999f
> +-0.999f
> +1.0f
> +-1.0f
> +1.001f
> +-1.001f
> +123.5f
> +-123.5f
> +12345.1f
> +-1000000.5f
> +1e15f
> +-1e30f
> diff --git a/benchtests/nearbyint-inputs b/benchtests/nearbyint-inputs
> new file mode 100644
> index 0000000000..49ff407a6a
> --- /dev/null
> +++ b/benchtests/nearbyint-inputs
> @@ -0,0 +1,22 @@
> +## args: double
> +## ret: double
> +## includes: math.h
> +0.0
> +-0.0
> +0.001
> +-0.001
> +0.5
> +-0.5
> +0.999
> +-0.999
> +1.0
> +-1.0
> +1.001
> +-1.001
> +123.5
> +-123.5
> +12345.1
> +-1000000.1
> +1e15
> +-1e30
> +1e200
> diff --git a/benchtests/nearbyintf-inputs b/benchtests/nearbyintf-inputs
> new file mode 100644
> index 0000000000..c37c5dacba
> --- /dev/null
> +++ b/benchtests/nearbyintf-inputs
> @@ -0,0 +1,21 @@
> +## args: float
> +## ret: float
> +## includes: math.h
> +0.0f
> +-0.0f
> +0.001f
> +-0.001f
> +0.5f
> +-0.5f
> +0.999f
> +-0.999f
> +1.0f
> +-1.0f
> +1.001f
> +-1.001f
> +123.5f
> +-123.5f
> +12345.1f
> +-1000000.5f
> +1e15f
> +-1e30f
> diff --git a/benchtests/rintf-inputs b/benchtests/rintf-inputs
> new file mode 100644
> index 0000000000..f7c6f06789
> --- /dev/null
> +++ b/benchtests/rintf-inputs
> @@ -0,0 +1,7 @@
> +## args: float
> +## ret: float
> +## includes: math.h
> +78.5
> +-78.5
> +4503599627370497.0
> +-4503599627370497.0


More information about the Libc-alpha mailing list