
(FPCore (cosTheta_i u1 u2) :precision binary32 (* (sqrt (/ u1 (- 1.0 u1))) (cos (* 6.28318530718 u2))))
float code(float cosTheta_i, float u1, float u2) {
return sqrtf((u1 / (1.0f - u1))) * cosf((6.28318530718f * u2));
}
module fmin_fmax_functions
implicit none
private
public fmax
public fmin
interface fmax
module procedure fmax88
module procedure fmax44
module procedure fmax84
module procedure fmax48
end interface
interface fmin
module procedure fmin88
module procedure fmin44
module procedure fmin84
module procedure fmin48
end interface
contains
real(8) function fmax88(x, y) result (res)
real(8), intent (in) :: x
real(8), intent (in) :: y
res = merge(y, merge(x, max(x, y), y /= y), x /= x)
end function
real(4) function fmax44(x, y) result (res)
real(4), intent (in) :: x
real(4), intent (in) :: y
res = merge(y, merge(x, max(x, y), y /= y), x /= x)
end function
real(8) function fmax84(x, y) result(res)
real(8), intent (in) :: x
real(4), intent (in) :: y
res = merge(dble(y), merge(x, max(x, dble(y)), y /= y), x /= x)
end function
real(8) function fmax48(x, y) result(res)
real(4), intent (in) :: x
real(8), intent (in) :: y
res = merge(y, merge(dble(x), max(dble(x), y), y /= y), x /= x)
end function
real(8) function fmin88(x, y) result (res)
real(8), intent (in) :: x
real(8), intent (in) :: y
res = merge(y, merge(x, min(x, y), y /= y), x /= x)
end function
real(4) function fmin44(x, y) result (res)
real(4), intent (in) :: x
real(4), intent (in) :: y
res = merge(y, merge(x, min(x, y), y /= y), x /= x)
end function
real(8) function fmin84(x, y) result(res)
real(8), intent (in) :: x
real(4), intent (in) :: y
res = merge(dble(y), merge(x, min(x, dble(y)), y /= y), x /= x)
end function
real(8) function fmin48(x, y) result(res)
real(4), intent (in) :: x
real(8), intent (in) :: y
res = merge(y, merge(dble(x), min(dble(x), y), y /= y), x /= x)
end function
end module
real(4) function code(costheta_i, u1, u2)
use fmin_fmax_functions
real(4), intent (in) :: costheta_i
real(4), intent (in) :: u1
real(4), intent (in) :: u2
code = sqrt((u1 / (1.0e0 - u1))) * cos((6.28318530718e0 * u2))
end function
function code(cosTheta_i, u1, u2) return Float32(sqrt(Float32(u1 / Float32(Float32(1.0) - u1))) * cos(Float32(Float32(6.28318530718) * u2))) end
function tmp = code(cosTheta_i, u1, u2) tmp = sqrt((u1 / (single(1.0) - u1))) * cos((single(6.28318530718) * u2)); end
\begin{array}{l}
\\
\sqrt{\frac{u1}{1 - u1}} \cdot \cos \left(6.28318530718 \cdot u2\right)
\end{array}
Herbie found 15 alternatives:
| Alternative | Accuracy | Speedup |
|---|
(FPCore (cosTheta_i u1 u2) :precision binary32 (* (sqrt (/ u1 (- 1.0 u1))) (cos (* 6.28318530718 u2))))
float code(float cosTheta_i, float u1, float u2) {
return sqrtf((u1 / (1.0f - u1))) * cosf((6.28318530718f * u2));
}
module fmin_fmax_functions
implicit none
private
public fmax
public fmin
interface fmax
module procedure fmax88
module procedure fmax44
module procedure fmax84
module procedure fmax48
end interface
interface fmin
module procedure fmin88
module procedure fmin44
module procedure fmin84
module procedure fmin48
end interface
contains
real(8) function fmax88(x, y) result (res)
real(8), intent (in) :: x
real(8), intent (in) :: y
res = merge(y, merge(x, max(x, y), y /= y), x /= x)
end function
real(4) function fmax44(x, y) result (res)
real(4), intent (in) :: x
real(4), intent (in) :: y
res = merge(y, merge(x, max(x, y), y /= y), x /= x)
end function
real(8) function fmax84(x, y) result(res)
real(8), intent (in) :: x
real(4), intent (in) :: y
res = merge(dble(y), merge(x, max(x, dble(y)), y /= y), x /= x)
end function
real(8) function fmax48(x, y) result(res)
real(4), intent (in) :: x
real(8), intent (in) :: y
res = merge(y, merge(dble(x), max(dble(x), y), y /= y), x /= x)
end function
real(8) function fmin88(x, y) result (res)
real(8), intent (in) :: x
real(8), intent (in) :: y
res = merge(y, merge(x, min(x, y), y /= y), x /= x)
end function
real(4) function fmin44(x, y) result (res)
real(4), intent (in) :: x
real(4), intent (in) :: y
res = merge(y, merge(x, min(x, y), y /= y), x /= x)
end function
real(8) function fmin84(x, y) result(res)
real(8), intent (in) :: x
real(4), intent (in) :: y
res = merge(dble(y), merge(x, min(x, dble(y)), y /= y), x /= x)
end function
real(8) function fmin48(x, y) result(res)
real(4), intent (in) :: x
real(8), intent (in) :: y
res = merge(y, merge(dble(x), min(dble(x), y), y /= y), x /= x)
end function
end module
real(4) function code(costheta_i, u1, u2)
use fmin_fmax_functions
real(4), intent (in) :: costheta_i
real(4), intent (in) :: u1
real(4), intent (in) :: u2
code = sqrt((u1 / (1.0e0 - u1))) * cos((6.28318530718e0 * u2))
end function
function code(cosTheta_i, u1, u2) return Float32(sqrt(Float32(u1 / Float32(Float32(1.0) - u1))) * cos(Float32(Float32(6.28318530718) * u2))) end
function tmp = code(cosTheta_i, u1, u2) tmp = sqrt((u1 / (single(1.0) - u1))) * cos((single(6.28318530718) * u2)); end
\begin{array}{l}
\\
\sqrt{\frac{u1}{1 - u1}} \cdot \cos \left(6.28318530718 \cdot u2\right)
\end{array}
(FPCore (cosTheta_i u1 u2) :precision binary32 (* (sqrt (/ u1 (- 1.0 u1))) (cos (* 6.28318530718 u2))))
float code(float cosTheta_i, float u1, float u2) {
return sqrtf((u1 / (1.0f - u1))) * cosf((6.28318530718f * u2));
}
module fmin_fmax_functions
implicit none
private
public fmax
public fmin
interface fmax
module procedure fmax88
module procedure fmax44
module procedure fmax84
module procedure fmax48
end interface
interface fmin
module procedure fmin88
module procedure fmin44
module procedure fmin84
module procedure fmin48
end interface
contains
real(8) function fmax88(x, y) result (res)
real(8), intent (in) :: x
real(8), intent (in) :: y
res = merge(y, merge(x, max(x, y), y /= y), x /= x)
end function
real(4) function fmax44(x, y) result (res)
real(4), intent (in) :: x
real(4), intent (in) :: y
res = merge(y, merge(x, max(x, y), y /= y), x /= x)
end function
real(8) function fmax84(x, y) result(res)
real(8), intent (in) :: x
real(4), intent (in) :: y
res = merge(dble(y), merge(x, max(x, dble(y)), y /= y), x /= x)
end function
real(8) function fmax48(x, y) result(res)
real(4), intent (in) :: x
real(8), intent (in) :: y
res = merge(y, merge(dble(x), max(dble(x), y), y /= y), x /= x)
end function
real(8) function fmin88(x, y) result (res)
real(8), intent (in) :: x
real(8), intent (in) :: y
res = merge(y, merge(x, min(x, y), y /= y), x /= x)
end function
real(4) function fmin44(x, y) result (res)
real(4), intent (in) :: x
real(4), intent (in) :: y
res = merge(y, merge(x, min(x, y), y /= y), x /= x)
end function
real(8) function fmin84(x, y) result(res)
real(8), intent (in) :: x
real(4), intent (in) :: y
res = merge(dble(y), merge(x, min(x, dble(y)), y /= y), x /= x)
end function
real(8) function fmin48(x, y) result(res)
real(4), intent (in) :: x
real(8), intent (in) :: y
res = merge(y, merge(dble(x), min(dble(x), y), y /= y), x /= x)
end function
end module
real(4) function code(costheta_i, u1, u2)
use fmin_fmax_functions
real(4), intent (in) :: costheta_i
real(4), intent (in) :: u1
real(4), intent (in) :: u2
code = sqrt((u1 / (1.0e0 - u1))) * cos((6.28318530718e0 * u2))
end function
function code(cosTheta_i, u1, u2) return Float32(sqrt(Float32(u1 / Float32(Float32(1.0) - u1))) * cos(Float32(Float32(6.28318530718) * u2))) end
function tmp = code(cosTheta_i, u1, u2) tmp = sqrt((u1 / (single(1.0) - u1))) * cos((single(6.28318530718) * u2)); end
\begin{array}{l}
\\
\sqrt{\frac{u1}{1 - u1}} \cdot \cos \left(6.28318530718 \cdot u2\right)
\end{array}
Initial program 99.0%
(FPCore (cosTheta_i u1 u2)
:precision binary32
(let* ((t_0 (sqrt (/ u1 (- 1.0 u1)))))
(if (<= (* t_0 (cos (* 6.28318530718 u2))) 0.0010499999625608325)
(* (sqrt u1) (fma (* u2 u2) -19.739208802181317 1.0))
t_0)))
float code(float cosTheta_i, float u1, float u2) {
float t_0 = sqrtf((u1 / (1.0f - u1)));
float tmp;
if ((t_0 * cosf((6.28318530718f * u2))) <= 0.0010499999625608325f) {
tmp = sqrtf(u1) * fmaf((u2 * u2), -19.739208802181317f, 1.0f);
} else {
tmp = t_0;
}
return tmp;
}
function code(cosTheta_i, u1, u2) t_0 = sqrt(Float32(u1 / Float32(Float32(1.0) - u1))) tmp = Float32(0.0) if (Float32(t_0 * cos(Float32(Float32(6.28318530718) * u2))) <= Float32(0.0010499999625608325)) tmp = Float32(sqrt(u1) * fma(Float32(u2 * u2), Float32(-19.739208802181317), Float32(1.0))); else tmp = t_0; end return tmp end
\begin{array}{l}
\\
\begin{array}{l}
t_0 := \sqrt{\frac{u1}{1 - u1}}\\
\mathbf{if}\;t\_0 \cdot \cos \left(6.28318530718 \cdot u2\right) \leq 0.0010499999625608325:\\
\;\;\;\;\sqrt{u1} \cdot \mathsf{fma}\left(u2 \cdot u2, -19.739208802181317, 1\right)\\
\mathbf{else}:\\
\;\;\;\;t\_0\\
\end{array}
\end{array}
if (*.f32 (sqrt.f32 (/.f32 u1 (-.f32 #s(literal 1 binary32) u1))) (cos.f32 (*.f32 #s(literal 314159265359/50000000000 binary32) u2))) < 0.00104999996Initial program 98.8%
Taylor expanded in u1 around 0
*-commutativeN/A
lower-*.f32N/A
+-commutativeN/A
*-commutativeN/A
lower-fma.f32N/A
lower-+.f3298.1
Applied rewrites98.1%
Taylor expanded in u2 around 0
Applied rewrites74.7%
Taylor expanded in u1 around 0
Applied rewrites74.0%
Taylor expanded in u2 around 0
+-commutativeN/A
*-commutativeN/A
lower-fma.f32N/A
pow2N/A
lift-*.f3282.3
Applied rewrites82.3%
if 0.00104999996 < (*.f32 (sqrt.f32 (/.f32 u1 (-.f32 #s(literal 1 binary32) u1))) (cos.f32 (*.f32 #s(literal 314159265359/50000000000 binary32) u2))) Initial program 99.1%
Taylor expanded in u2 around 0
lift-/.f32N/A
lift--.f32N/A
lift-sqrt.f3283.7
Applied rewrites83.7%
(FPCore (cosTheta_i u1 u2)
:precision binary32
(if (<= u2 0.125)
(*
(sqrt (/ u1 (- 1.0 u1)))
(fma
(-
(*
(/
-4217.124896033587
(- (* (* u2 u2) -85.45681720672748) 64.93939402268539))
(* u2 u2))
19.739208802181317)
(* u2 u2)
1.0))
(* (sqrt (* (fma (+ 1.0 u1) u1 1.0) u1)) (cos (* 6.28318530718 u2)))))
float code(float cosTheta_i, float u1, float u2) {
float tmp;
if (u2 <= 0.125f) {
tmp = sqrtf((u1 / (1.0f - u1))) * fmaf((((-4217.124896033587f / (((u2 * u2) * -85.45681720672748f) - 64.93939402268539f)) * (u2 * u2)) - 19.739208802181317f), (u2 * u2), 1.0f);
} else {
tmp = sqrtf((fmaf((1.0f + u1), u1, 1.0f) * u1)) * cosf((6.28318530718f * u2));
}
return tmp;
}
function code(cosTheta_i, u1, u2) tmp = Float32(0.0) if (u2 <= Float32(0.125)) tmp = Float32(sqrt(Float32(u1 / Float32(Float32(1.0) - u1))) * fma(Float32(Float32(Float32(Float32(-4217.124896033587) / Float32(Float32(Float32(u2 * u2) * Float32(-85.45681720672748)) - Float32(64.93939402268539))) * Float32(u2 * u2)) - Float32(19.739208802181317)), Float32(u2 * u2), Float32(1.0))); else tmp = Float32(sqrt(Float32(fma(Float32(Float32(1.0) + u1), u1, Float32(1.0)) * u1)) * cos(Float32(Float32(6.28318530718) * u2))); end return tmp end
\begin{array}{l}
\\
\begin{array}{l}
\mathbf{if}\;u2 \leq 0.125:\\
\;\;\;\;\sqrt{\frac{u1}{1 - u1}} \cdot \mathsf{fma}\left(\frac{-4217.124896033587}{\left(u2 \cdot u2\right) \cdot -85.45681720672748 - 64.93939402268539} \cdot \left(u2 \cdot u2\right) - 19.739208802181317, u2 \cdot u2, 1\right)\\
\mathbf{else}:\\
\;\;\;\;\sqrt{\mathsf{fma}\left(1 + u1, u1, 1\right) \cdot u1} \cdot \cos \left(6.28318530718 \cdot u2\right)\\
\end{array}
\end{array}
if u2 < 0.125Initial program 99.3%
Taylor expanded in u2 around 0
+-commutativeN/A
*-commutativeN/A
lower-fma.f32N/A
lower--.f32N/A
*-commutativeN/A
lower-*.f32N/A
+-commutativeN/A
lower-fma.f32N/A
unpow2N/A
lower-*.f32N/A
unpow2N/A
lower-*.f32N/A
unpow2N/A
lower-*.f3299.1
Applied rewrites99.1%
lift-fma.f32N/A
flip-+N/A
lower-/.f32N/A
lift-*.f32N/A
pow2N/A
lift-*.f32N/A
pow2N/A
lower--.f32N/A
lower-*.f32N/A
*-commutativeN/A
lower-*.f32N/A
pow2N/A
lift-*.f32N/A
*-commutativeN/A
lower-*.f32N/A
pow2N/A
lift-*.f32N/A
metadata-evalN/A
lift-*.f32N/A
pow2N/A
Applied rewrites99.1%
Taylor expanded in u2 around 0
Applied rewrites99.1%
if 0.125 < u2 Initial program 95.7%
Taylor expanded in u1 around 0
*-commutativeN/A
lower-*.f32N/A
+-commutativeN/A
*-commutativeN/A
lower-fma.f32N/A
lower-+.f3288.5
Applied rewrites88.5%
(FPCore (cosTheta_i u1 u2)
:precision binary32
(if (<= u2 0.125)
(*
(sqrt (/ u1 (- 1.0 u1)))
(fma
(-
(*
(/
-4217.124896033587
(- (* (* u2 u2) -85.45681720672748) 64.93939402268539))
(* u2 u2))
19.739208802181317)
(* u2 u2)
1.0))
(* (sqrt (* (+ 1.0 u1) u1)) (cos (* 6.28318530718 u2)))))
float code(float cosTheta_i, float u1, float u2) {
float tmp;
if (u2 <= 0.125f) {
tmp = sqrtf((u1 / (1.0f - u1))) * fmaf((((-4217.124896033587f / (((u2 * u2) * -85.45681720672748f) - 64.93939402268539f)) * (u2 * u2)) - 19.739208802181317f), (u2 * u2), 1.0f);
} else {
tmp = sqrtf(((1.0f + u1) * u1)) * cosf((6.28318530718f * u2));
}
return tmp;
}
function code(cosTheta_i, u1, u2) tmp = Float32(0.0) if (u2 <= Float32(0.125)) tmp = Float32(sqrt(Float32(u1 / Float32(Float32(1.0) - u1))) * fma(Float32(Float32(Float32(Float32(-4217.124896033587) / Float32(Float32(Float32(u2 * u2) * Float32(-85.45681720672748)) - Float32(64.93939402268539))) * Float32(u2 * u2)) - Float32(19.739208802181317)), Float32(u2 * u2), Float32(1.0))); else tmp = Float32(sqrt(Float32(Float32(Float32(1.0) + u1) * u1)) * cos(Float32(Float32(6.28318530718) * u2))); end return tmp end
\begin{array}{l}
\\
\begin{array}{l}
\mathbf{if}\;u2 \leq 0.125:\\
\;\;\;\;\sqrt{\frac{u1}{1 - u1}} \cdot \mathsf{fma}\left(\frac{-4217.124896033587}{\left(u2 \cdot u2\right) \cdot -85.45681720672748 - 64.93939402268539} \cdot \left(u2 \cdot u2\right) - 19.739208802181317, u2 \cdot u2, 1\right)\\
\mathbf{else}:\\
\;\;\;\;\sqrt{\left(1 + u1\right) \cdot u1} \cdot \cos \left(6.28318530718 \cdot u2\right)\\
\end{array}
\end{array}
if u2 < 0.125Initial program 99.3%
Taylor expanded in u2 around 0
+-commutativeN/A
*-commutativeN/A
lower-fma.f32N/A
lower--.f32N/A
*-commutativeN/A
lower-*.f32N/A
+-commutativeN/A
lower-fma.f32N/A
unpow2N/A
lower-*.f32N/A
unpow2N/A
lower-*.f32N/A
unpow2N/A
lower-*.f3299.1
Applied rewrites99.1%
lift-fma.f32N/A
flip-+N/A
lower-/.f32N/A
lift-*.f32N/A
pow2N/A
lift-*.f32N/A
pow2N/A
lower--.f32N/A
lower-*.f32N/A
*-commutativeN/A
lower-*.f32N/A
pow2N/A
lift-*.f32N/A
*-commutativeN/A
lower-*.f32N/A
pow2N/A
lift-*.f32N/A
metadata-evalN/A
lift-*.f32N/A
pow2N/A
Applied rewrites99.1%
Taylor expanded in u2 around 0
Applied rewrites99.1%
if 0.125 < u2 Initial program 95.7%
Taylor expanded in u1 around 0
*-commutativeN/A
lower-*.f32N/A
lower-+.f3284.7
Applied rewrites84.7%
(FPCore (cosTheta_i u1 u2)
:precision binary32
(if (<= u2 0.125)
(*
(sqrt (/ u1 (- 1.0 u1)))
(fma
(-
(*
(/
-4217.124896033587
(- (* (* u2 u2) -85.45681720672748) 64.93939402268539))
(* u2 u2))
19.739208802181317)
(* u2 u2)
1.0))
(* (sqrt u1) (cos (* 6.28318530718 u2)))))
float code(float cosTheta_i, float u1, float u2) {
float tmp;
if (u2 <= 0.125f) {
tmp = sqrtf((u1 / (1.0f - u1))) * fmaf((((-4217.124896033587f / (((u2 * u2) * -85.45681720672748f) - 64.93939402268539f)) * (u2 * u2)) - 19.739208802181317f), (u2 * u2), 1.0f);
} else {
tmp = sqrtf(u1) * cosf((6.28318530718f * u2));
}
return tmp;
}
function code(cosTheta_i, u1, u2) tmp = Float32(0.0) if (u2 <= Float32(0.125)) tmp = Float32(sqrt(Float32(u1 / Float32(Float32(1.0) - u1))) * fma(Float32(Float32(Float32(Float32(-4217.124896033587) / Float32(Float32(Float32(u2 * u2) * Float32(-85.45681720672748)) - Float32(64.93939402268539))) * Float32(u2 * u2)) - Float32(19.739208802181317)), Float32(u2 * u2), Float32(1.0))); else tmp = Float32(sqrt(u1) * cos(Float32(Float32(6.28318530718) * u2))); end return tmp end
\begin{array}{l}
\\
\begin{array}{l}
\mathbf{if}\;u2 \leq 0.125:\\
\;\;\;\;\sqrt{\frac{u1}{1 - u1}} \cdot \mathsf{fma}\left(\frac{-4217.124896033587}{\left(u2 \cdot u2\right) \cdot -85.45681720672748 - 64.93939402268539} \cdot \left(u2 \cdot u2\right) - 19.739208802181317, u2 \cdot u2, 1\right)\\
\mathbf{else}:\\
\;\;\;\;\sqrt{u1} \cdot \cos \left(6.28318530718 \cdot u2\right)\\
\end{array}
\end{array}
if u2 < 0.125Initial program 99.3%
Taylor expanded in u2 around 0
+-commutativeN/A
*-commutativeN/A
lower-fma.f32N/A
lower--.f32N/A
*-commutativeN/A
lower-*.f32N/A
+-commutativeN/A
lower-fma.f32N/A
unpow2N/A
lower-*.f32N/A
unpow2N/A
lower-*.f32N/A
unpow2N/A
lower-*.f3299.1
Applied rewrites99.1%
lift-fma.f32N/A
flip-+N/A
lower-/.f32N/A
lift-*.f32N/A
pow2N/A
lift-*.f32N/A
pow2N/A
lower--.f32N/A
lower-*.f32N/A
*-commutativeN/A
lower-*.f32N/A
pow2N/A
lift-*.f32N/A
*-commutativeN/A
lower-*.f32N/A
pow2N/A
lift-*.f32N/A
metadata-evalN/A
lift-*.f32N/A
pow2N/A
Applied rewrites99.1%
Taylor expanded in u2 around 0
Applied rewrites99.1%
if 0.125 < u2 Initial program 95.7%
Taylor expanded in u1 around 0
Applied rewrites73.0%
(FPCore (cosTheta_i u1 u2)
:precision binary32
(let* ((t_0 (* (* u2 u2) -85.45681720672748)))
(*
(sqrt (/ u1 (- 1.0 u1)))
(fma
(-
(*
(/ (- (* t_0 t_0) 4217.124896033587) (- t_0 64.93939402268539))
(* u2 u2))
19.739208802181317)
(* u2 u2)
1.0))))
float code(float cosTheta_i, float u1, float u2) {
float t_0 = (u2 * u2) * -85.45681720672748f;
return sqrtf((u1 / (1.0f - u1))) * fmaf((((((t_0 * t_0) - 4217.124896033587f) / (t_0 - 64.93939402268539f)) * (u2 * u2)) - 19.739208802181317f), (u2 * u2), 1.0f);
}
function code(cosTheta_i, u1, u2) t_0 = Float32(Float32(u2 * u2) * Float32(-85.45681720672748)) return Float32(sqrt(Float32(u1 / Float32(Float32(1.0) - u1))) * fma(Float32(Float32(Float32(Float32(Float32(t_0 * t_0) - Float32(4217.124896033587)) / Float32(t_0 - Float32(64.93939402268539))) * Float32(u2 * u2)) - Float32(19.739208802181317)), Float32(u2 * u2), Float32(1.0))) end
\begin{array}{l}
\\
\begin{array}{l}
t_0 := \left(u2 \cdot u2\right) \cdot -85.45681720672748\\
\sqrt{\frac{u1}{1 - u1}} \cdot \mathsf{fma}\left(\frac{t\_0 \cdot t\_0 - 4217.124896033587}{t\_0 - 64.93939402268539} \cdot \left(u2 \cdot u2\right) - 19.739208802181317, u2 \cdot u2, 1\right)
\end{array}
\end{array}
Initial program 99.0%
Taylor expanded in u2 around 0
+-commutativeN/A
*-commutativeN/A
lower-fma.f32N/A
lower--.f32N/A
*-commutativeN/A
lower-*.f32N/A
+-commutativeN/A
lower-fma.f32N/A
unpow2N/A
lower-*.f32N/A
unpow2N/A
lower-*.f32N/A
unpow2N/A
lower-*.f3293.4
Applied rewrites93.4%
lift-fma.f32N/A
flip-+N/A
lower-/.f32N/A
lift-*.f32N/A
pow2N/A
lift-*.f32N/A
pow2N/A
lower--.f32N/A
lower-*.f32N/A
*-commutativeN/A
lower-*.f32N/A
pow2N/A
lift-*.f32N/A
*-commutativeN/A
lower-*.f32N/A
pow2N/A
lift-*.f32N/A
metadata-evalN/A
lift-*.f32N/A
pow2N/A
Applied rewrites93.4%
(FPCore (cosTheta_i u1 u2)
:precision binary32
(*
(sqrt (/ u1 (- 1.0 u1)))
(fma
(-
(* (* (fma (* u2 u2) -85.45681720672748 64.93939402268539) u2) u2)
19.739208802181317)
(* u2 u2)
1.0)))
float code(float cosTheta_i, float u1, float u2) {
return sqrtf((u1 / (1.0f - u1))) * fmaf((((fmaf((u2 * u2), -85.45681720672748f, 64.93939402268539f) * u2) * u2) - 19.739208802181317f), (u2 * u2), 1.0f);
}
function code(cosTheta_i, u1, u2) return Float32(sqrt(Float32(u1 / Float32(Float32(1.0) - u1))) * fma(Float32(Float32(Float32(fma(Float32(u2 * u2), Float32(-85.45681720672748), Float32(64.93939402268539)) * u2) * u2) - Float32(19.739208802181317)), Float32(u2 * u2), Float32(1.0))) end
\begin{array}{l}
\\
\sqrt{\frac{u1}{1 - u1}} \cdot \mathsf{fma}\left(\left(\mathsf{fma}\left(u2 \cdot u2, -85.45681720672748, 64.93939402268539\right) \cdot u2\right) \cdot u2 - 19.739208802181317, u2 \cdot u2, 1\right)
\end{array}
Initial program 99.0%
Taylor expanded in u2 around 0
+-commutativeN/A
*-commutativeN/A
lower-fma.f32N/A
lower--.f32N/A
*-commutativeN/A
lower-*.f32N/A
+-commutativeN/A
lower-fma.f32N/A
unpow2N/A
lower-*.f32N/A
unpow2N/A
lower-*.f32N/A
unpow2N/A
lower-*.f3293.4
Applied rewrites93.4%
Applied rewrites93.4%
(FPCore (cosTheta_i u1 u2)
:precision binary32
(*
(sqrt (/ u1 (- 1.0 u1)))
(fma
(*
(-
(* (* (fma (* u2 u2) -85.45681720672748 64.93939402268539) u2) u2)
19.739208802181317)
u2)
u2
1.0)))
float code(float cosTheta_i, float u1, float u2) {
return sqrtf((u1 / (1.0f - u1))) * fmaf(((((fmaf((u2 * u2), -85.45681720672748f, 64.93939402268539f) * u2) * u2) - 19.739208802181317f) * u2), u2, 1.0f);
}
function code(cosTheta_i, u1, u2) return Float32(sqrt(Float32(u1 / Float32(Float32(1.0) - u1))) * fma(Float32(Float32(Float32(Float32(fma(Float32(u2 * u2), Float32(-85.45681720672748), Float32(64.93939402268539)) * u2) * u2) - Float32(19.739208802181317)) * u2), u2, Float32(1.0))) end
\begin{array}{l}
\\
\sqrt{\frac{u1}{1 - u1}} \cdot \mathsf{fma}\left(\left(\left(\mathsf{fma}\left(u2 \cdot u2, -85.45681720672748, 64.93939402268539\right) \cdot u2\right) \cdot u2 - 19.739208802181317\right) \cdot u2, u2, 1\right)
\end{array}
Initial program 99.0%
Taylor expanded in u2 around 0
+-commutativeN/A
*-commutativeN/A
lower-fma.f32N/A
lower--.f32N/A
*-commutativeN/A
lower-*.f32N/A
+-commutativeN/A
lower-fma.f32N/A
unpow2N/A
lower-*.f32N/A
unpow2N/A
lower-*.f32N/A
unpow2N/A
lower-*.f3293.4
Applied rewrites93.4%
Applied rewrites93.4%
(FPCore (cosTheta_i u1 u2) :precision binary32 (* (sqrt (/ u1 (- 1.0 u1))) (fma (- (* 64.93939402268539 (* u2 u2)) 19.739208802181317) (* u2 u2) 1.0)))
float code(float cosTheta_i, float u1, float u2) {
return sqrtf((u1 / (1.0f - u1))) * fmaf(((64.93939402268539f * (u2 * u2)) - 19.739208802181317f), (u2 * u2), 1.0f);
}
function code(cosTheta_i, u1, u2) return Float32(sqrt(Float32(u1 / Float32(Float32(1.0) - u1))) * fma(Float32(Float32(Float32(64.93939402268539) * Float32(u2 * u2)) - Float32(19.739208802181317)), Float32(u2 * u2), Float32(1.0))) end
\begin{array}{l}
\\
\sqrt{\frac{u1}{1 - u1}} \cdot \mathsf{fma}\left(64.93939402268539 \cdot \left(u2 \cdot u2\right) - 19.739208802181317, u2 \cdot u2, 1\right)
\end{array}
Initial program 99.0%
Taylor expanded in u2 around 0
+-commutativeN/A
*-commutativeN/A
lower-fma.f32N/A
lower--.f32N/A
lower-*.f32N/A
unpow2N/A
lower-*.f32N/A
unpow2N/A
lower-*.f3291.4
Applied rewrites91.4%
(FPCore (cosTheta_i u1 u2)
:precision binary32
(if (<= u2 5.500000042957254e-5)
(sqrt (/ u1 (- 1.0 u1)))
(*
(sqrt (* (fma (+ 1.0 u1) u1 1.0) u1))
(fma (* u2 u2) -19.739208802181317 1.0))))
float code(float cosTheta_i, float u1, float u2) {
float tmp;
if (u2 <= 5.500000042957254e-5f) {
tmp = sqrtf((u1 / (1.0f - u1)));
} else {
tmp = sqrtf((fmaf((1.0f + u1), u1, 1.0f) * u1)) * fmaf((u2 * u2), -19.739208802181317f, 1.0f);
}
return tmp;
}
function code(cosTheta_i, u1, u2) tmp = Float32(0.0) if (u2 <= Float32(5.500000042957254e-5)) tmp = sqrt(Float32(u1 / Float32(Float32(1.0) - u1))); else tmp = Float32(sqrt(Float32(fma(Float32(Float32(1.0) + u1), u1, Float32(1.0)) * u1)) * fma(Float32(u2 * u2), Float32(-19.739208802181317), Float32(1.0))); end return tmp end
\begin{array}{l}
\\
\begin{array}{l}
\mathbf{if}\;u2 \leq 5.500000042957254 \cdot 10^{-5}:\\
\;\;\;\;\sqrt{\frac{u1}{1 - u1}}\\
\mathbf{else}:\\
\;\;\;\;\sqrt{\mathsf{fma}\left(1 + u1, u1, 1\right) \cdot u1} \cdot \mathsf{fma}\left(u2 \cdot u2, -19.739208802181317, 1\right)\\
\end{array}
\end{array}
if u2 < 5.50000004e-5Initial program 99.6%
Taylor expanded in u2 around 0
lift-/.f32N/A
lift--.f32N/A
lift-sqrt.f3299.6
Applied rewrites99.6%
if 5.50000004e-5 < u2 Initial program 98.3%
Taylor expanded in u1 around 0
*-commutativeN/A
lower-*.f32N/A
+-commutativeN/A
*-commutativeN/A
lower-fma.f32N/A
lower-+.f3290.3
Applied rewrites90.3%
Taylor expanded in u2 around 0
+-commutativeN/A
*-commutativeN/A
lower-fma.f32N/A
pow2N/A
lift-*.f3269.2
Applied rewrites69.2%
(FPCore (cosTheta_i u1 u2)
:precision binary32
(if (<= u2 0.0007999999797903001)
(sqrt (/ u1 (- 1.0 u1)))
(*
(sqrt u1)
(fma
(- (* (* u2 u2) 64.93939402268539) 19.739208802181317)
(* u2 u2)
1.0))))
float code(float cosTheta_i, float u1, float u2) {
float tmp;
if (u2 <= 0.0007999999797903001f) {
tmp = sqrtf((u1 / (1.0f - u1)));
} else {
tmp = sqrtf(u1) * fmaf((((u2 * u2) * 64.93939402268539f) - 19.739208802181317f), (u2 * u2), 1.0f);
}
return tmp;
}
function code(cosTheta_i, u1, u2) tmp = Float32(0.0) if (u2 <= Float32(0.0007999999797903001)) tmp = sqrt(Float32(u1 / Float32(Float32(1.0) - u1))); else tmp = Float32(sqrt(u1) * fma(Float32(Float32(Float32(u2 * u2) * Float32(64.93939402268539)) - Float32(19.739208802181317)), Float32(u2 * u2), Float32(1.0))); end return tmp end
\begin{array}{l}
\\
\begin{array}{l}
\mathbf{if}\;u2 \leq 0.0007999999797903001:\\
\;\;\;\;\sqrt{\frac{u1}{1 - u1}}\\
\mathbf{else}:\\
\;\;\;\;\sqrt{u1} \cdot \mathsf{fma}\left(\left(u2 \cdot u2\right) \cdot 64.93939402268539 - 19.739208802181317, u2 \cdot u2, 1\right)\\
\end{array}
\end{array}
if u2 < 7.9999998e-4Initial program 99.5%
Taylor expanded in u2 around 0
lift-/.f32N/A
lift--.f32N/A
lift-sqrt.f3297.6
Applied rewrites97.6%
if 7.9999998e-4 < u2 Initial program 98.0%
Taylor expanded in u1 around 0
*-commutativeN/A
lower-*.f32N/A
+-commutativeN/A
*-commutativeN/A
lower-fma.f32N/A
lower-+.f3289.8
Applied rewrites89.8%
Taylor expanded in u2 around 0
Applied rewrites42.6%
Taylor expanded in u1 around 0
Applied rewrites41.0%
Taylor expanded in u2 around 0
+-commutativeN/A
*-commutativeN/A
lower-fma.f32N/A
lower--.f32N/A
*-commutativeN/A
lower-*.f32N/A
pow2N/A
lift-*.f32N/A
pow2N/A
lift-*.f3259.4
Applied rewrites59.4%
(FPCore (cosTheta_i u1 u2) :precision binary32 (* (sqrt (/ u1 (- 1.0 u1))) (fma (* u2 u2) -19.739208802181317 1.0)))
float code(float cosTheta_i, float u1, float u2) {
return sqrtf((u1 / (1.0f - u1))) * fmaf((u2 * u2), -19.739208802181317f, 1.0f);
}
function code(cosTheta_i, u1, u2) return Float32(sqrt(Float32(u1 / Float32(Float32(1.0) - u1))) * fma(Float32(u2 * u2), Float32(-19.739208802181317), Float32(1.0))) end
\begin{array}{l}
\\
\sqrt{\frac{u1}{1 - u1}} \cdot \mathsf{fma}\left(u2 \cdot u2, -19.739208802181317, 1\right)
\end{array}
Initial program 99.0%
Taylor expanded in u2 around 0
+-commutativeN/A
*-commutativeN/A
lower-fma.f32N/A
unpow2N/A
lower-*.f3288.2
Applied rewrites88.2%
(FPCore (cosTheta_i u1 u2) :precision binary32 (sqrt (/ u1 (- 1.0 u1))))
float code(float cosTheta_i, float u1, float u2) {
return sqrtf((u1 / (1.0f - u1)));
}
module fmin_fmax_functions
implicit none
private
public fmax
public fmin
interface fmax
module procedure fmax88
module procedure fmax44
module procedure fmax84
module procedure fmax48
end interface
interface fmin
module procedure fmin88
module procedure fmin44
module procedure fmin84
module procedure fmin48
end interface
contains
real(8) function fmax88(x, y) result (res)
real(8), intent (in) :: x
real(8), intent (in) :: y
res = merge(y, merge(x, max(x, y), y /= y), x /= x)
end function
real(4) function fmax44(x, y) result (res)
real(4), intent (in) :: x
real(4), intent (in) :: y
res = merge(y, merge(x, max(x, y), y /= y), x /= x)
end function
real(8) function fmax84(x, y) result(res)
real(8), intent (in) :: x
real(4), intent (in) :: y
res = merge(dble(y), merge(x, max(x, dble(y)), y /= y), x /= x)
end function
real(8) function fmax48(x, y) result(res)
real(4), intent (in) :: x
real(8), intent (in) :: y
res = merge(y, merge(dble(x), max(dble(x), y), y /= y), x /= x)
end function
real(8) function fmin88(x, y) result (res)
real(8), intent (in) :: x
real(8), intent (in) :: y
res = merge(y, merge(x, min(x, y), y /= y), x /= x)
end function
real(4) function fmin44(x, y) result (res)
real(4), intent (in) :: x
real(4), intent (in) :: y
res = merge(y, merge(x, min(x, y), y /= y), x /= x)
end function
real(8) function fmin84(x, y) result(res)
real(8), intent (in) :: x
real(4), intent (in) :: y
res = merge(dble(y), merge(x, min(x, dble(y)), y /= y), x /= x)
end function
real(8) function fmin48(x, y) result(res)
real(4), intent (in) :: x
real(8), intent (in) :: y
res = merge(y, merge(dble(x), min(dble(x), y), y /= y), x /= x)
end function
end module
real(4) function code(costheta_i, u1, u2)
use fmin_fmax_functions
real(4), intent (in) :: costheta_i
real(4), intent (in) :: u1
real(4), intent (in) :: u2
code = sqrt((u1 / (1.0e0 - u1)))
end function
function code(cosTheta_i, u1, u2) return sqrt(Float32(u1 / Float32(Float32(1.0) - u1))) end
function tmp = code(cosTheta_i, u1, u2) tmp = sqrt((u1 / (single(1.0) - u1))); end
\begin{array}{l}
\\
\sqrt{\frac{u1}{1 - u1}}
\end{array}
Initial program 99.0%
Taylor expanded in u2 around 0
lift-/.f32N/A
lift--.f32N/A
lift-sqrt.f3280.0
Applied rewrites80.0%
(FPCore (cosTheta_i u1 u2) :precision binary32 (* (sqrt (* (+ 1.0 u1) u1)) 1.0))
float code(float cosTheta_i, float u1, float u2) {
return sqrtf(((1.0f + u1) * u1)) * 1.0f;
}
module fmin_fmax_functions
implicit none
private
public fmax
public fmin
interface fmax
module procedure fmax88
module procedure fmax44
module procedure fmax84
module procedure fmax48
end interface
interface fmin
module procedure fmin88
module procedure fmin44
module procedure fmin84
module procedure fmin48
end interface
contains
real(8) function fmax88(x, y) result (res)
real(8), intent (in) :: x
real(8), intent (in) :: y
res = merge(y, merge(x, max(x, y), y /= y), x /= x)
end function
real(4) function fmax44(x, y) result (res)
real(4), intent (in) :: x
real(4), intent (in) :: y
res = merge(y, merge(x, max(x, y), y /= y), x /= x)
end function
real(8) function fmax84(x, y) result(res)
real(8), intent (in) :: x
real(4), intent (in) :: y
res = merge(dble(y), merge(x, max(x, dble(y)), y /= y), x /= x)
end function
real(8) function fmax48(x, y) result(res)
real(4), intent (in) :: x
real(8), intent (in) :: y
res = merge(y, merge(dble(x), max(dble(x), y), y /= y), x /= x)
end function
real(8) function fmin88(x, y) result (res)
real(8), intent (in) :: x
real(8), intent (in) :: y
res = merge(y, merge(x, min(x, y), y /= y), x /= x)
end function
real(4) function fmin44(x, y) result (res)
real(4), intent (in) :: x
real(4), intent (in) :: y
res = merge(y, merge(x, min(x, y), y /= y), x /= x)
end function
real(8) function fmin84(x, y) result(res)
real(8), intent (in) :: x
real(4), intent (in) :: y
res = merge(dble(y), merge(x, min(x, dble(y)), y /= y), x /= x)
end function
real(8) function fmin48(x, y) result(res)
real(4), intent (in) :: x
real(8), intent (in) :: y
res = merge(y, merge(dble(x), min(dble(x), y), y /= y), x /= x)
end function
end module
real(4) function code(costheta_i, u1, u2)
use fmin_fmax_functions
real(4), intent (in) :: costheta_i
real(4), intent (in) :: u1
real(4), intent (in) :: u2
code = sqrt(((1.0e0 + u1) * u1)) * 1.0e0
end function
function code(cosTheta_i, u1, u2) return Float32(sqrt(Float32(Float32(Float32(1.0) + u1) * u1)) * Float32(1.0)) end
function tmp = code(cosTheta_i, u1, u2) tmp = sqrt(((single(1.0) + u1) * u1)) * single(1.0); end
\begin{array}{l}
\\
\sqrt{\left(1 + u1\right) \cdot u1} \cdot 1
\end{array}
Initial program 99.0%
Taylor expanded in u1 around 0
*-commutativeN/A
lower-*.f32N/A
+-commutativeN/A
*-commutativeN/A
lower-fma.f32N/A
lower-+.f3290.7
Applied rewrites90.7%
Taylor expanded in u2 around 0
Applied rewrites74.4%
Taylor expanded in u1 around 0
lift-+.f3271.6
Applied rewrites71.6%
(FPCore (cosTheta_i u1 u2) :precision binary32 (* (sqrt u1) 1.0))
float code(float cosTheta_i, float u1, float u2) {
return sqrtf(u1) * 1.0f;
}
module fmin_fmax_functions
implicit none
private
public fmax
public fmin
interface fmax
module procedure fmax88
module procedure fmax44
module procedure fmax84
module procedure fmax48
end interface
interface fmin
module procedure fmin88
module procedure fmin44
module procedure fmin84
module procedure fmin48
end interface
contains
real(8) function fmax88(x, y) result (res)
real(8), intent (in) :: x
real(8), intent (in) :: y
res = merge(y, merge(x, max(x, y), y /= y), x /= x)
end function
real(4) function fmax44(x, y) result (res)
real(4), intent (in) :: x
real(4), intent (in) :: y
res = merge(y, merge(x, max(x, y), y /= y), x /= x)
end function
real(8) function fmax84(x, y) result(res)
real(8), intent (in) :: x
real(4), intent (in) :: y
res = merge(dble(y), merge(x, max(x, dble(y)), y /= y), x /= x)
end function
real(8) function fmax48(x, y) result(res)
real(4), intent (in) :: x
real(8), intent (in) :: y
res = merge(y, merge(dble(x), max(dble(x), y), y /= y), x /= x)
end function
real(8) function fmin88(x, y) result (res)
real(8), intent (in) :: x
real(8), intent (in) :: y
res = merge(y, merge(x, min(x, y), y /= y), x /= x)
end function
real(4) function fmin44(x, y) result (res)
real(4), intent (in) :: x
real(4), intent (in) :: y
res = merge(y, merge(x, min(x, y), y /= y), x /= x)
end function
real(8) function fmin84(x, y) result(res)
real(8), intent (in) :: x
real(4), intent (in) :: y
res = merge(dble(y), merge(x, min(x, dble(y)), y /= y), x /= x)
end function
real(8) function fmin48(x, y) result(res)
real(4), intent (in) :: x
real(8), intent (in) :: y
res = merge(y, merge(dble(x), min(dble(x), y), y /= y), x /= x)
end function
end module
real(4) function code(costheta_i, u1, u2)
use fmin_fmax_functions
real(4), intent (in) :: costheta_i
real(4), intent (in) :: u1
real(4), intent (in) :: u2
code = sqrt(u1) * 1.0e0
end function
function code(cosTheta_i, u1, u2) return Float32(sqrt(u1) * Float32(1.0)) end
function tmp = code(cosTheta_i, u1, u2) tmp = sqrt(u1) * single(1.0); end
\begin{array}{l}
\\
\sqrt{u1} \cdot 1
\end{array}
Initial program 99.0%
Taylor expanded in u1 around 0
*-commutativeN/A
lower-*.f32N/A
+-commutativeN/A
*-commutativeN/A
lower-fma.f32N/A
lower-+.f3290.7
Applied rewrites90.7%
Taylor expanded in u2 around 0
Applied rewrites74.4%
Taylor expanded in u1 around 0
Applied rewrites63.3%
herbie shell --seed 2025086
(FPCore (cosTheta_i u1 u2)
:name "Trowbridge-Reitz Sample, near normal, slope_x"
:precision binary32
:pre (and (and (and (> cosTheta_i 0.9999) (<= cosTheta_i 1.0)) (and (<= 2.328306437e-10 u1) (<= u1 1.0))) (and (<= 2.328306437e-10 u2) (<= u2 1.0)))
(* (sqrt (/ u1 (- 1.0 u1))) (cos (* 6.28318530718 u2))))