
(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));
}
real(4) function code(costheta_i, u1, u2)
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}
Sampling outcomes in binary32 precision:
Herbie found 8 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));
}
real(4) function code(costheta_i, u1, u2)
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 (* (pow (+ (/ 1.0 u1) -1.0) -0.5) (cos (* 6.28318530718 u2))))
float code(float cosTheta_i, float u1, float u2) {
return powf(((1.0f / u1) + -1.0f), -0.5f) * cosf((6.28318530718f * u2));
}
real(4) function code(costheta_i, u1, u2)
real(4), intent (in) :: costheta_i
real(4), intent (in) :: u1
real(4), intent (in) :: u2
code = (((1.0e0 / u1) + (-1.0e0)) ** (-0.5e0)) * cos((6.28318530718e0 * u2))
end function
function code(cosTheta_i, u1, u2) return Float32((Float32(Float32(Float32(1.0) / u1) + Float32(-1.0)) ^ Float32(-0.5)) * cos(Float32(Float32(6.28318530718) * u2))) end
function tmp = code(cosTheta_i, u1, u2) tmp = (((single(1.0) / u1) + single(-1.0)) ^ single(-0.5)) * cos((single(6.28318530718) * u2)); end
\begin{array}{l}
\\
{\left(\frac{1}{u1} + -1\right)}^{-0.5} \cdot \cos \left(6.28318530718 \cdot u2\right)
\end{array}
Initial program 99.0%
add-cbrt-cube99.0%
add-cbrt-cube99.0%
cbrt-unprod99.0%
add-sqr-sqrt99.0%
pow199.0%
pow1/299.0%
pow-prod-up99.0%
metadata-eval99.0%
pow399.0%
Applied egg-rr99.0%
add-sqr-sqrt95.2%
sqrt-unprod96.2%
cbrt-prod96.2%
rem-cbrt-cube96.2%
cbrt-prod96.1%
rem-cbrt-cube96.1%
swap-sqr96.0%
Applied egg-rr98.6%
associate-/r/98.7%
pow1/298.7%
pow-flip99.1%
div-sub99.1%
*-inverses99.1%
sub-neg99.1%
metadata-eval99.1%
metadata-eval99.1%
Applied egg-rr99.1%
Final simplification99.1%
(FPCore (cosTheta_i u1 u2) :precision binary32 (if (<= (* 6.28318530718 u2) 0.004600000102072954) (pow (+ (/ 1.0 u1) -1.0) -0.5) (* (cos (* 6.28318530718 u2)) (sqrt u1))))
float code(float cosTheta_i, float u1, float u2) {
float tmp;
if ((6.28318530718f * u2) <= 0.004600000102072954f) {
tmp = powf(((1.0f / u1) + -1.0f), -0.5f);
} else {
tmp = cosf((6.28318530718f * u2)) * sqrtf(u1);
}
return tmp;
}
real(4) function code(costheta_i, u1, u2)
real(4), intent (in) :: costheta_i
real(4), intent (in) :: u1
real(4), intent (in) :: u2
real(4) :: tmp
if ((6.28318530718e0 * u2) <= 0.004600000102072954e0) then
tmp = ((1.0e0 / u1) + (-1.0e0)) ** (-0.5e0)
else
tmp = cos((6.28318530718e0 * u2)) * sqrt(u1)
end if
code = tmp
end function
function code(cosTheta_i, u1, u2) tmp = Float32(0.0) if (Float32(Float32(6.28318530718) * u2) <= Float32(0.004600000102072954)) tmp = Float32(Float32(Float32(1.0) / u1) + Float32(-1.0)) ^ Float32(-0.5); else tmp = Float32(cos(Float32(Float32(6.28318530718) * u2)) * sqrt(u1)); end return tmp end
function tmp_2 = code(cosTheta_i, u1, u2) tmp = single(0.0); if ((single(6.28318530718) * u2) <= single(0.004600000102072954)) tmp = ((single(1.0) / u1) + single(-1.0)) ^ single(-0.5); else tmp = cos((single(6.28318530718) * u2)) * sqrt(u1); end tmp_2 = tmp; end
\begin{array}{l}
\\
\begin{array}{l}
\mathbf{if}\;6.28318530718 \cdot u2 \leq 0.004600000102072954:\\
\;\;\;\;{\left(\frac{1}{u1} + -1\right)}^{-0.5}\\
\mathbf{else}:\\
\;\;\;\;\cos \left(6.28318530718 \cdot u2\right) \cdot \sqrt{u1}\\
\end{array}
\end{array}
if (*.f32 314159265359/50000000000 u2) < 0.0046000001Initial program 99.4%
Taylor expanded in u2 around 0 97.9%
add-cbrt-cube97.9%
add-sqr-sqrt97.9%
pow197.9%
pow1/297.9%
pow-prod-up97.8%
metadata-eval97.8%
Applied egg-rr97.8%
pow1/395.0%
pow-pow97.9%
metadata-eval97.9%
pow1/297.9%
clear-num97.8%
sqrt-div97.4%
metadata-eval97.4%
pow1/297.4%
pow-flip97.8%
div-sub97.9%
*-inverses97.9%
sub-neg97.9%
metadata-eval97.9%
metadata-eval97.9%
Applied egg-rr97.9%
if 0.0046000001 < (*.f32 314159265359/50000000000 u2) Initial program 98.2%
Taylor expanded in u1 around 0 77.4%
Final simplification91.8%
(FPCore (cosTheta_i u1 u2) :precision binary32 (* (cos (* 6.28318530718 u2)) (sqrt (/ u1 (- 1.0 u1)))))
float code(float cosTheta_i, float u1, float u2) {
return cosf((6.28318530718f * u2)) * sqrtf((u1 / (1.0f - u1)));
}
real(4) function code(costheta_i, u1, u2)
real(4), intent (in) :: costheta_i
real(4), intent (in) :: u1
real(4), intent (in) :: u2
code = cos((6.28318530718e0 * u2)) * sqrt((u1 / (1.0e0 - u1)))
end function
function code(cosTheta_i, u1, u2) return Float32(cos(Float32(Float32(6.28318530718) * u2)) * sqrt(Float32(u1 / Float32(Float32(1.0) - u1)))) end
function tmp = code(cosTheta_i, u1, u2) tmp = cos((single(6.28318530718) * u2)) * sqrt((u1 / (single(1.0) - u1))); end
\begin{array}{l}
\\
\cos \left(6.28318530718 \cdot u2\right) \cdot \sqrt{\frac{u1}{1 - u1}}
\end{array}
Initial program 99.0%
Final simplification99.0%
(FPCore (cosTheta_i u1 u2) :precision binary32 (pow (+ (/ 1.0 u1) -1.0) -0.5))
float code(float cosTheta_i, float u1, float u2) {
return powf(((1.0f / u1) + -1.0f), -0.5f);
}
real(4) function code(costheta_i, u1, u2)
real(4), intent (in) :: costheta_i
real(4), intent (in) :: u1
real(4), intent (in) :: u2
code = ((1.0e0 / u1) + (-1.0e0)) ** (-0.5e0)
end function
function code(cosTheta_i, u1, u2) return Float32(Float32(Float32(1.0) / u1) + Float32(-1.0)) ^ Float32(-0.5) end
function tmp = code(cosTheta_i, u1, u2) tmp = ((single(1.0) / u1) + single(-1.0)) ^ single(-0.5); end
\begin{array}{l}
\\
{\left(\frac{1}{u1} + -1\right)}^{-0.5}
\end{array}
Initial program 99.0%
Taylor expanded in u2 around 0 82.9%
add-cbrt-cube82.8%
add-sqr-sqrt82.8%
pow182.8%
pow1/282.8%
pow-prod-up82.8%
metadata-eval82.8%
Applied egg-rr82.8%
pow1/380.8%
pow-pow82.9%
metadata-eval82.9%
pow1/282.9%
clear-num82.8%
sqrt-div82.5%
metadata-eval82.5%
pow1/282.5%
pow-flip82.8%
div-sub82.9%
*-inverses82.9%
sub-neg82.9%
metadata-eval82.9%
metadata-eval82.9%
Applied egg-rr82.9%
Final simplification82.9%
(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)));
}
real(4) function code(costheta_i, u1, u2)
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{u1 \cdot \left(1 + u1\right)}
\end{array}
Initial program 99.0%
Taylor expanded in u2 around 0 82.9%
Taylor expanded in u1 around 0 76.4%
+-commutative76.4%
unpow276.4%
cube-mult76.4%
unpow276.4%
distribute-lft-out76.4%
unpow276.4%
distribute-rgt1-in76.4%
*-commutative76.4%
*-rgt-identity76.4%
distribute-lft-in76.3%
fma-udef76.3%
Simplified76.3%
Taylor expanded in u1 around 0 73.1%
Final simplification73.1%
(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)));
}
real(4) function code(costheta_i, u1, u2)
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 82.9%
Final simplification82.9%
(FPCore (cosTheta_i u1 u2) :precision binary32 (sqrt u1))
float code(float cosTheta_i, float u1, float u2) {
return sqrtf(u1);
}
real(4) function code(costheta_i, u1, u2)
real(4), intent (in) :: costheta_i
real(4), intent (in) :: u1
real(4), intent (in) :: u2
code = sqrt(u1)
end function
function code(cosTheta_i, u1, u2) return sqrt(u1) end
function tmp = code(cosTheta_i, u1, u2) tmp = sqrt(u1); end
\begin{array}{l}
\\
\sqrt{u1}
\end{array}
Initial program 99.0%
Taylor expanded in u2 around 0 82.9%
Taylor expanded in u1 around 0 64.0%
Final simplification64.0%
(FPCore (cosTheta_i u1 u2) :precision binary32 (+ u1 0.5))
float code(float cosTheta_i, float u1, float u2) {
return u1 + 0.5f;
}
real(4) function code(costheta_i, u1, u2)
real(4), intent (in) :: costheta_i
real(4), intent (in) :: u1
real(4), intent (in) :: u2
code = u1 + 0.5e0
end function
function code(cosTheta_i, u1, u2) return Float32(u1 + Float32(0.5)) end
function tmp = code(cosTheta_i, u1, u2) tmp = u1 + single(0.5); end
\begin{array}{l}
\\
u1 + 0.5
\end{array}
Initial program 99.0%
Taylor expanded in u2 around 0 82.9%
Taylor expanded in u1 around 0 76.4%
+-commutative76.4%
unpow276.4%
cube-mult76.4%
unpow276.4%
distribute-lft-out76.4%
unpow276.4%
distribute-rgt1-in76.4%
*-commutative76.4%
*-rgt-identity76.4%
distribute-lft-in76.3%
fma-udef76.3%
Simplified76.3%
Taylor expanded in u1 around 0 73.1%
Taylor expanded in u1 around inf 20.7%
+-commutative20.7%
Simplified20.7%
Final simplification20.7%
herbie shell --seed 2023319
(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))))