
(FPCore (cosTheta_i u1 u2) :precision binary32 (* (sqrt (- (log (- 1.0 u1)))) (cos (* (* 2.0 PI) u2))))
float code(float cosTheta_i, float u1, float u2) {
return sqrtf(-logf((1.0f - u1))) * cosf(((2.0f * ((float) M_PI)) * u2));
}
function code(cosTheta_i, u1, u2) return Float32(sqrt(Float32(-log(Float32(Float32(1.0) - u1)))) * cos(Float32(Float32(Float32(2.0) * Float32(pi)) * u2))) end
function tmp = code(cosTheta_i, u1, u2) tmp = sqrt(-log((single(1.0) - u1))) * cos(((single(2.0) * single(pi)) * u2)); end
\begin{array}{l}
\\
\sqrt{-\log \left(1 - u1\right)} \cdot \cos \left(\left(2 \cdot \pi\right) \cdot u2\right)
\end{array}
Sampling outcomes in binary32 precision:
Herbie found 13 alternatives:
| Alternative | Accuracy | Speedup |
|---|
(FPCore (cosTheta_i u1 u2) :precision binary32 (* (sqrt (- (log (- 1.0 u1)))) (cos (* (* 2.0 PI) u2))))
float code(float cosTheta_i, float u1, float u2) {
return sqrtf(-logf((1.0f - u1))) * cosf(((2.0f * ((float) M_PI)) * u2));
}
function code(cosTheta_i, u1, u2) return Float32(sqrt(Float32(-log(Float32(Float32(1.0) - u1)))) * cos(Float32(Float32(Float32(2.0) * Float32(pi)) * u2))) end
function tmp = code(cosTheta_i, u1, u2) tmp = sqrt(-log((single(1.0) - u1))) * cos(((single(2.0) * single(pi)) * u2)); end
\begin{array}{l}
\\
\sqrt{-\log \left(1 - u1\right)} \cdot \cos \left(\left(2 \cdot \pi\right) \cdot u2\right)
\end{array}
(FPCore (cosTheta_i u1 u2)
:precision binary32
(let* ((t_0 (sin (* u2 PI))))
(*
(sqrt (- (log1p (- u1))))
(+ 0.5 (fma t_0 (- t_0) (* 0.5 (cos (* (* u2 PI) 2.0))))))))
float code(float cosTheta_i, float u1, float u2) {
float t_0 = sinf((u2 * ((float) M_PI)));
return sqrtf(-log1pf(-u1)) * (0.5f + fmaf(t_0, -t_0, (0.5f * cosf(((u2 * ((float) M_PI)) * 2.0f)))));
}
function code(cosTheta_i, u1, u2) t_0 = sin(Float32(u2 * Float32(pi))) return Float32(sqrt(Float32(-log1p(Float32(-u1)))) * Float32(Float32(0.5) + fma(t_0, Float32(-t_0), Float32(Float32(0.5) * cos(Float32(Float32(u2 * Float32(pi)) * Float32(2.0))))))) end
\begin{array}{l}
\\
\begin{array}{l}
t_0 := \sin \left(u2 \cdot \pi\right)\\
\sqrt{-\mathsf{log1p}\left(-u1\right)} \cdot \left(0.5 + \mathsf{fma}\left(t\_0, -t\_0, 0.5 \cdot \cos \left(\left(u2 \cdot \pi\right) \cdot 2\right)\right)\right)
\end{array}
\end{array}
Initial program 56.3%
*-lowering-*.f32N/A
sqrt-lowering-sqrt.f32N/A
neg-lowering-neg.f32N/A
sub-negN/A
log1p-defineN/A
log1p-lowering-log1p.f32N/A
neg-lowering-neg.f32N/A
cos-lowering-cos.f32N/A
associate-*l*N/A
*-lowering-*.f32N/A
*-lowering-*.f32N/A
PI-lowering-PI.f3299.0%
Simplified99.0%
cos-2N/A
sub-negN/A
+-commutativeN/A
distribute-rgt-neg-inN/A
fma-defineN/A
fma-lowering-fma.f32N/A
sin-lowering-sin.f32N/A
*-lowering-*.f32N/A
PI-lowering-PI.f32N/A
neg-lowering-neg.f32N/A
sin-lowering-sin.f32N/A
*-lowering-*.f32N/A
PI-lowering-PI.f32N/A
sqr-cos-aN/A
Applied egg-rr98.9%
Taylor expanded in u2 around inf
+-lowering-+.f32N/A
+-commutativeN/A
mul-1-negN/A
unsub-negN/A
--lowering--.f32N/A
*-commutativeN/A
*-lowering-*.f32N/A
cos-lowering-cos.f32N/A
*-lowering-*.f32N/A
*-lowering-*.f32N/A
PI-lowering-PI.f32N/A
pow-lowering-pow.f32N/A
sin-lowering-sin.f32N/A
*-lowering-*.f32N/A
PI-lowering-PI.f3299.0%
Simplified99.0%
sub-negN/A
+-commutativeN/A
unpow2N/A
distribute-rgt-neg-inN/A
fma-defineN/A
fma-lowering-fma.f32N/A
sin-lowering-sin.f32N/A
*-lowering-*.f32N/A
PI-lowering-PI.f32N/A
neg-lowering-neg.f32N/A
sin-lowering-sin.f32N/A
*-lowering-*.f32N/A
PI-lowering-PI.f32N/A
*-commutativeN/A
*-lowering-*.f32N/A
cos-lowering-cos.f32N/A
Applied egg-rr99.0%
Final simplification99.0%
(FPCore (cosTheta_i u1 u2) :precision binary32 (* (sqrt (- (log1p (- u1)))) (cos (* (* u2 PI) 2.0))))
float code(float cosTheta_i, float u1, float u2) {
return sqrtf(-log1pf(-u1)) * cosf(((u2 * ((float) M_PI)) * 2.0f));
}
function code(cosTheta_i, u1, u2) return Float32(sqrt(Float32(-log1p(Float32(-u1)))) * cos(Float32(Float32(u2 * Float32(pi)) * Float32(2.0)))) end
\begin{array}{l}
\\
\sqrt{-\mathsf{log1p}\left(-u1\right)} \cdot \cos \left(\left(u2 \cdot \pi\right) \cdot 2\right)
\end{array}
Initial program 56.3%
*-lowering-*.f32N/A
sqrt-lowering-sqrt.f32N/A
neg-lowering-neg.f32N/A
sub-negN/A
log1p-defineN/A
log1p-lowering-log1p.f32N/A
neg-lowering-neg.f32N/A
cos-lowering-cos.f32N/A
associate-*l*N/A
*-lowering-*.f32N/A
*-lowering-*.f32N/A
PI-lowering-PI.f3299.0%
Simplified99.0%
Final simplification99.0%
(FPCore (cosTheta_i u1 u2)
:precision binary32
(let* ((t_0 (* u2 (* PI 2.0))))
(if (<= t_0 0.07000000029802322)
(* (sqrt (- (log1p (- u1)))) (+ 1.0 (* (* u2 u2) (* PI (* PI -2.0)))))
(*
(sqrt
(* u1 (+ 1.0 (* u1 (+ 0.5 (* u1 (+ 0.3333333333333333 (* u1 0.25))))))))
(cos t_0)))))
float code(float cosTheta_i, float u1, float u2) {
float t_0 = u2 * (((float) M_PI) * 2.0f);
float tmp;
if (t_0 <= 0.07000000029802322f) {
tmp = sqrtf(-log1pf(-u1)) * (1.0f + ((u2 * u2) * (((float) M_PI) * (((float) M_PI) * -2.0f))));
} else {
tmp = sqrtf((u1 * (1.0f + (u1 * (0.5f + (u1 * (0.3333333333333333f + (u1 * 0.25f)))))))) * cosf(t_0);
}
return tmp;
}
function code(cosTheta_i, u1, u2) t_0 = Float32(u2 * Float32(Float32(pi) * Float32(2.0))) tmp = Float32(0.0) if (t_0 <= Float32(0.07000000029802322)) tmp = Float32(sqrt(Float32(-log1p(Float32(-u1)))) * Float32(Float32(1.0) + Float32(Float32(u2 * u2) * Float32(Float32(pi) * Float32(Float32(pi) * Float32(-2.0)))))); else tmp = Float32(sqrt(Float32(u1 * Float32(Float32(1.0) + Float32(u1 * Float32(Float32(0.5) + Float32(u1 * Float32(Float32(0.3333333333333333) + Float32(u1 * Float32(0.25))))))))) * cos(t_0)); end return tmp end
\begin{array}{l}
\\
\begin{array}{l}
t_0 := u2 \cdot \left(\pi \cdot 2\right)\\
\mathbf{if}\;t\_0 \leq 0.07000000029802322:\\
\;\;\;\;\sqrt{-\mathsf{log1p}\left(-u1\right)} \cdot \left(1 + \left(u2 \cdot u2\right) \cdot \left(\pi \cdot \left(\pi \cdot -2\right)\right)\right)\\
\mathbf{else}:\\
\;\;\;\;\sqrt{u1 \cdot \left(1 + u1 \cdot \left(0.5 + u1 \cdot \left(0.3333333333333333 + u1 \cdot 0.25\right)\right)\right)} \cdot \cos t\_0\\
\end{array}
\end{array}
if (*.f32 (*.f32 #s(literal 2 binary32) (PI.f32)) u2) < 0.0700000003Initial program 59.1%
*-lowering-*.f32N/A
sqrt-lowering-sqrt.f32N/A
neg-lowering-neg.f32N/A
sub-negN/A
log1p-defineN/A
log1p-lowering-log1p.f32N/A
neg-lowering-neg.f32N/A
cos-lowering-cos.f32N/A
associate-*l*N/A
*-lowering-*.f32N/A
*-lowering-*.f32N/A
PI-lowering-PI.f3299.4%
Simplified99.4%
Taylor expanded in u2 around 0
*-commutativeN/A
associate-*r*N/A
+-lowering-+.f32N/A
*-commutativeN/A
*-lowering-*.f32N/A
unpow2N/A
*-lowering-*.f32N/A
*-commutativeN/A
unpow2N/A
associate-*l*N/A
*-lowering-*.f32N/A
PI-lowering-PI.f32N/A
*-lowering-*.f32N/A
PI-lowering-PI.f3299.1%
Simplified99.1%
if 0.0700000003 < (*.f32 (*.f32 #s(literal 2 binary32) (PI.f32)) u2) Initial program 45.2%
Taylor expanded in u1 around 0
*-lowering-*.f32N/A
+-lowering-+.f32N/A
*-lowering-*.f32N/A
+-lowering-+.f32N/A
*-lowering-*.f32N/A
+-lowering-+.f32N/A
*-commutativeN/A
*-lowering-*.f3295.9%
Simplified95.9%
Final simplification98.4%
(FPCore (cosTheta_i u1 u2)
:precision binary32
(let* ((t_0 (* u2 (* PI 2.0))))
(if (<= t_0 0.07000000029802322)
(* (sqrt (- (log1p (- u1)))) (+ 1.0 (* (* u2 u2) (* PI (* PI -2.0)))))
(*
(cos t_0)
(sqrt (* u1 (+ 1.0 (* u1 (+ 0.5 (* u1 0.3333333333333333))))))))))
float code(float cosTheta_i, float u1, float u2) {
float t_0 = u2 * (((float) M_PI) * 2.0f);
float tmp;
if (t_0 <= 0.07000000029802322f) {
tmp = sqrtf(-log1pf(-u1)) * (1.0f + ((u2 * u2) * (((float) M_PI) * (((float) M_PI) * -2.0f))));
} else {
tmp = cosf(t_0) * sqrtf((u1 * (1.0f + (u1 * (0.5f + (u1 * 0.3333333333333333f))))));
}
return tmp;
}
function code(cosTheta_i, u1, u2) t_0 = Float32(u2 * Float32(Float32(pi) * Float32(2.0))) tmp = Float32(0.0) if (t_0 <= Float32(0.07000000029802322)) tmp = Float32(sqrt(Float32(-log1p(Float32(-u1)))) * Float32(Float32(1.0) + Float32(Float32(u2 * u2) * Float32(Float32(pi) * Float32(Float32(pi) * Float32(-2.0)))))); else tmp = Float32(cos(t_0) * sqrt(Float32(u1 * Float32(Float32(1.0) + Float32(u1 * Float32(Float32(0.5) + Float32(u1 * Float32(0.3333333333333333)))))))); end return tmp end
\begin{array}{l}
\\
\begin{array}{l}
t_0 := u2 \cdot \left(\pi \cdot 2\right)\\
\mathbf{if}\;t\_0 \leq 0.07000000029802322:\\
\;\;\;\;\sqrt{-\mathsf{log1p}\left(-u1\right)} \cdot \left(1 + \left(u2 \cdot u2\right) \cdot \left(\pi \cdot \left(\pi \cdot -2\right)\right)\right)\\
\mathbf{else}:\\
\;\;\;\;\cos t\_0 \cdot \sqrt{u1 \cdot \left(1 + u1 \cdot \left(0.5 + u1 \cdot 0.3333333333333333\right)\right)}\\
\end{array}
\end{array}
if (*.f32 (*.f32 #s(literal 2 binary32) (PI.f32)) u2) < 0.0700000003Initial program 59.1%
*-lowering-*.f32N/A
sqrt-lowering-sqrt.f32N/A
neg-lowering-neg.f32N/A
sub-negN/A
log1p-defineN/A
log1p-lowering-log1p.f32N/A
neg-lowering-neg.f32N/A
cos-lowering-cos.f32N/A
associate-*l*N/A
*-lowering-*.f32N/A
*-lowering-*.f32N/A
PI-lowering-PI.f3299.4%
Simplified99.4%
Taylor expanded in u2 around 0
*-commutativeN/A
associate-*r*N/A
+-lowering-+.f32N/A
*-commutativeN/A
*-lowering-*.f32N/A
unpow2N/A
*-lowering-*.f32N/A
*-commutativeN/A
unpow2N/A
associate-*l*N/A
*-lowering-*.f32N/A
PI-lowering-PI.f32N/A
*-lowering-*.f32N/A
PI-lowering-PI.f3299.1%
Simplified99.1%
if 0.0700000003 < (*.f32 (*.f32 #s(literal 2 binary32) (PI.f32)) u2) Initial program 45.2%
Taylor expanded in u1 around 0
*-lowering-*.f32N/A
+-lowering-+.f32N/A
*-lowering-*.f32N/A
+-lowering-+.f32N/A
*-commutativeN/A
*-lowering-*.f3294.9%
Simplified94.9%
Final simplification98.2%
(FPCore (cosTheta_i u1 u2)
:precision binary32
(let* ((t_0 (* u2 (* PI 2.0))))
(if (<= t_0 0.07000000029802322)
(* (sqrt (- (log1p (- u1)))) (+ 1.0 (* (* u2 u2) (* PI (* PI -2.0)))))
(* (cos t_0) (sqrt (* u1 (+ 1.0 (* u1 0.5))))))))
float code(float cosTheta_i, float u1, float u2) {
float t_0 = u2 * (((float) M_PI) * 2.0f);
float tmp;
if (t_0 <= 0.07000000029802322f) {
tmp = sqrtf(-log1pf(-u1)) * (1.0f + ((u2 * u2) * (((float) M_PI) * (((float) M_PI) * -2.0f))));
} else {
tmp = cosf(t_0) * sqrtf((u1 * (1.0f + (u1 * 0.5f))));
}
return tmp;
}
function code(cosTheta_i, u1, u2) t_0 = Float32(u2 * Float32(Float32(pi) * Float32(2.0))) tmp = Float32(0.0) if (t_0 <= Float32(0.07000000029802322)) tmp = Float32(sqrt(Float32(-log1p(Float32(-u1)))) * Float32(Float32(1.0) + Float32(Float32(u2 * u2) * Float32(Float32(pi) * Float32(Float32(pi) * Float32(-2.0)))))); else tmp = Float32(cos(t_0) * sqrt(Float32(u1 * Float32(Float32(1.0) + Float32(u1 * Float32(0.5)))))); end return tmp end
\begin{array}{l}
\\
\begin{array}{l}
t_0 := u2 \cdot \left(\pi \cdot 2\right)\\
\mathbf{if}\;t\_0 \leq 0.07000000029802322:\\
\;\;\;\;\sqrt{-\mathsf{log1p}\left(-u1\right)} \cdot \left(1 + \left(u2 \cdot u2\right) \cdot \left(\pi \cdot \left(\pi \cdot -2\right)\right)\right)\\
\mathbf{else}:\\
\;\;\;\;\cos t\_0 \cdot \sqrt{u1 \cdot \left(1 + u1 \cdot 0.5\right)}\\
\end{array}
\end{array}
if (*.f32 (*.f32 #s(literal 2 binary32) (PI.f32)) u2) < 0.0700000003Initial program 59.1%
*-lowering-*.f32N/A
sqrt-lowering-sqrt.f32N/A
neg-lowering-neg.f32N/A
sub-negN/A
log1p-defineN/A
log1p-lowering-log1p.f32N/A
neg-lowering-neg.f32N/A
cos-lowering-cos.f32N/A
associate-*l*N/A
*-lowering-*.f32N/A
*-lowering-*.f32N/A
PI-lowering-PI.f3299.4%
Simplified99.4%
Taylor expanded in u2 around 0
*-commutativeN/A
associate-*r*N/A
+-lowering-+.f32N/A
*-commutativeN/A
*-lowering-*.f32N/A
unpow2N/A
*-lowering-*.f32N/A
*-commutativeN/A
unpow2N/A
associate-*l*N/A
*-lowering-*.f32N/A
PI-lowering-PI.f32N/A
*-lowering-*.f32N/A
PI-lowering-PI.f3299.1%
Simplified99.1%
if 0.0700000003 < (*.f32 (*.f32 #s(literal 2 binary32) (PI.f32)) u2) Initial program 45.2%
Taylor expanded in u1 around 0
*-lowering-*.f32N/A
+-lowering-+.f32N/A
*-commutativeN/A
*-lowering-*.f3292.3%
Simplified92.3%
Final simplification97.7%
(FPCore (cosTheta_i u1 u2)
:precision binary32
(let* ((t_0 (* u2 (* PI 2.0))))
(if (<= t_0 0.0010000000474974513)
(sqrt (- (log1p (- u1))))
(* (cos t_0) (sqrt (* u1 (+ 1.0 (* u1 0.5))))))))
float code(float cosTheta_i, float u1, float u2) {
float t_0 = u2 * (((float) M_PI) * 2.0f);
float tmp;
if (t_0 <= 0.0010000000474974513f) {
tmp = sqrtf(-log1pf(-u1));
} else {
tmp = cosf(t_0) * sqrtf((u1 * (1.0f + (u1 * 0.5f))));
}
return tmp;
}
function code(cosTheta_i, u1, u2) t_0 = Float32(u2 * Float32(Float32(pi) * Float32(2.0))) tmp = Float32(0.0) if (t_0 <= Float32(0.0010000000474974513)) tmp = sqrt(Float32(-log1p(Float32(-u1)))); else tmp = Float32(cos(t_0) * sqrt(Float32(u1 * Float32(Float32(1.0) + Float32(u1 * Float32(0.5)))))); end return tmp end
\begin{array}{l}
\\
\begin{array}{l}
t_0 := u2 \cdot \left(\pi \cdot 2\right)\\
\mathbf{if}\;t\_0 \leq 0.0010000000474974513:\\
\;\;\;\;\sqrt{-\mathsf{log1p}\left(-u1\right)}\\
\mathbf{else}:\\
\;\;\;\;\cos t\_0 \cdot \sqrt{u1 \cdot \left(1 + u1 \cdot 0.5\right)}\\
\end{array}
\end{array}
if (*.f32 (*.f32 #s(literal 2 binary32) (PI.f32)) u2) < 0.00100000005Initial program 59.7%
*-lowering-*.f32N/A
sqrt-lowering-sqrt.f32N/A
neg-lowering-neg.f32N/A
sub-negN/A
log1p-defineN/A
log1p-lowering-log1p.f32N/A
neg-lowering-neg.f32N/A
cos-lowering-cos.f32N/A
associate-*l*N/A
*-lowering-*.f32N/A
*-lowering-*.f32N/A
PI-lowering-PI.f3299.6%
Simplified99.6%
Taylor expanded in u2 around 0
Simplified99.2%
*-rgt-identityN/A
sqrt-lowering-sqrt.f32N/A
neg-lowering-neg.f32N/A
log1p-defineN/A
log1p-lowering-log1p.f32N/A
neg-lowering-neg.f3299.2%
Applied egg-rr99.2%
if 0.00100000005 < (*.f32 (*.f32 #s(literal 2 binary32) (PI.f32)) u2) Initial program 51.0%
Taylor expanded in u1 around 0
*-lowering-*.f32N/A
+-lowering-+.f32N/A
*-commutativeN/A
*-lowering-*.f3290.2%
Simplified90.2%
Final simplification95.6%
(FPCore (cosTheta_i u1 u2)
:precision binary32
(let* ((t_0 (* u2 (* PI 2.0))))
(if (<= t_0 0.013000000268220901)
(sqrt (- (log1p (- u1))))
(* (cos t_0) (sqrt u1)))))
float code(float cosTheta_i, float u1, float u2) {
float t_0 = u2 * (((float) M_PI) * 2.0f);
float tmp;
if (t_0 <= 0.013000000268220901f) {
tmp = sqrtf(-log1pf(-u1));
} else {
tmp = cosf(t_0) * sqrtf(u1);
}
return tmp;
}
function code(cosTheta_i, u1, u2) t_0 = Float32(u2 * Float32(Float32(pi) * Float32(2.0))) tmp = Float32(0.0) if (t_0 <= Float32(0.013000000268220901)) tmp = sqrt(Float32(-log1p(Float32(-u1)))); else tmp = Float32(cos(t_0) * sqrt(u1)); end return tmp end
\begin{array}{l}
\\
\begin{array}{l}
t_0 := u2 \cdot \left(\pi \cdot 2\right)\\
\mathbf{if}\;t\_0 \leq 0.013000000268220901:\\
\;\;\;\;\sqrt{-\mathsf{log1p}\left(-u1\right)}\\
\mathbf{else}:\\
\;\;\;\;\cos t\_0 \cdot \sqrt{u1}\\
\end{array}
\end{array}
if (*.f32 (*.f32 #s(literal 2 binary32) (PI.f32)) u2) < 0.0130000003Initial program 60.0%
*-lowering-*.f32N/A
sqrt-lowering-sqrt.f32N/A
neg-lowering-neg.f32N/A
sub-negN/A
log1p-defineN/A
log1p-lowering-log1p.f32N/A
neg-lowering-neg.f32N/A
cos-lowering-cos.f32N/A
associate-*l*N/A
*-lowering-*.f32N/A
*-lowering-*.f32N/A
PI-lowering-PI.f3299.5%
Simplified99.5%
Taylor expanded in u2 around 0
Simplified96.6%
*-rgt-identityN/A
sqrt-lowering-sqrt.f32N/A
neg-lowering-neg.f32N/A
log1p-defineN/A
log1p-lowering-log1p.f32N/A
neg-lowering-neg.f3296.6%
Applied egg-rr96.6%
if 0.0130000003 < (*.f32 (*.f32 #s(literal 2 binary32) (PI.f32)) u2) Initial program 47.2%
Taylor expanded in u1 around 0
Simplified80.6%
Final simplification92.0%
(FPCore (cosTheta_i u1 u2)
:precision binary32
(if (<= (* u2 (* PI 2.0)) 0.0006399999838322401)
(sqrt (- (log1p (- u1))))
(*
(sqrt
(+ (+ 0.25 (/ 0.3333333333333333 u1)) (/ (+ 0.5 (/ 1.0 u1)) (* u1 u1))))
(+ (* u1 u1) (* (* -2.0 (* u1 u1)) (* u2 (* u2 (* PI PI))))))))
float code(float cosTheta_i, float u1, float u2) {
float tmp;
if ((u2 * (((float) M_PI) * 2.0f)) <= 0.0006399999838322401f) {
tmp = sqrtf(-log1pf(-u1));
} else {
tmp = sqrtf(((0.25f + (0.3333333333333333f / u1)) + ((0.5f + (1.0f / u1)) / (u1 * u1)))) * ((u1 * u1) + ((-2.0f * (u1 * u1)) * (u2 * (u2 * (((float) M_PI) * ((float) M_PI))))));
}
return tmp;
}
function code(cosTheta_i, u1, u2) tmp = Float32(0.0) if (Float32(u2 * Float32(Float32(pi) * Float32(2.0))) <= Float32(0.0006399999838322401)) tmp = sqrt(Float32(-log1p(Float32(-u1)))); else tmp = Float32(sqrt(Float32(Float32(Float32(0.25) + Float32(Float32(0.3333333333333333) / u1)) + Float32(Float32(Float32(0.5) + Float32(Float32(1.0) / u1)) / Float32(u1 * u1)))) * Float32(Float32(u1 * u1) + Float32(Float32(Float32(-2.0) * Float32(u1 * u1)) * Float32(u2 * Float32(u2 * Float32(Float32(pi) * Float32(pi))))))); end return tmp end
\begin{array}{l}
\\
\begin{array}{l}
\mathbf{if}\;u2 \cdot \left(\pi \cdot 2\right) \leq 0.0006399999838322401:\\
\;\;\;\;\sqrt{-\mathsf{log1p}\left(-u1\right)}\\
\mathbf{else}:\\
\;\;\;\;\sqrt{\left(0.25 + \frac{0.3333333333333333}{u1}\right) + \frac{0.5 + \frac{1}{u1}}{u1 \cdot u1}} \cdot \left(u1 \cdot u1 + \left(-2 \cdot \left(u1 \cdot u1\right)\right) \cdot \left(u2 \cdot \left(u2 \cdot \left(\pi \cdot \pi\right)\right)\right)\right)\\
\end{array}
\end{array}
if (*.f32 (*.f32 #s(literal 2 binary32) (PI.f32)) u2) < 6.39999984e-4Initial program 59.5%
*-lowering-*.f32N/A
sqrt-lowering-sqrt.f32N/A
neg-lowering-neg.f32N/A
sub-negN/A
log1p-defineN/A
log1p-lowering-log1p.f32N/A
neg-lowering-neg.f32N/A
cos-lowering-cos.f32N/A
associate-*l*N/A
*-lowering-*.f32N/A
*-lowering-*.f32N/A
PI-lowering-PI.f3299.6%
Simplified99.6%
Taylor expanded in u2 around 0
Simplified99.4%
*-rgt-identityN/A
sqrt-lowering-sqrt.f32N/A
neg-lowering-neg.f32N/A
log1p-defineN/A
log1p-lowering-log1p.f32N/A
neg-lowering-neg.f3299.4%
Applied egg-rr99.4%
if 6.39999984e-4 < (*.f32 (*.f32 #s(literal 2 binary32) (PI.f32)) u2) Initial program 51.9%
Taylor expanded in u1 around 0
*-lowering-*.f32N/A
+-lowering-+.f32N/A
*-lowering-*.f32N/A
+-lowering-+.f32N/A
*-lowering-*.f32N/A
+-lowering-+.f32N/A
*-commutativeN/A
*-lowering-*.f3294.4%
Simplified94.4%
Taylor expanded in u1 around -inf
*-lowering-*.f32N/A
pow-lowering-pow.f32N/A
mul-1-negN/A
unsub-negN/A
--lowering--.f32N/A
/-lowering-/.f32N/A
Simplified94.2%
Taylor expanded in u2 around 0
+-commutativeN/A
associate-*r*N/A
distribute-rgt-outN/A
*-lowering-*.f32N/A
Simplified69.1%
Final simplification86.6%
(FPCore (cosTheta_i u1 u2) :precision binary32 (* (sqrt (+ (+ 0.25 (/ 0.3333333333333333 u1)) (/ (+ 0.5 (/ 1.0 u1)) (* u1 u1)))) (+ (* u1 u1) (* (* -2.0 (* u1 u1)) (* u2 (* u2 (* PI PI)))))))
float code(float cosTheta_i, float u1, float u2) {
return sqrtf(((0.25f + (0.3333333333333333f / u1)) + ((0.5f + (1.0f / u1)) / (u1 * u1)))) * ((u1 * u1) + ((-2.0f * (u1 * u1)) * (u2 * (u2 * (((float) M_PI) * ((float) M_PI))))));
}
function code(cosTheta_i, u1, u2) return Float32(sqrt(Float32(Float32(Float32(0.25) + Float32(Float32(0.3333333333333333) / u1)) + Float32(Float32(Float32(0.5) + Float32(Float32(1.0) / u1)) / Float32(u1 * u1)))) * Float32(Float32(u1 * u1) + Float32(Float32(Float32(-2.0) * Float32(u1 * u1)) * Float32(u2 * Float32(u2 * Float32(Float32(pi) * Float32(pi))))))) end
function tmp = code(cosTheta_i, u1, u2) tmp = sqrt(((single(0.25) + (single(0.3333333333333333) / u1)) + ((single(0.5) + (single(1.0) / u1)) / (u1 * u1)))) * ((u1 * u1) + ((single(-2.0) * (u1 * u1)) * (u2 * (u2 * (single(pi) * single(pi)))))); end
\begin{array}{l}
\\
\sqrt{\left(0.25 + \frac{0.3333333333333333}{u1}\right) + \frac{0.5 + \frac{1}{u1}}{u1 \cdot u1}} \cdot \left(u1 \cdot u1 + \left(-2 \cdot \left(u1 \cdot u1\right)\right) \cdot \left(u2 \cdot \left(u2 \cdot \left(\pi \cdot \pi\right)\right)\right)\right)
\end{array}
Initial program 56.3%
Taylor expanded in u1 around 0
*-lowering-*.f32N/A
+-lowering-+.f32N/A
*-lowering-*.f32N/A
+-lowering-+.f32N/A
*-lowering-*.f32N/A
+-lowering-+.f32N/A
*-commutativeN/A
*-lowering-*.f3293.5%
Simplified93.5%
Taylor expanded in u1 around -inf
*-lowering-*.f32N/A
pow-lowering-pow.f32N/A
mul-1-negN/A
unsub-negN/A
--lowering--.f32N/A
/-lowering-/.f32N/A
Simplified93.3%
Taylor expanded in u2 around 0
+-commutativeN/A
associate-*r*N/A
distribute-rgt-outN/A
*-lowering-*.f32N/A
Simplified82.6%
(FPCore (cosTheta_i u1 u2) :precision binary32 (sqrt (* u1 (+ 1.0 (* u1 (+ 0.5 (* u1 (+ 0.3333333333333333 (* u1 0.25)))))))))
float code(float cosTheta_i, float u1, float u2) {
return sqrtf((u1 * (1.0f + (u1 * (0.5f + (u1 * (0.3333333333333333f + (u1 * 0.25f))))))));
}
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 * (0.5e0 + (u1 * (0.3333333333333333e0 + (u1 * 0.25e0))))))))
end function
function code(cosTheta_i, u1, u2) return sqrt(Float32(u1 * Float32(Float32(1.0) + Float32(u1 * Float32(Float32(0.5) + Float32(u1 * Float32(Float32(0.3333333333333333) + Float32(u1 * Float32(0.25))))))))) end
function tmp = code(cosTheta_i, u1, u2) tmp = sqrt((u1 * (single(1.0) + (u1 * (single(0.5) + (u1 * (single(0.3333333333333333) + (u1 * single(0.25))))))))); end
\begin{array}{l}
\\
\sqrt{u1 \cdot \left(1 + u1 \cdot \left(0.5 + u1 \cdot \left(0.3333333333333333 + u1 \cdot 0.25\right)\right)\right)}
\end{array}
Initial program 56.3%
Taylor expanded in u1 around 0
*-lowering-*.f32N/A
+-lowering-+.f32N/A
*-lowering-*.f32N/A
+-lowering-+.f32N/A
*-lowering-*.f32N/A
+-lowering-+.f32N/A
*-commutativeN/A
*-lowering-*.f3293.5%
Simplified93.5%
Taylor expanded in u2 around 0
sqrt-lowering-sqrt.f32N/A
*-lowering-*.f32N/A
+-lowering-+.f32N/A
*-lowering-*.f32N/A
+-lowering-+.f32N/A
*-lowering-*.f32N/A
+-lowering-+.f32N/A
*-commutativeN/A
*-lowering-*.f3276.1%
Simplified76.1%
(FPCore (cosTheta_i u1 u2) :precision binary32 (sqrt (* u1 (+ 1.0 (* u1 (+ 0.5 (* u1 0.3333333333333333)))))))
float code(float cosTheta_i, float u1, float u2) {
return sqrtf((u1 * (1.0f + (u1 * (0.5f + (u1 * 0.3333333333333333f))))));
}
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 * (0.5e0 + (u1 * 0.3333333333333333e0))))))
end function
function code(cosTheta_i, u1, u2) return sqrt(Float32(u1 * Float32(Float32(1.0) + Float32(u1 * Float32(Float32(0.5) + Float32(u1 * Float32(0.3333333333333333))))))) end
function tmp = code(cosTheta_i, u1, u2) tmp = sqrt((u1 * (single(1.0) + (u1 * (single(0.5) + (u1 * single(0.3333333333333333))))))); end
\begin{array}{l}
\\
\sqrt{u1 \cdot \left(1 + u1 \cdot \left(0.5 + u1 \cdot 0.3333333333333333\right)\right)}
\end{array}
Initial program 56.3%
*-lowering-*.f32N/A
sqrt-lowering-sqrt.f32N/A
neg-lowering-neg.f32N/A
sub-negN/A
log1p-defineN/A
log1p-lowering-log1p.f32N/A
neg-lowering-neg.f32N/A
cos-lowering-cos.f32N/A
associate-*l*N/A
*-lowering-*.f32N/A
*-lowering-*.f32N/A
PI-lowering-PI.f3299.0%
Simplified99.0%
Taylor expanded in u2 around 0
Simplified80.3%
Taylor expanded in u1 around 0
*-lowering-*.f32N/A
+-lowering-+.f32N/A
*-lowering-*.f32N/A
+-lowering-+.f32N/A
*-lowering-*.f3274.8%
Simplified74.8%
Final simplification74.8%
(FPCore (cosTheta_i u1 u2) :precision binary32 (sqrt (* u1 (+ 1.0 (* u1 0.5)))))
float code(float cosTheta_i, float u1, float u2) {
return sqrtf((u1 * (1.0f + (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 = sqrt((u1 * (1.0e0 + (u1 * 0.5e0))))
end function
function code(cosTheta_i, u1, u2) return sqrt(Float32(u1 * Float32(Float32(1.0) + Float32(u1 * Float32(0.5))))) end
function tmp = code(cosTheta_i, u1, u2) tmp = sqrt((u1 * (single(1.0) + (u1 * single(0.5))))); end
\begin{array}{l}
\\
\sqrt{u1 \cdot \left(1 + u1 \cdot 0.5\right)}
\end{array}
Initial program 56.3%
*-lowering-*.f32N/A
sqrt-lowering-sqrt.f32N/A
neg-lowering-neg.f32N/A
sub-negN/A
log1p-defineN/A
log1p-lowering-log1p.f32N/A
neg-lowering-neg.f32N/A
cos-lowering-cos.f32N/A
associate-*l*N/A
*-lowering-*.f32N/A
*-lowering-*.f32N/A
PI-lowering-PI.f3299.0%
Simplified99.0%
Taylor expanded in u2 around 0
Simplified80.3%
Taylor expanded in u1 around 0
*-lowering-*.f32N/A
+-lowering-+.f32N/A
*-lowering-*.f3272.2%
Simplified72.2%
Final simplification72.2%
(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 56.3%
*-lowering-*.f32N/A
sqrt-lowering-sqrt.f32N/A
neg-lowering-neg.f32N/A
sub-negN/A
log1p-defineN/A
log1p-lowering-log1p.f32N/A
neg-lowering-neg.f32N/A
cos-lowering-cos.f32N/A
associate-*l*N/A
*-lowering-*.f32N/A
*-lowering-*.f32N/A
PI-lowering-PI.f3299.0%
Simplified99.0%
Taylor expanded in u2 around 0
Simplified80.3%
flip-+N/A
clear-numN/A
log-recN/A
neg-lowering-neg.f32N/A
log-lowering-log.f32N/A
/-lowering-/.f32N/A
neg-sub0N/A
associate--r-N/A
metadata-evalN/A
+-lowering-+.f32N/A
metadata-evalN/A
--lowering--.f32N/A
sqr-negN/A
*-lowering-*.f3247.8%
Applied egg-rr47.8%
Taylor expanded in u1 around 0
sqrt-lowering-sqrt.f3264.3%
Simplified64.3%
herbie shell --seed 2024288
(FPCore (cosTheta_i u1 u2)
:name "Beckmann 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 (- (log (- 1.0 u1)))) (cos (* (* 2.0 PI) u2))))