
(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 (* (sqrt (- (log1p (- u1)))) (cos (* (* 2.0 PI) u2))))
float code(float cosTheta_i, float u1, float u2) {
return sqrtf(-log1pf(-u1)) * cosf(((2.0f * ((float) M_PI)) * u2));
}
function code(cosTheta_i, u1, u2) return Float32(sqrt(Float32(-log1p(Float32(-u1)))) * cos(Float32(Float32(Float32(2.0) * Float32(pi)) * u2))) end
\begin{array}{l}
\\
\sqrt{-\mathsf{log1p}\left(-u1\right)} \cdot \cos \left(\left(2 \cdot \pi\right) \cdot u2\right)
\end{array}
Initial program 54.7%
sub-negN/A
accelerator-lowering-log1p.f32N/A
neg-lowering-neg.f3299.2%
Applied egg-rr99.2%
(FPCore (cosTheta_i u1 u2)
:precision binary32
(let* ((t_0 (cos (* (* 2.0 PI) u2))))
(if (<= t_0 0.9990000128746033)
(* t_0 (sqrt (* u1 (+ 1.0 (* u1 0.5)))))
(* (sqrt (- (log1p (- u1)))) (+ 1.0 (* u2 (* u2 (* -2.0 (* PI PI)))))))))
float code(float cosTheta_i, float u1, float u2) {
float t_0 = cosf(((2.0f * ((float) M_PI)) * u2));
float tmp;
if (t_0 <= 0.9990000128746033f) {
tmp = t_0 * sqrtf((u1 * (1.0f + (u1 * 0.5f))));
} else {
tmp = sqrtf(-log1pf(-u1)) * (1.0f + (u2 * (u2 * (-2.0f * (((float) M_PI) * ((float) M_PI))))));
}
return tmp;
}
function code(cosTheta_i, u1, u2) t_0 = cos(Float32(Float32(Float32(2.0) * Float32(pi)) * u2)) tmp = Float32(0.0) if (t_0 <= Float32(0.9990000128746033)) tmp = Float32(t_0 * sqrt(Float32(u1 * Float32(Float32(1.0) + Float32(u1 * Float32(0.5)))))); else tmp = Float32(sqrt(Float32(-log1p(Float32(-u1)))) * Float32(Float32(1.0) + Float32(u2 * Float32(u2 * Float32(Float32(-2.0) * Float32(Float32(pi) * Float32(pi))))))); end return tmp end
\begin{array}{l}
\\
\begin{array}{l}
t_0 := \cos \left(\left(2 \cdot \pi\right) \cdot u2\right)\\
\mathbf{if}\;t\_0 \leq 0.9990000128746033:\\
\;\;\;\;t\_0 \cdot \sqrt{u1 \cdot \left(1 + u1 \cdot 0.5\right)}\\
\mathbf{else}:\\
\;\;\;\;\sqrt{-\mathsf{log1p}\left(-u1\right)} \cdot \left(1 + u2 \cdot \left(u2 \cdot \left(-2 \cdot \left(\pi \cdot \pi\right)\right)\right)\right)\\
\end{array}
\end{array}
if (cos.f32 (*.f32 (*.f32 #s(literal 2 binary32) (PI.f32)) u2)) < 0.999000013Initial program 49.7%
Taylor expanded in u1 around 0
*-lowering-*.f32N/A
+-lowering-+.f32N/A
*-commutativeN/A
*-lowering-*.f3290.0%
Simplified90.0%
if 0.999000013 < (cos.f32 (*.f32 (*.f32 #s(literal 2 binary32) (PI.f32)) u2)) Initial program 56.5%
sub-negN/A
accelerator-lowering-log1p.f32N/A
neg-lowering-neg.f3299.5%
Applied egg-rr99.5%
Taylor expanded in u2 around 0
*-commutativeN/A
associate-*r*N/A
+-lowering-+.f32N/A
associate-*r*N/A
*-commutativeN/A
*-commutativeN/A
associate-*r*N/A
*-commutativeN/A
unpow2N/A
associate-*l*N/A
*-lowering-*.f32N/A
*-lowering-*.f32N/A
*-lowering-*.f32N/A
unpow2N/A
*-lowering-*.f32N/A
PI-lowering-PI.f32N/A
PI-lowering-PI.f3299.4%
Simplified99.4%
Final simplification97.0%
(FPCore (cosTheta_i u1 u2)
:precision binary32
(let* ((t_0 (cos (* (* 2.0 PI) u2))))
(if (<= t_0 0.9999995231628418)
(* t_0 (sqrt (* u1 (+ 1.0 (* u1 0.5)))))
(sqrt (- (log1p (- u1)))))))
float code(float cosTheta_i, float u1, float u2) {
float t_0 = cosf(((2.0f * ((float) M_PI)) * u2));
float tmp;
if (t_0 <= 0.9999995231628418f) {
tmp = t_0 * sqrtf((u1 * (1.0f + (u1 * 0.5f))));
} else {
tmp = sqrtf(-log1pf(-u1));
}
return tmp;
}
function code(cosTheta_i, u1, u2) t_0 = cos(Float32(Float32(Float32(2.0) * Float32(pi)) * u2)) tmp = Float32(0.0) if (t_0 <= Float32(0.9999995231628418)) tmp = Float32(t_0 * sqrt(Float32(u1 * Float32(Float32(1.0) + Float32(u1 * Float32(0.5)))))); else tmp = sqrt(Float32(-log1p(Float32(-u1)))); end return tmp end
\begin{array}{l}
\\
\begin{array}{l}
t_0 := \cos \left(\left(2 \cdot \pi\right) \cdot u2\right)\\
\mathbf{if}\;t\_0 \leq 0.9999995231628418:\\
\;\;\;\;t\_0 \cdot \sqrt{u1 \cdot \left(1 + u1 \cdot 0.5\right)}\\
\mathbf{else}:\\
\;\;\;\;\sqrt{-\mathsf{log1p}\left(-u1\right)}\\
\end{array}
\end{array}
if (cos.f32 (*.f32 (*.f32 #s(literal 2 binary32) (PI.f32)) u2)) < 0.999999523Initial program 51.0%
Taylor expanded in u1 around 0
*-lowering-*.f32N/A
+-lowering-+.f32N/A
*-commutativeN/A
*-lowering-*.f3289.9%
Simplified89.9%
if 0.999999523 < (cos.f32 (*.f32 (*.f32 #s(literal 2 binary32) (PI.f32)) u2)) Initial program 57.4%
sub-negN/A
accelerator-lowering-log1p.f32N/A
neg-lowering-neg.f3299.6%
Applied egg-rr99.6%
Taylor expanded in u2 around 0
*-commutativeN/A
associate-*r*N/A
+-lowering-+.f32N/A
associate-*r*N/A
*-commutativeN/A
*-commutativeN/A
associate-*r*N/A
*-commutativeN/A
unpow2N/A
associate-*l*N/A
*-lowering-*.f32N/A
*-lowering-*.f32N/A
*-lowering-*.f32N/A
unpow2N/A
*-lowering-*.f32N/A
PI-lowering-PI.f32N/A
PI-lowering-PI.f3299.6%
Simplified99.6%
sqrt-lowering-sqrt.f32N/A
neg-logN/A
log-lowering-log.f32N/A
/-lowering-/.f32N/A
unsub-negN/A
--lowering--.f3255.3%
Applied egg-rr55.3%
Taylor expanded in u2 around 0
sqrt-lowering-sqrt.f32N/A
log-recN/A
neg-lowering-neg.f32N/A
sub-negN/A
accelerator-lowering-log1p.f32N/A
neg-lowering-neg.f3299.5%
Simplified99.5%
Final simplification95.5%
(FPCore (cosTheta_i u1 u2)
:precision binary32
(let* ((t_0 (cos (* (* 2.0 PI) u2))))
(if (<= t_0 0.9999961853027344)
(* t_0 (sqrt u1))
(sqrt (- (log1p (- u1)))))))
float code(float cosTheta_i, float u1, float u2) {
float t_0 = cosf(((2.0f * ((float) M_PI)) * u2));
float tmp;
if (t_0 <= 0.9999961853027344f) {
tmp = t_0 * sqrtf(u1);
} else {
tmp = sqrtf(-log1pf(-u1));
}
return tmp;
}
function code(cosTheta_i, u1, u2) t_0 = cos(Float32(Float32(Float32(2.0) * Float32(pi)) * u2)) tmp = Float32(0.0) if (t_0 <= Float32(0.9999961853027344)) tmp = Float32(t_0 * sqrt(u1)); else tmp = sqrt(Float32(-log1p(Float32(-u1)))); end return tmp end
\begin{array}{l}
\\
\begin{array}{l}
t_0 := \cos \left(\left(2 \cdot \pi\right) \cdot u2\right)\\
\mathbf{if}\;t\_0 \leq 0.9999961853027344:\\
\;\;\;\;t\_0 \cdot \sqrt{u1}\\
\mathbf{else}:\\
\;\;\;\;\sqrt{-\mathsf{log1p}\left(-u1\right)}\\
\end{array}
\end{array}
if (cos.f32 (*.f32 (*.f32 #s(literal 2 binary32) (PI.f32)) u2)) < 0.99999619Initial program 50.5%
Taylor expanded in u1 around 0
Simplified80.0%
if 0.99999619 < (cos.f32 (*.f32 (*.f32 #s(literal 2 binary32) (PI.f32)) u2)) Initial program 57.2%
sub-negN/A
accelerator-lowering-log1p.f32N/A
neg-lowering-neg.f3299.5%
Applied egg-rr99.5%
Taylor expanded in u2 around 0
*-commutativeN/A
associate-*r*N/A
+-lowering-+.f32N/A
associate-*r*N/A
*-commutativeN/A
*-commutativeN/A
associate-*r*N/A
*-commutativeN/A
unpow2N/A
associate-*l*N/A
*-lowering-*.f32N/A
*-lowering-*.f32N/A
*-lowering-*.f32N/A
unpow2N/A
*-lowering-*.f32N/A
PI-lowering-PI.f32N/A
PI-lowering-PI.f3299.5%
Simplified99.5%
sqrt-lowering-sqrt.f32N/A
neg-logN/A
log-lowering-log.f32N/A
/-lowering-/.f32N/A
unsub-negN/A
--lowering--.f3255.1%
Applied egg-rr55.1%
Taylor expanded in u2 around 0
sqrt-lowering-sqrt.f32N/A
log-recN/A
neg-lowering-neg.f32N/A
sub-negN/A
accelerator-lowering-log1p.f32N/A
neg-lowering-neg.f3298.5%
Simplified98.5%
Final simplification91.7%
(FPCore (cosTheta_i u1 u2)
:precision binary32
(let* ((t_0 (* (* 2.0 PI) u2)))
(if (<= t_0 0.03500000014901161)
(* (sqrt (- (log1p (- u1)))) (+ 1.0 (* u2 (* u2 (* -2.0 (* PI PI))))))
(*
(cos t_0)
(sqrt
(*
u1
(+ 1.0 (* u1 (+ 0.5 (* u1 (+ 0.3333333333333333 (* u1 0.25))))))))))))
float code(float cosTheta_i, float u1, float u2) {
float t_0 = (2.0f * ((float) M_PI)) * u2;
float tmp;
if (t_0 <= 0.03500000014901161f) {
tmp = sqrtf(-log1pf(-u1)) * (1.0f + (u2 * (u2 * (-2.0f * (((float) M_PI) * ((float) M_PI))))));
} else {
tmp = cosf(t_0) * sqrtf((u1 * (1.0f + (u1 * (0.5f + (u1 * (0.3333333333333333f + (u1 * 0.25f))))))));
}
return tmp;
}
function code(cosTheta_i, u1, u2) t_0 = Float32(Float32(Float32(2.0) * Float32(pi)) * u2) tmp = Float32(0.0) if (t_0 <= Float32(0.03500000014901161)) tmp = Float32(sqrt(Float32(-log1p(Float32(-u1)))) * Float32(Float32(1.0) + Float32(u2 * Float32(u2 * Float32(Float32(-2.0) * Float32(Float32(pi) * Float32(pi))))))); else tmp = Float32(cos(t_0) * 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 return tmp end
\begin{array}{l}
\\
\begin{array}{l}
t_0 := \left(2 \cdot \pi\right) \cdot u2\\
\mathbf{if}\;t\_0 \leq 0.03500000014901161:\\
\;\;\;\;\sqrt{-\mathsf{log1p}\left(-u1\right)} \cdot \left(1 + u2 \cdot \left(u2 \cdot \left(-2 \cdot \left(\pi \cdot \pi\right)\right)\right)\right)\\
\mathbf{else}:\\
\;\;\;\;\cos t\_0 \cdot \sqrt{u1 \cdot \left(1 + u1 \cdot \left(0.5 + u1 \cdot \left(0.3333333333333333 + u1 \cdot 0.25\right)\right)\right)}\\
\end{array}
\end{array}
if (*.f32 (*.f32 #s(literal 2 binary32) (PI.f32)) u2) < 0.0350000001Initial program 56.2%
sub-negN/A
accelerator-lowering-log1p.f32N/A
neg-lowering-neg.f3299.5%
Applied egg-rr99.5%
Taylor expanded in u2 around 0
*-commutativeN/A
associate-*r*N/A
+-lowering-+.f32N/A
associate-*r*N/A
*-commutativeN/A
*-commutativeN/A
associate-*r*N/A
*-commutativeN/A
unpow2N/A
associate-*l*N/A
*-lowering-*.f32N/A
*-lowering-*.f32N/A
*-lowering-*.f32N/A
unpow2N/A
*-lowering-*.f32N/A
PI-lowering-PI.f32N/A
PI-lowering-PI.f3299.5%
Simplified99.5%
if 0.0350000001 < (*.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
*-lowering-*.f32N/A
+-lowering-+.f32N/A
*-lowering-*.f32N/A
+-lowering-+.f32N/A
*-commutativeN/A
*-lowering-*.f3296.0%
Simplified96.0%
Final simplification98.5%
(FPCore (cosTheta_i u1 u2)
:precision binary32
(let* ((t_0 (* (* 2.0 PI) u2)))
(if (<= t_0 0.03500000014901161)
(* (sqrt (- (log1p (- u1)))) (+ 1.0 (* u2 (* u2 (* -2.0 (* PI PI))))))
(*
(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 = (2.0f * ((float) M_PI)) * u2;
float tmp;
if (t_0 <= 0.03500000014901161f) {
tmp = sqrtf(-log1pf(-u1)) * (1.0f + (u2 * (u2 * (-2.0f * (((float) M_PI) * ((float) M_PI))))));
} 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(Float32(Float32(2.0) * Float32(pi)) * u2) tmp = Float32(0.0) if (t_0 <= Float32(0.03500000014901161)) tmp = Float32(sqrt(Float32(-log1p(Float32(-u1)))) * Float32(Float32(1.0) + Float32(u2 * Float32(u2 * Float32(Float32(-2.0) * Float32(Float32(pi) * Float32(pi))))))); 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 := \left(2 \cdot \pi\right) \cdot u2\\
\mathbf{if}\;t\_0 \leq 0.03500000014901161:\\
\;\;\;\;\sqrt{-\mathsf{log1p}\left(-u1\right)} \cdot \left(1 + u2 \cdot \left(u2 \cdot \left(-2 \cdot \left(\pi \cdot \pi\right)\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) < 0.0350000001Initial program 56.2%
sub-negN/A
accelerator-lowering-log1p.f32N/A
neg-lowering-neg.f3299.5%
Applied egg-rr99.5%
Taylor expanded in u2 around 0
*-commutativeN/A
associate-*r*N/A
+-lowering-+.f32N/A
associate-*r*N/A
*-commutativeN/A
*-commutativeN/A
associate-*r*N/A
*-commutativeN/A
unpow2N/A
associate-*l*N/A
*-lowering-*.f32N/A
*-lowering-*.f32N/A
*-lowering-*.f32N/A
unpow2N/A
*-lowering-*.f32N/A
PI-lowering-PI.f32N/A
PI-lowering-PI.f3299.5%
Simplified99.5%
if 0.0350000001 < (*.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
*-lowering-*.f32N/A
+-lowering-+.f32N/A
*-commutativeN/A
*-lowering-*.f3294.1%
Simplified94.1%
Final simplification98.0%
(FPCore (cosTheta_i u1 u2)
:precision binary32
(if (<= (* (* 2.0 PI) u2) 0.0010000000474974513)
(sqrt (- (log1p (- u1))))
(*
(+ 1.0 (* u2 (* u2 (* -2.0 (* PI PI)))))
(sqrt
(*
u1
(+ 1.0 (* u1 (+ 0.5 (* u1 (+ 0.3333333333333333 (* u1 0.25)))))))))))
float code(float cosTheta_i, float u1, float u2) {
float tmp;
if (((2.0f * ((float) M_PI)) * u2) <= 0.0010000000474974513f) {
tmp = sqrtf(-log1pf(-u1));
} else {
tmp = (1.0f + (u2 * (u2 * (-2.0f * (((float) M_PI) * ((float) M_PI)))))) * sqrtf((u1 * (1.0f + (u1 * (0.5f + (u1 * (0.3333333333333333f + (u1 * 0.25f))))))));
}
return tmp;
}
function code(cosTheta_i, u1, u2) tmp = Float32(0.0) if (Float32(Float32(Float32(2.0) * Float32(pi)) * u2) <= Float32(0.0010000000474974513)) tmp = sqrt(Float32(-log1p(Float32(-u1)))); else tmp = Float32(Float32(Float32(1.0) + Float32(u2 * Float32(u2 * Float32(Float32(-2.0) * Float32(Float32(pi) * Float32(pi)))))) * 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 return tmp end
\begin{array}{l}
\\
\begin{array}{l}
\mathbf{if}\;\left(2 \cdot \pi\right) \cdot u2 \leq 0.0010000000474974513:\\
\;\;\;\;\sqrt{-\mathsf{log1p}\left(-u1\right)}\\
\mathbf{else}:\\
\;\;\;\;\left(1 + u2 \cdot \left(u2 \cdot \left(-2 \cdot \left(\pi \cdot \pi\right)\right)\right)\right) \cdot \sqrt{u1 \cdot \left(1 + u1 \cdot \left(0.5 + u1 \cdot \left(0.3333333333333333 + u1 \cdot 0.25\right)\right)\right)}\\
\end{array}
\end{array}
if (*.f32 (*.f32 #s(literal 2 binary32) (PI.f32)) u2) < 0.00100000005Initial program 57.7%
sub-negN/A
accelerator-lowering-log1p.f32N/A
neg-lowering-neg.f3299.6%
Applied egg-rr99.6%
Taylor expanded in u2 around 0
*-commutativeN/A
associate-*r*N/A
+-lowering-+.f32N/A
associate-*r*N/A
*-commutativeN/A
*-commutativeN/A
associate-*r*N/A
*-commutativeN/A
unpow2N/A
associate-*l*N/A
*-lowering-*.f32N/A
*-lowering-*.f32N/A
*-lowering-*.f32N/A
unpow2N/A
*-lowering-*.f32N/A
PI-lowering-PI.f32N/A
PI-lowering-PI.f3299.6%
Simplified99.6%
sqrt-lowering-sqrt.f32N/A
neg-logN/A
log-lowering-log.f32N/A
/-lowering-/.f32N/A
unsub-negN/A
--lowering--.f3255.6%
Applied egg-rr55.6%
Taylor expanded in u2 around 0
sqrt-lowering-sqrt.f32N/A
log-recN/A
neg-lowering-neg.f32N/A
sub-negN/A
accelerator-lowering-log1p.f32N/A
neg-lowering-neg.f3299.4%
Simplified99.4%
if 0.00100000005 < (*.f32 (*.f32 #s(literal 2 binary32) (PI.f32)) u2) Initial program 50.6%
sub-negN/A
accelerator-lowering-log1p.f32N/A
neg-lowering-neg.f3298.8%
Applied egg-rr98.8%
Taylor expanded in u2 around 0
*-commutativeN/A
associate-*r*N/A
+-lowering-+.f32N/A
associate-*r*N/A
*-commutativeN/A
*-commutativeN/A
associate-*r*N/A
*-commutativeN/A
unpow2N/A
associate-*l*N/A
*-lowering-*.f32N/A
*-lowering-*.f32N/A
*-lowering-*.f32N/A
unpow2N/A
*-lowering-*.f32N/A
PI-lowering-PI.f32N/A
PI-lowering-PI.f3272.2%
Simplified72.2%
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-*.f3271.0%
Simplified71.0%
Final simplification87.6%
(FPCore (cosTheta_i u1 u2) :precision binary32 (* (+ 1.0 (* u2 (* u2 (* -2.0 (* PI PI))))) (sqrt (* u1 (+ 1.0 (* u1 (+ 0.5 (* u1 (+ 0.3333333333333333 (* u1 0.25))))))))))
float code(float cosTheta_i, float u1, float u2) {
return (1.0f + (u2 * (u2 * (-2.0f * (((float) M_PI) * ((float) M_PI)))))) * sqrtf((u1 * (1.0f + (u1 * (0.5f + (u1 * (0.3333333333333333f + (u1 * 0.25f))))))));
}
function code(cosTheta_i, u1, u2) return Float32(Float32(Float32(1.0) + Float32(u2 * Float32(u2 * Float32(Float32(-2.0) * Float32(Float32(pi) * Float32(pi)))))) * 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 = (single(1.0) + (u2 * (u2 * (single(-2.0) * (single(pi) * single(pi)))))) * sqrt((u1 * (single(1.0) + (u1 * (single(0.5) + (u1 * (single(0.3333333333333333) + (u1 * single(0.25))))))))); end
\begin{array}{l}
\\
\left(1 + u2 \cdot \left(u2 \cdot \left(-2 \cdot \left(\pi \cdot \pi\right)\right)\right)\right) \cdot \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 54.7%
sub-negN/A
accelerator-lowering-log1p.f32N/A
neg-lowering-neg.f3299.2%
Applied egg-rr99.2%
Taylor expanded in u2 around 0
*-commutativeN/A
associate-*r*N/A
+-lowering-+.f32N/A
associate-*r*N/A
*-commutativeN/A
*-commutativeN/A
associate-*r*N/A
*-commutativeN/A
unpow2N/A
associate-*l*N/A
*-lowering-*.f32N/A
*-lowering-*.f32N/A
*-lowering-*.f32N/A
unpow2N/A
*-lowering-*.f32N/A
PI-lowering-PI.f32N/A
PI-lowering-PI.f3288.2%
Simplified88.2%
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-*.f3284.4%
Simplified84.4%
Final simplification84.4%
(FPCore (cosTheta_i u1 u2) :precision binary32 (* (+ 1.0 (* u2 (* u2 (* -2.0 (* PI PI))))) (sqrt (* u1 (+ 1.0 (* u1 (+ 0.5 (* u1 0.3333333333333333))))))))
float code(float cosTheta_i, float u1, float u2) {
return (1.0f + (u2 * (u2 * (-2.0f * (((float) M_PI) * ((float) M_PI)))))) * sqrtf((u1 * (1.0f + (u1 * (0.5f + (u1 * 0.3333333333333333f))))));
}
function code(cosTheta_i, u1, u2) return Float32(Float32(Float32(1.0) + Float32(u2 * Float32(u2 * Float32(Float32(-2.0) * Float32(Float32(pi) * Float32(pi)))))) * 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 = (single(1.0) + (u2 * (u2 * (single(-2.0) * (single(pi) * single(pi)))))) * sqrt((u1 * (single(1.0) + (u1 * (single(0.5) + (u1 * single(0.3333333333333333))))))); end
\begin{array}{l}
\\
\left(1 + u2 \cdot \left(u2 \cdot \left(-2 \cdot \left(\pi \cdot \pi\right)\right)\right)\right) \cdot \sqrt{u1 \cdot \left(1 + u1 \cdot \left(0.5 + u1 \cdot 0.3333333333333333\right)\right)}
\end{array}
Initial program 54.7%
sub-negN/A
accelerator-lowering-log1p.f32N/A
neg-lowering-neg.f3299.2%
Applied egg-rr99.2%
Taylor expanded in u2 around 0
*-commutativeN/A
associate-*r*N/A
+-lowering-+.f32N/A
associate-*r*N/A
*-commutativeN/A
*-commutativeN/A
associate-*r*N/A
*-commutativeN/A
unpow2N/A
associate-*l*N/A
*-lowering-*.f32N/A
*-lowering-*.f32N/A
*-lowering-*.f32N/A
unpow2N/A
*-lowering-*.f32N/A
PI-lowering-PI.f32N/A
PI-lowering-PI.f3288.2%
Simplified88.2%
Taylor expanded in u1 around 0
*-lowering-*.f32N/A
+-lowering-+.f32N/A
*-lowering-*.f32N/A
+-lowering-+.f32N/A
*-commutativeN/A
*-lowering-*.f3283.0%
Simplified83.0%
Final simplification83.0%
(FPCore (cosTheta_i u1 u2) :precision binary32 (* (+ 1.0 (* u2 (* u2 (* -2.0 (* PI PI))))) (sqrt (* u1 (+ 1.0 (* u1 0.5))))))
float code(float cosTheta_i, float u1, float u2) {
return (1.0f + (u2 * (u2 * (-2.0f * (((float) M_PI) * ((float) M_PI)))))) * sqrtf((u1 * (1.0f + (u1 * 0.5f))));
}
function code(cosTheta_i, u1, u2) return Float32(Float32(Float32(1.0) + Float32(u2 * Float32(u2 * Float32(Float32(-2.0) * Float32(Float32(pi) * Float32(pi)))))) * sqrt(Float32(u1 * Float32(Float32(1.0) + Float32(u1 * Float32(0.5)))))) end
function tmp = code(cosTheta_i, u1, u2) tmp = (single(1.0) + (u2 * (u2 * (single(-2.0) * (single(pi) * single(pi)))))) * sqrt((u1 * (single(1.0) + (u1 * single(0.5))))); end
\begin{array}{l}
\\
\left(1 + u2 \cdot \left(u2 \cdot \left(-2 \cdot \left(\pi \cdot \pi\right)\right)\right)\right) \cdot \sqrt{u1 \cdot \left(1 + u1 \cdot 0.5\right)}
\end{array}
Initial program 54.7%
sub-negN/A
accelerator-lowering-log1p.f32N/A
neg-lowering-neg.f3299.2%
Applied egg-rr99.2%
Taylor expanded in u2 around 0
*-commutativeN/A
associate-*r*N/A
+-lowering-+.f32N/A
associate-*r*N/A
*-commutativeN/A
*-commutativeN/A
associate-*r*N/A
*-commutativeN/A
unpow2N/A
associate-*l*N/A
*-lowering-*.f32N/A
*-lowering-*.f32N/A
*-lowering-*.f32N/A
unpow2N/A
*-lowering-*.f32N/A
PI-lowering-PI.f32N/A
PI-lowering-PI.f3288.2%
Simplified88.2%
Taylor expanded in u1 around 0
*-lowering-*.f32N/A
+-lowering-+.f32N/A
*-commutativeN/A
*-lowering-*.f3279.9%
Simplified79.9%
Final simplification79.9%
(FPCore (cosTheta_i u1 u2) :precision binary32 (* (sqrt u1) (+ 1.0 (* (* -2.0 (* PI PI)) (* u2 u2)))))
float code(float cosTheta_i, float u1, float u2) {
return sqrtf(u1) * (1.0f + ((-2.0f * (((float) M_PI) * ((float) M_PI))) * (u2 * u2)));
}
function code(cosTheta_i, u1, u2) return Float32(sqrt(u1) * Float32(Float32(1.0) + Float32(Float32(Float32(-2.0) * Float32(Float32(pi) * Float32(pi))) * Float32(u2 * u2)))) end
function tmp = code(cosTheta_i, u1, u2) tmp = sqrt(u1) * (single(1.0) + ((single(-2.0) * (single(pi) * single(pi))) * (u2 * u2))); end
\begin{array}{l}
\\
\sqrt{u1} \cdot \left(1 + \left(-2 \cdot \left(\pi \cdot \pi\right)\right) \cdot \left(u2 \cdot u2\right)\right)
\end{array}
Initial program 54.7%
Taylor expanded in u1 around 0
Simplified78.1%
pow1/2N/A
sqr-powN/A
pow2N/A
pow-lowering-pow.f32N/A
pow-lowering-pow.f32N/A
metadata-eval77.7%
Applied egg-rr77.7%
Taylor expanded in u2 around 0
*-commutativeN/A
associate-*r*N/A
*-lft-identityN/A
distribute-rgt-inN/A
*-lowering-*.f32N/A
sqrt-lowering-sqrt.f32N/A
+-lowering-+.f32N/A
*-commutativeN/A
associate-*r*N/A
*-commutativeN/A
*-lowering-*.f32N/A
unpow2N/A
*-lowering-*.f32N/A
*-lowering-*.f32N/A
unpow2N/A
*-lowering-*.f32N/A
PI-lowering-PI.f32N/A
PI-lowering-PI.f3270.6%
Simplified70.6%
Final simplification70.6%
(FPCore (cosTheta_i u1 u2) :precision binary32 (* (+ 1.0 (* u2 (* u2 (* -2.0 (* PI PI))))) (sqrt u1)))
float code(float cosTheta_i, float u1, float u2) {
return (1.0f + (u2 * (u2 * (-2.0f * (((float) M_PI) * ((float) M_PI)))))) * sqrtf(u1);
}
function code(cosTheta_i, u1, u2) return Float32(Float32(Float32(1.0) + Float32(u2 * Float32(u2 * Float32(Float32(-2.0) * Float32(Float32(pi) * Float32(pi)))))) * sqrt(u1)) end
function tmp = code(cosTheta_i, u1, u2) tmp = (single(1.0) + (u2 * (u2 * (single(-2.0) * (single(pi) * single(pi)))))) * sqrt(u1); end
\begin{array}{l}
\\
\left(1 + u2 \cdot \left(u2 \cdot \left(-2 \cdot \left(\pi \cdot \pi\right)\right)\right)\right) \cdot \sqrt{u1}
\end{array}
Initial program 54.7%
Taylor expanded in u1 around 0
Simplified78.1%
Taylor expanded in u2 around 0
*-commutativeN/A
associate-*r*N/A
+-lowering-+.f32N/A
associate-*r*N/A
*-commutativeN/A
*-commutativeN/A
associate-*r*N/A
*-commutativeN/A
unpow2N/A
associate-*l*N/A
*-lowering-*.f32N/A
*-lowering-*.f32N/A
*-lowering-*.f32N/A
unpow2N/A
*-lowering-*.f32N/A
PI-lowering-PI.f32N/A
PI-lowering-PI.f3270.6%
Simplified70.6%
Final simplification70.6%
(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 54.7%
Taylor expanded in u1 around 0
Simplified78.1%
Taylor expanded in u2 around 0
sqrt-lowering-sqrt.f3264.5%
Simplified64.5%
herbie shell --seed 2024191
(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))))