Disney BSSRDF, sample scattering profile, lower

Percentage Accurate: 61.1% → 99.4%

Time: 8.6s

Alternatives: 10

Speedup: 21.8×

Specification

\[\left(0 \leq s \land s \leq 256\right) \land \left(2.328306437 \cdot 10^{-10} \leq u \land u \leq 0.25\right)\]

Math

\[\begin{array}{l} \\ s \cdot \log \left(\frac{1}{1 - 4 \cdot u}\right) \end{array} \]

FPCore

(FPCore (s u) :precision binary32 (* s (log (/ 1.0 (- 1.0 (* 4.0 u))))))

C

float code(float s, float u) {
	return s * logf((1.0f / (1.0f - (4.0f * u))));
}

Fortran

real(4) function code(s, u)
    real(4), intent (in) :: s
    real(4), intent (in) :: u
    code = s * log((1.0e0 / (1.0e0 - (4.0e0 * u))))
end function

Julia

function code(s, u)
	return Float32(s * log(Float32(Float32(1.0) / Float32(Float32(1.0) - Float32(Float32(4.0) * u)))))
end

MATLAB

function tmp = code(s, u)
	tmp = s * log((single(1.0) / (single(1.0) - (single(4.0) * u))));
end

Initial Program: 61.1% accurate, 1.0× speedup

Math

\[\begin{array}{l} \\ s \cdot \log \left(\frac{1}{1 - 4 \cdot u}\right) \end{array} \]

FPCore

(FPCore (s u) :precision binary32 (* s (log (/ 1.0 (- 1.0 (* 4.0 u))))))

C

float code(float s, float u) {
	return s * logf((1.0f / (1.0f - (4.0f * u))));
}

Fortran

real(4) function code(s, u)
    real(4), intent (in) :: s
    real(4), intent (in) :: u
    code = s * log((1.0e0 / (1.0e0 - (4.0e0 * u))))
end function

Julia

function code(s, u)
	return Float32(s * log(Float32(Float32(1.0) / Float32(Float32(1.0) - Float32(Float32(4.0) * u)))))
end

MATLAB

function tmp = code(s, u)
	tmp = s * log((single(1.0) / (single(1.0) - (single(4.0) * u))));
end

Alternative 1: 99.4% accurate, 1.0× speedup

Math

\[\begin{array}{l} \\ \left(-s\right) \cdot \mathsf{log1p}\left(u \cdot -4\right) \end{array} \]

FPCore

(FPCore (s u) :precision binary32 (* (- s) (log1p (* u -4.0))))

C

float code(float s, float u) {
	return -s * log1pf((u * -4.0f));
}

Julia

function code(s, u)
	return Float32(Float32(-s) * log1p(Float32(u * Float32(-4.0))))
end

Derivation

1. Initial program 60.2%

   \[s \cdot \log \left(\frac{1}{1 - 4 \cdot u}\right) \]
2. Step-by-step derivation

   1. log-rec 62.6%

      \[\leadsto s \cdot \color{blue}{\left(-\log \left(1 - 4 \cdot u\right)\right)} \]

   2. distribute-rgt-neg-out 62.6%

      \[\leadsto \color{blue}{-s \cdot \log \left(1 - 4 \cdot u\right)} \]

   3. distribute-lft-neg-out 62.6%

      \[\leadsto \color{blue}{\left(-s\right) \cdot \log \left(1 - 4 \cdot u\right)} \]

   4. cancel-sign-sub-inv 62.6%

      \[\leadsto \left(-s\right) \cdot \log \color{blue}{\left(1 + \left(-4\right) \cdot u\right)} \]

   5. log1p-define 99.4%

      \[\leadsto \left(-s\right) \cdot \color{blue}{\mathsf{log1p}\left(\left(-4\right) \cdot u\right)} \]

   6. *-commutative 99.4%

      \[\leadsto \left(-s\right) \cdot \mathsf{log1p}\left(\color{blue}{u \cdot \left(-4\right)}\right) \]

   7. metadata-eval 99.4%

      \[\leadsto \left(-s\right) \cdot \mathsf{log1p}\left(u \cdot \color{blue}{-4}\right) \]

3. Simplified 99.4%

   \[\leadsto \color{blue}{\left(-s\right) \cdot \mathsf{log1p}\left(u \cdot -4\right)} \]
4. Add Preprocessing
5. Add Preprocessing

Alternative 2: 93.2% accurate, 4.7× speedup

Math

\[\begin{array}{l} \\ u \cdot \left(s \cdot 4 + u \cdot \left(s \cdot 8 + u \cdot \left(s \cdot 21.333333333333332 + 64 \cdot \left(s \cdot u\right)\right)\right)\right) \end{array} \]

FPCore

(FPCore (s u)
 :precision binary32
 (*
  u
  (+
   (* s 4.0)
   (* u (+ (* s 8.0) (* u (+ (* s 21.333333333333332) (* 64.0 (* s u)))))))))

C

float code(float s, float u) {
	return u * ((s * 4.0f) + (u * ((s * 8.0f) + (u * ((s * 21.333333333333332f) + (64.0f * (s * u)))))));
}

Fortran

real(4) function code(s, u)
    real(4), intent (in) :: s
    real(4), intent (in) :: u
    code = u * ((s * 4.0e0) + (u * ((s * 8.0e0) + (u * ((s * 21.333333333333332e0) + (64.0e0 * (s * u)))))))
end function

Julia

function code(s, u)
	return Float32(u * Float32(Float32(s * Float32(4.0)) + Float32(u * Float32(Float32(s * Float32(8.0)) + Float32(u * Float32(Float32(s * Float32(21.333333333333332)) + Float32(Float32(64.0) * Float32(s * u))))))))
end

MATLAB

function tmp = code(s, u)
	tmp = u * ((s * single(4.0)) + (u * ((s * single(8.0)) + (u * ((s * single(21.333333333333332)) + (single(64.0) * (s * u)))))));
end

Derivation

1. Initial program 60.2%

   \[s \cdot \log \left(\frac{1}{1 - 4 \cdot u}\right) \]
2. Add Preprocessing

3. Taylor expanded in u around 0 93.5%

   \[\leadsto \color{blue}{u \cdot \left(4 \cdot s + u \cdot \left(8 \cdot s + u \cdot \left(21.333333333333332 \cdot s + 64 \cdot \left(s \cdot u\right)\right)\right)\right)} \]

4. Final simplification 93.5%

   \[\leadsto u \cdot \left(s \cdot 4 + u \cdot \left(s \cdot 8 + u \cdot \left(s \cdot 21.333333333333332 + 64 \cdot \left(s \cdot u\right)\right)\right)\right) \]
5. Add Preprocessing

Alternative 3: 92.9% accurate, 6.4× speedup

Math

\[\begin{array}{l} \\ s \cdot \left(u \cdot \left(4 + u \cdot \left(8 + u \cdot \left(21.333333333333332 + u \cdot 64\right)\right)\right)\right) \end{array} \]

FPCore

(FPCore (s u)
 :precision binary32
 (* s (* u (+ 4.0 (* u (+ 8.0 (* u (+ 21.333333333333332 (* u 64.0)))))))))

C

float code(float s, float u) {
	return s * (u * (4.0f + (u * (8.0f + (u * (21.333333333333332f + (u * 64.0f)))))));
}

Fortran

real(4) function code(s, u)
    real(4), intent (in) :: s
    real(4), intent (in) :: u
    code = s * (u * (4.0e0 + (u * (8.0e0 + (u * (21.333333333333332e0 + (u * 64.0e0)))))))
end function

Julia

function code(s, u)
	return Float32(s * Float32(u * Float32(Float32(4.0) + Float32(u * Float32(Float32(8.0) + Float32(u * Float32(Float32(21.333333333333332) + Float32(u * Float32(64.0)))))))))
end

MATLAB

function tmp = code(s, u)
	tmp = s * (u * (single(4.0) + (u * (single(8.0) + (u * (single(21.333333333333332) + (u * single(64.0))))))));
end

Derivation

1. Initial program 60.2%

   \[s \cdot \log \left(\frac{1}{1 - 4 \cdot u}\right) \]
2. Add Preprocessing

3. Taylor expanded in u around 0 93.3%

   \[\leadsto s \cdot \color{blue}{\left(u \cdot \left(4 + u \cdot \left(8 + u \cdot \left(21.333333333333332 + 64 \cdot u\right)\right)\right)\right)} \]
4. Step-by-step derivation

   1. *-commutative 93.3%

      \[\leadsto s \cdot \left(u \cdot \left(4 + u \cdot \left(8 + u \cdot \left(21.333333333333332 + \color{blue}{u \cdot 64}\right)\right)\right)\right) \]

5. Simplified 93.3%

   \[\leadsto s \cdot \color{blue}{\left(u \cdot \left(4 + u \cdot \left(8 + u \cdot \left(21.333333333333332 + u \cdot 64\right)\right)\right)\right)} \]
6. Add Preprocessing

Alternative 4: 91.1% accurate, 7.3× speedup

Math

\[\begin{array}{l} \\ u \cdot \left(s \cdot 4 + s \cdot \left(u \cdot \left(8 + u \cdot 21.333333333333332\right)\right)\right) \end{array} \]

FPCore

(FPCore (s u)
 :precision binary32
 (* u (+ (* s 4.0) (* s (* u (+ 8.0 (* u 21.333333333333332)))))))

C

float code(float s, float u) {
	return u * ((s * 4.0f) + (s * (u * (8.0f + (u * 21.333333333333332f)))));
}

Fortran

real(4) function code(s, u)
    real(4), intent (in) :: s
    real(4), intent (in) :: u
    code = u * ((s * 4.0e0) + (s * (u * (8.0e0 + (u * 21.333333333333332e0)))))
end function

Julia

function code(s, u)
	return Float32(u * Float32(Float32(s * Float32(4.0)) + Float32(s * Float32(u * Float32(Float32(8.0) + Float32(u * Float32(21.333333333333332)))))))
end

MATLAB

function tmp = code(s, u)
	tmp = u * ((s * single(4.0)) + (s * (u * (single(8.0) + (u * single(21.333333333333332))))));
end

Derivation

1. Initial program 60.2%

   \[s \cdot \log \left(\frac{1}{1 - 4 \cdot u}\right) \]
2. Add Preprocessing

3. Taylor expanded in u around 0 91.4%

   \[\leadsto \color{blue}{u \cdot \left(4 \cdot s + u \cdot \left(8 \cdot s + 21.333333333333332 \cdot \left(s \cdot u\right)\right)\right)} \]
4. Step-by-step derivation

   1. distribute-lft-in 91.4%

      \[\leadsto \color{blue}{u \cdot \left(4 \cdot s\right) + u \cdot \left(u \cdot \left(8 \cdot s + 21.333333333333332 \cdot \left(s \cdot u\right)\right)\right)} \]

   2. *-commutative 91.4%

      \[\leadsto u \cdot \color{blue}{\left(s \cdot 4\right)} + u \cdot \left(u \cdot \left(8 \cdot s + 21.333333333333332 \cdot \left(s \cdot u\right)\right)\right) \]

   3. +-commutative 91.4%

      \[\leadsto u \cdot \left(s \cdot 4\right) + u \cdot \left(u \cdot \color{blue}{\left(21.333333333333332 \cdot \left(s \cdot u\right) + 8 \cdot s\right)}\right) \]

   4. fma-define 91.4%

      \[\leadsto u \cdot \left(s \cdot 4\right) + u \cdot \left(u \cdot \color{blue}{\mathsf{fma}\left(21.333333333333332, s \cdot u, 8 \cdot s\right)}\right) \]

   5. *-commutative 91.4%

      \[\leadsto u \cdot \left(s \cdot 4\right) + u \cdot \left(u \cdot \mathsf{fma}\left(21.333333333333332, s \cdot u, \color{blue}{s \cdot 8}\right)\right) \]

5. Applied egg-rr 91.4%

   \[\leadsto \color{blue}{u \cdot \left(s \cdot 4\right) + u \cdot \left(u \cdot \mathsf{fma}\left(21.333333333333332, s \cdot u, s \cdot 8\right)\right)} \]
6. Step-by-step derivation

   1. +-commutative 91.4%

      \[\leadsto \color{blue}{u \cdot \left(u \cdot \mathsf{fma}\left(21.333333333333332, s \cdot u, s \cdot 8\right)\right) + u \cdot \left(s \cdot 4\right)} \]

   2. distribute-lft-out 91.4%

      \[\leadsto \color{blue}{u \cdot \left(u \cdot \mathsf{fma}\left(21.333333333333332, s \cdot u, s \cdot 8\right) + s \cdot 4\right)} \]

   3. fma-undefine 91.4%

      \[\leadsto u \cdot \left(u \cdot \color{blue}{\left(21.333333333333332 \cdot \left(s \cdot u\right) + s \cdot 8\right)} + s \cdot 4\right) \]

   4. distribute-lft-out 91.4%

      \[\leadsto u \cdot \left(\color{blue}{\left(u \cdot \left(21.333333333333332 \cdot \left(s \cdot u\right)\right) + u \cdot \left(s \cdot 8\right)\right)} + s \cdot 4\right) \]

   5. associate-*r* 91.4%

      \[\leadsto u \cdot \left(\left(\color{blue}{\left(u \cdot 21.333333333333332\right) \cdot \left(s \cdot u\right)} + u \cdot \left(s \cdot 8\right)\right) + s \cdot 4\right) \]

   6. *-commutative 91.4%

      \[\leadsto u \cdot \left(\left(\color{blue}{\left(21.333333333333332 \cdot u\right)} \cdot \left(s \cdot u\right) + u \cdot \left(s \cdot 8\right)\right) + s \cdot 4\right) \]

   7. *-commutative 91.4%

      \[\leadsto u \cdot \left(\left(\left(21.333333333333332 \cdot u\right) \cdot \left(s \cdot u\right) + \color{blue}{\left(s \cdot 8\right) \cdot u}\right) + s \cdot 4\right) \]

   8. *-commutative 91.4%

      \[\leadsto u \cdot \left(\left(\left(21.333333333333332 \cdot u\right) \cdot \left(s \cdot u\right) + \color{blue}{\left(8 \cdot s\right)} \cdot u\right) + s \cdot 4\right) \]

   9. associate-*r* 91.4%

      \[\leadsto u \cdot \left(\left(\left(21.333333333333332 \cdot u\right) \cdot \left(s \cdot u\right) + \color{blue}{8 \cdot \left(s \cdot u\right)}\right) + s \cdot 4\right) \]

   10. distribute-rgt-in 91.4%

       \[\leadsto u \cdot \left(\color{blue}{\left(s \cdot u\right) \cdot \left(21.333333333333332 \cdot u + 8\right)} + s \cdot 4\right) \]

   11. +-commutative 91.4%

       \[\leadsto u \cdot \left(\left(s \cdot u\right) \cdot \color{blue}{\left(8 + 21.333333333333332 \cdot u\right)} + s \cdot 4\right) \]

   12. associate-*r* 91.4%

       \[\leadsto u \cdot \left(\color{blue}{s \cdot \left(u \cdot \left(8 + 21.333333333333332 \cdot u\right)\right)} + s \cdot 4\right) \]

   13. *-commutative 91.4%

       \[\leadsto u \cdot \left(s \cdot \left(u \cdot \left(8 + \color{blue}{u \cdot 21.333333333333332}\right)\right) + s \cdot 4\right) \]

7. Simplified 91.4%

   \[\leadsto \color{blue}{u \cdot \left(s \cdot \left(u \cdot \left(8 + u \cdot 21.333333333333332\right)\right) + s \cdot 4\right)} \]

8. Final simplification 91.4%

   \[\leadsto u \cdot \left(s \cdot 4 + s \cdot \left(u \cdot \left(8 + u \cdot 21.333333333333332\right)\right)\right) \]
9. Add Preprocessing

Alternative 5: 90.8% accurate, 8.4× speedup

Math

\[\begin{array}{l} \\ s \cdot \left(u \cdot \left(4 + u \cdot \left(8 + u \cdot 21.333333333333332\right)\right)\right) \end{array} \]

FPCore

(FPCore (s u)
 :precision binary32
 (* s (* u (+ 4.0 (* u (+ 8.0 (* u 21.333333333333332)))))))

C

float code(float s, float u) {
	return s * (u * (4.0f + (u * (8.0f + (u * 21.333333333333332f)))));
}

Fortran

real(4) function code(s, u)
    real(4), intent (in) :: s
    real(4), intent (in) :: u
    code = s * (u * (4.0e0 + (u * (8.0e0 + (u * 21.333333333333332e0)))))
end function

Julia

function code(s, u)
	return Float32(s * Float32(u * Float32(Float32(4.0) + Float32(u * Float32(Float32(8.0) + Float32(u * Float32(21.333333333333332)))))))
end

MATLAB

function tmp = code(s, u)
	tmp = s * (u * (single(4.0) + (u * (single(8.0) + (u * single(21.333333333333332))))));
end

Derivation

1. Initial program 60.2%

   \[s \cdot \log \left(\frac{1}{1 - 4 \cdot u}\right) \]
2. Add Preprocessing

3. Taylor expanded in u around 0 91.2%

   \[\leadsto s \cdot \color{blue}{\left(u \cdot \left(4 + u \cdot \left(8 + 21.333333333333332 \cdot u\right)\right)\right)} \]
4. Step-by-step derivation

   1. *-commutative 91.2%

      \[\leadsto s \cdot \left(u \cdot \left(4 + u \cdot \left(8 + \color{blue}{u \cdot 21.333333333333332}\right)\right)\right) \]

5. Simplified 91.2%

   \[\leadsto s \cdot \color{blue}{\left(u \cdot \left(4 + u \cdot \left(8 + u \cdot 21.333333333333332\right)\right)\right)} \]
6. Add Preprocessing

Alternative 6: 86.8% accurate, 9.9× speedup

Math

\[\begin{array}{l} \\ s \cdot \left(u \cdot \left(u \cdot 8\right) + \frac{u}{0.25}\right) \end{array} \]

FPCore

(FPCore (s u) :precision binary32 (* s (+ (* u (* u 8.0)) (/ u 0.25))))

C

float code(float s, float u) {
	return s * ((u * (u * 8.0f)) + (u / 0.25f));
}

Fortran

real(4) function code(s, u)
    real(4), intent (in) :: s
    real(4), intent (in) :: u
    code = s * ((u * (u * 8.0e0)) + (u / 0.25e0))
end function

Julia

function code(s, u)
	return Float32(s * Float32(Float32(u * Float32(u * Float32(8.0))) + Float32(u / Float32(0.25))))
end

MATLAB

function tmp = code(s, u)
	tmp = s * ((u * (u * single(8.0))) + (u / single(0.25)));
end

Derivation

1. Initial program 60.2%

   \[s \cdot \log \left(\frac{1}{1 - 4 \cdot u}\right) \]
2. Add Preprocessing

3. Taylor expanded in u around 0 86.9%

   \[\leadsto s \cdot \color{blue}{\left(u \cdot \left(4 + 8 \cdot u\right)\right)} \]
4. Step-by-step derivation

   1. *-commutative 86.9%

      \[\leadsto s \cdot \left(u \cdot \left(4 + \color{blue}{u \cdot 8}\right)\right) \]

5. Simplified 86.9%

   \[\leadsto s \cdot \color{blue}{\left(u \cdot \left(4 + u \cdot 8\right)\right)} \]

6. Taylor expanded in u around inf 86.7%

   \[\leadsto s \cdot \left(u \cdot \color{blue}{\left(u \cdot \left(8 + 4 \cdot \frac{1}{u}\right)\right)}\right) \]
7. Step-by-step derivation

   1. associate-*r/ 86.7%

      \[\leadsto s \cdot \left(u \cdot \left(u \cdot \left(8 + \color{blue}{\frac{4 \cdot 1}{u}}\right)\right)\right) \]

   2. metadata-eval 86.7%

      \[\leadsto s \cdot \left(u \cdot \left(u \cdot \left(8 + \frac{\color{blue}{4}}{u}\right)\right)\right) \]

8. Simplified 86.7%

   \[\leadsto s \cdot \left(u \cdot \color{blue}{\left(u \cdot \left(8 + \frac{4}{u}\right)\right)}\right) \]
9. Step-by-step derivation

   1. distribute-lft-in 86.7%

      \[\leadsto s \cdot \left(u \cdot \color{blue}{\left(u \cdot 8 + u \cdot \frac{4}{u}\right)}\right) \]

   2. distribute-rgt-in 86.7%

      \[\leadsto s \cdot \color{blue}{\left(\left(u \cdot 8\right) \cdot u + \left(u \cdot \frac{4}{u}\right) \cdot u\right)} \]

   3. clear-num 86.7%

      \[\leadsto s \cdot \left(\left(u \cdot 8\right) \cdot u + \left(u \cdot \color{blue}{\frac{1}{\frac{u}{4}}}\right) \cdot u\right) \]

   4. un-div-inv 87.0%

      \[\leadsto s \cdot \left(\left(u \cdot 8\right) \cdot u + \color{blue}{\frac{u}{\frac{u}{4}}} \cdot u\right) \]

   5. div-inv 87.0%

      \[\leadsto s \cdot \left(\left(u \cdot 8\right) \cdot u + \frac{u}{\color{blue}{u \cdot \frac{1}{4}}} \cdot u\right) \]

   6. metadata-eval 87.0%

      \[\leadsto s \cdot \left(\left(u \cdot 8\right) \cdot u + \frac{u}{u \cdot \color{blue}{0.25}} \cdot u\right) \]

10. Applied egg-rr 87.0%

    \[\leadsto s \cdot \color{blue}{\left(\left(u \cdot 8\right) \cdot u + \frac{u}{u \cdot 0.25} \cdot u\right)} \]
11. Step-by-step derivation

    1. *-commutative 87.0%

       \[\leadsto s \cdot \left(\left(u \cdot 8\right) \cdot u + \color{blue}{u \cdot \frac{u}{u \cdot 0.25}}\right) \]

    2. associate-/r* 87.0%

       \[\leadsto s \cdot \left(\left(u \cdot 8\right) \cdot u + u \cdot \color{blue}{\frac{\frac{u}{u}}{0.25}}\right) \]

    3. *-inverses 87.0%

       \[\leadsto s \cdot \left(\left(u \cdot 8\right) \cdot u + u \cdot \frac{\color{blue}{1}}{0.25}\right) \]

    4. un-div-inv 87.0%

       \[\leadsto s \cdot \left(\left(u \cdot 8\right) \cdot u + \color{blue}{\frac{u}{0.25}}\right) \]

12. Applied egg-rr 87.0%

    \[\leadsto s \cdot \left(\left(u \cdot 8\right) \cdot u + \color{blue}{\frac{u}{0.25}}\right) \]

13. Final simplification 87.0%

    \[\leadsto s \cdot \left(u \cdot \left(u \cdot 8\right) + \frac{u}{0.25}\right) \]
14. Add Preprocessing

Alternative 7: 86.6% accurate, 12.1× speedup

Math

\[\begin{array}{l} \\ s \cdot \left(u \cdot \left(4 + u \cdot 8\right)\right) \end{array} \]

FPCore

(FPCore (s u) :precision binary32 (* s (* u (+ 4.0 (* u 8.0)))))

C

float code(float s, float u) {
	return s * (u * (4.0f + (u * 8.0f)));
}

Fortran

real(4) function code(s, u)
    real(4), intent (in) :: s
    real(4), intent (in) :: u
    code = s * (u * (4.0e0 + (u * 8.0e0)))
end function

Julia

function code(s, u)
	return Float32(s * Float32(u * Float32(Float32(4.0) + Float32(u * Float32(8.0)))))
end

MATLAB

function tmp = code(s, u)
	tmp = s * (u * (single(4.0) + (u * single(8.0))));
end

Derivation

1. Initial program 60.2%

   \[s \cdot \log \left(\frac{1}{1 - 4 \cdot u}\right) \]
2. Add Preprocessing

3. Taylor expanded in u around 0 86.9%

   \[\leadsto s \cdot \color{blue}{\left(u \cdot \left(4 + 8 \cdot u\right)\right)} \]
4. Step-by-step derivation

   1. *-commutative 86.9%

      \[\leadsto s \cdot \left(u \cdot \left(4 + \color{blue}{u \cdot 8}\right)\right) \]

5. Simplified 86.9%

   \[\leadsto s \cdot \color{blue}{\left(u \cdot \left(4 + u \cdot 8\right)\right)} \]
6. Add Preprocessing

Alternative 8: 74.1% accurate, 21.8× speedup

Math

\[\begin{array}{l} \\ u \cdot \left(s \cdot 4\right) \end{array} \]

FPCore

(FPCore (s u) :precision binary32 (* u (* s 4.0)))

C

float code(float s, float u) {
	return u * (s * 4.0f);
}

Fortran

real(4) function code(s, u)
    real(4), intent (in) :: s
    real(4), intent (in) :: u
    code = u * (s * 4.0e0)
end function

Julia

function code(s, u)
	return Float32(u * Float32(s * Float32(4.0)))
end

MATLAB

function tmp = code(s, u)
	tmp = u * (s * single(4.0));
end

Derivation

1. Initial program 60.2%

   \[s \cdot \log \left(\frac{1}{1 - 4 \cdot u}\right) \]
2. Add Preprocessing

3. Taylor expanded in u around 0 74.7%

   \[\leadsto \color{blue}{4 \cdot \left(s \cdot u\right)} \]
4. Step-by-step derivation

   1. associate-*r* 74.9%

      \[\leadsto \color{blue}{\left(4 \cdot s\right) \cdot u} \]

5. Simplified 74.9%

   \[\leadsto \color{blue}{\left(4 \cdot s\right) \cdot u} \]

6. Final simplification 74.9%

   \[\leadsto u \cdot \left(s \cdot 4\right) \]
7. Add Preprocessing

Alternative 9: 73.9% accurate, 21.8× speedup

Math

\[\begin{array}{l} \\ 4 \cdot \left(s \cdot u\right) \end{array} \]

FPCore

(FPCore (s u) :precision binary32 (* 4.0 (* s u)))

C

float code(float s, float u) {
	return 4.0f * (s * u);
}

Fortran

real(4) function code(s, u)
    real(4), intent (in) :: s
    real(4), intent (in) :: u
    code = 4.0e0 * (s * u)
end function

Julia

function code(s, u)
	return Float32(Float32(4.0) * Float32(s * u))
end

MATLAB

function tmp = code(s, u)
	tmp = single(4.0) * (s * u);
end

Derivation

1. Initial program 60.2%

   \[s \cdot \log \left(\frac{1}{1 - 4 \cdot u}\right) \]
2. Add Preprocessing

3. Taylor expanded in u around 0 74.7%

   \[\leadsto \color{blue}{4 \cdot \left(s \cdot u\right)} \]
4. Add Preprocessing

Alternative 10: 16.8% accurate, 109.0× speedup

Math

\[\begin{array}{l} \\ 0 \end{array} \]

FPCore

(FPCore (s u) :precision binary32 0.0)

C

float code(float s, float u) {
	return 0.0f;
}

Fortran

real(4) function code(s, u)
    real(4), intent (in) :: s
    real(4), intent (in) :: u
    code = 0.0e0
end function

Julia

function code(s, u)
	return Float32(0.0)
end

MATLAB

function tmp = code(s, u)
	tmp = single(0.0);
end

Derivation

1. Initial program 60.2%

   \[s \cdot \log \left(\frac{1}{1 - 4 \cdot u}\right) \]
2. Add Preprocessing

4. Applied egg-rr 16.7%

   \[\leadsto s \cdot \color{blue}{\left(0.5 \cdot \mathsf{log1p}\left(4 \cdot u\right) - 0.5 \cdot \mathsf{log1p}\left(4 \cdot u\right)\right)} \]
5. Step-by-step derivation

   1. +-inverses 16.7%

      \[\leadsto s \cdot \color{blue}{0} \]

6. Simplified 16.7%

   \[\leadsto s \cdot \color{blue}{0} \]

7. Taylor expanded in s around 0 16.7%

   \[\leadsto \color{blue}{0} \]
8. Add Preprocessing

Reproduce

herbie shell --seed 2024130 
(FPCore (s u)
  :name "Disney BSSRDF, sample scattering profile, lower"
  :precision binary32
  :pre (and (and (<= 0.0 s) (<= s 256.0)) (and (<= 2.328306437e-10 u) (<= u 0.25)))
  (* s (log (/ 1.0 (- 1.0 (* 4.0 u))))))