calculus - Contour Integral for Cosine and a rational function

I've been trying to figure out this integral via use of residues:

$$\int_{-\infty}^{\infty} \displaystyle \frac{\cos{5x}}{x^4+1}dx$$

The usual semicircle contour wont work for this guy as the integrated is unbounded.
My idea came from a book I was reading on contour integration, where we let

$$f(z) = \displaystyle\frac{e^{(-5iz)}}{2(z^4+1)}+\displaystyle\frac{e^{(5iz)}}{2(z^4+1)}$$

And do the integral in the complex play as follows:

$\gamma_{1}= \text{The contour taken to be the top half of the circle in the counter clockwise direction}$ This contour uses the second term in $f(z)$

$\gamma_{2}= \text{The contour taken from $-R$ to $R$ on the real axis}$

$\gamma_{3}= \text{The contour taken to be the bottom half of the circle in the clockwise direction}$ This uses the first term in $f(z)$.

In the end, the contours $\gamma_{1}$ and $\gamma_{3}$ are bounded and will tend to $0$ as $R$ goes to infinity, so that we're left with the two integrals that we want.

My issue now is that when computing residues..everything seems to be cancelling out and I'm getting $0$. Should I take different contour? I'm really not sure what I did wrong.

Answer

I've given the skeleton of my work below. Fill in any missing pieces and check your answer against mine.

Using $\gamma=[-R,R]\cup Re^{i[0,\pi]}$ and the simple poles at $\frac{1+i}{\sqrt2}$ and $\frac{-1+i}{\sqrt2}$ inside $\gamma$
$$
\begin{align}
\int_{-\infty}^\infty\frac{\cos(5x)}{x^4+1}\mathrm{d}x
&=\mathrm{Re}\left(\int_\gamma\frac{e^{i5z}}{z^4+1}\mathrm{d}z\right)\\
&=\mathrm{Re}\left(2\pi i\left(\left[\frac{e^{i5z}}{4z^3}\right]_{z=\frac{1+i}{\sqrt2}}

+\left[\frac{e^{i5z}}{4z^3}\right]_{z=\frac{-1+i}{\sqrt2}}\right)\right)\\
&=\mathrm{Re}\left(\frac{\pi}{2i}e^{-5/\sqrt2}\left(\frac{1+i}{\sqrt2}e^{i5/\sqrt2}
-\frac{1-i}{\sqrt2}e^{-i5/\sqrt2}
\right)\right)\\
&=\pi e^{-5/\sqrt2}\mathrm{Im}\left(\frac{1+i}{\sqrt2}e^{i5/\sqrt2}\right)\\
&=\pi e^{-5/\sqrt2}\mathrm{Im}\left(e^{i(5/\sqrt2+\pi/4)}\right)\\
&=\pi e^{-5/\sqrt2}\sin\left(\frac5{\sqrt2}+\frac\pi4\right)
\end{align}
$$
Mathematica 8 agrees numerically, but its closed form involves complex functions and looks nothing like what I have above.