I ran across the following conjecture, which I checked numerically and seemed to check out. For $x\to \infty$ the sum $$\Theta(x)=\sum_{i=1,3,5..} \text{Ei}{\left(-\frac{i^2\pi^3}{4 x^2}\right)},$$ has the asymptotic expression $$\Theta(x)_{x\to \infty}=\ln{2}-\frac{x}{\pi},$$ where $\text{Ei}(x)$ is defined as $$\text{Ei}(x)=\int_{-\infty}^x\frac{e^t}{t}\text{d}t.$$
Showing that this is actually true proves to be a significantly hard task that once again, I'm asking for your help. How would one go about proving this?
Assume that $x>0$. Using the alternative notation $E_1$, we find \begin{align*} \Theta (x) & = - \sum\limits_{n = 0}^\infty {E_1 \!\left( {\frac{{\pi ^3 }}{{4x^2 }}(2n + 1)^2 } \right)} = - \sum\limits_{n = 0}^\infty {\int_{(\pi /x)^2 }^{ + \infty } {\frac{{e^{ - (\pi /4)(2n + 1)^2 t} }}{t}dt} } \\ & = - \frac{1}{2}\int_{(\pi /x)^2 }^{ + \infty } {\frac{{\theta _2 (0|it)}}{t}dt} = - \int_{\pi /x}^{ + \infty } {\frac{{\theta _2 (0|is^2 )}}{s}ds} \\ & = - \int_{\pi /x}^{ + \infty } {\frac{{\theta _4 (0|is^{ - 2} )}}{{s^2 }}ds} = - \int_0^{x/\pi } {\theta _4 (0|it^2 )dt} \\ & = \int_0^{x/\pi } {(1 - \theta _4 (0|it^2 ))dt} - \frac{x}{\pi } \end{align*} where $\theta_i$ are the theta functions and we used the known transformation formula. Hence, $$ \mathop {\lim }\limits_{x \to + \infty } \left( {\Theta (x) + \frac{x}{\pi }} \right) = \int_0^{ + \infty } {(1 - \theta _4 (0|it^2 ))dt} $$ provided the improper integral on the right-hand side exists. Your claim is that the integral exists and it is $\log 2$. It is known that (see $(1.14.29)$ in F. Oberhettinger's Tables of Mellin Transforms) $$ \int_0^{ + \infty } {t^{s - 1} (1 - \theta _4 (0|it^2 ))dt} = \frac{{\Gamma (s/2)}}{{\pi ^{s/2} }}\sum\limits_{n = 1}^\infty {\frac{{( - 1)^{n + 1} }}{{n^s }}} $$ provided $\Re s>0$. This can be proved for $\Re s>1$ via term-by-term integration and can be extended to $\Re s>0$ via analytic continuation. Thus, with $s=1$, $$ \int_0^{ + \infty } {(1 - \theta _4 (0|it^2 ))dt} = \frac{{\Gamma (1/2)}}{{\pi ^{1/2} }}\sum\limits_{n = 1}^\infty {\frac{{( - 1)^{n + 1} }}{n}} = \sum\limits_{n = 1}^\infty {\frac{{( - 1)^{n + 1} }}{n}} = \log 2 $$ which is the desired result.
Note that we have an exact expansion for $x>0$ in terms of the complementary error function: $$ \Theta (x) = - \frac{x}{\pi } + \log 2 + \sum\limits_{n = 1}^\infty {\frac{{( - 1)^n }}{n}\operatorname{erfc}\left( {\frac{x}{{\sqrt \pi }}n} \right)} . $$