Interesting question. As Julio correctly pointed out, $d\vec{T}=d\vec{r}\times \vec{F} + \vec{r}\times d\vec{F}$. (Note that the original question had the order of $\vec{r}$ and $\vec{F}$ reversed.)
How do we interpret these two parts? The first term $(1)=d\vec{r}\times \vec{F}$ contains the change in torque due to a change in the moment arm $\vec{r}$, while the second term $(2)=\vec{r}\times d\vec{F}$ would correspond the change in torque due to a change in a force $\vec{F}$.
If you have forces distributed over a volume $\mathcal{V}$, and would like to calculate net torque, then the relevant term is the second one. E.g. if we want to calculate torque to the weight of an object, the infinitesimal force $d\vec{F}=-\rho(\vec{r})g\hat{z}dV$
\begin{equation}
\vec{T}=\int_{\mathcal{V}}\vec{r}\times \left(-\rho(\vec{r})g\hat{z}\right)dV
\end{equation}
We can of course move the constant stuff outside of the integral
\begin{aligned}
\vec{T}&=-g\left[\int_{\mathcal{V}}\vec{r}\rho(\vec{r})dV\right]\times\hat{z} \\
&= - gM\vec{r}_{CM} \times \hat{z} \\
&= \vec{r}_{CM} \times \left(-Mg\hat{z}\right)
\end{aligned}
And this last expression is the standard result that when you calculate the torque due to the weight of an object, you can treat it as if the weight was applied at the center of mass.
Anyhow this is an example where the second term is the relevant one. But I'm sure you can cook up an example where the first term is relevant. E.g. if you are applying a constant force at a point that is moving relative the the origin, at what rate is the Torque changing?