In my case, each sum contains 264 summands, so results that only hold in the limit of infinitely-many summands may also be useful if the finite case can't be conveniently described.
This is a special case of What's the distribution of $L_1$ norm of Gaussian random variables? (which does not have an answer).