I am doing an analysis of experimental results in which I need to repeat the same GaussianFilter hundred of times on different data. As explained in the documentation, GaussianFilter just convolves the data with a Gaussian kernel. Does it recompute the kernel every time I call the function, or will it somehow preserve and reuse the previous kernel? Would it be more efficient computationally for me to precompute the kernel (which I could do easily by applying GaussianFilter to a KroneckerDelta array), then do hundreds of ListConvolves instead of hundreds of GaussianFilters?