I'm playing around with some ideas in my head and thus I'm trying to do some fast motion simulations of solution spectra at low (2-5 GHz specifically) frequencies with molecules that have large hyperfine coupling interactions. I'm trying to use garlic and it begins to crash when I move to lower frequencies. The position at which the program stops computing spectra depends on the size of the hyperfine coupling interaction: a lower hyperfine coupling enables simulations at lower frequencies. Is the crashing at lower frequencies a bug? Maybe I am using the wrong program and chili would be better?
The algorithm used to calculate the peak positions in garlic runs into problems with extremely large hyperfine couplings, especially at the low frequencies you're using. You're better off using pepper, it does a full matrix diagonalization and can more easily handle the weird stuff that occurs when the hyperfine is on the same order as the electron Zeeman.
But garlic does need to have better error messages for these situations.