Is the Sherrington-Kirkpatrick model relevant for real spin glasses?

I describe the results of numerical simulations which test whether the Sherrington-Kirkpatrick model, which was solved by Parisi, applies to short-range spin glasses. I conclude that it probably does for dimension d greater than 6, but there appear to be some differences in lower dimensions. In particular, there does not appear to be the line of transitions in a magnetic field ( AT line) which is seen in the SK model.