Remind me what the blocker on Windows high AM integrals is, again?

No particular high AM blocker, but there is a high derivative blocker for Windows. Basically, it compiles ~1099 files in 3h, then hangs on the final file and fails to build under the 6h. Jet saw this, too, when he was building L2 on his local Windows box. It worked when left overnight, but on Azure through c-f it's a strict 6h limit. No lowering of AM for the Hessian ints helped any.

Since psi has so few analytic Hessians and they add so much expense to the L2 build in general, I figured the most helpful sol'n was to allow psi to live w/o them w/o too much fuss.