Turning off Heuristic processing as suggested by eschwartz and DoctorOhh did the trick. It converted in less than 10 seconds this time where as I let it run for 6 hours after posting this thread last night and it still hadn't budged off 1%.
I got my wife a new Kindle Fire HDX 8.9 for Christmas to replace her old first gen Nook. I had to convert her entire library of 1860 EPUBs (she's a MAJOR bookworm!) to AZW3. I did all of them with Heuristic processing turned on as I found it produced a better end result.
I've used the old calibre book editor to fix the editing errors of others on my wife's EPUBs, so I'm pretty familiar with how they were built. This PENGUIN ebook didn't seem unusual until I discovered that it contained a 50 level deep series of nested <Div> statements that spanned the entire length of the ebook. I now suspect this is what caused calibre to hang.
theducks, why did you say that calibre doesn't use multiple cores ? There are settings under Advanced/Miscellaneous for using multiple cores to process simultaneous conversion jobs. I've monitored calibre's core usage with 100 cued conversions and it appeared to be using every core -- completing them all successfully in just a few minutes. Am I misunderstanding how this setting is used ?