if that's true, it seems really wasteful honestly. why not reimplement the functionality using a native instruction set rather then emulating some other processor?
That seems extremely efficient to me. That'd be a bad way to build a brand new calculator, perhaps, but the quickest way to get an existing, what, 40 year old?, firmware up and running with the least number of gotchas.
I doubt there are competent and cost-effective engineering teams in existence who could exactly match HP's numeric libraries in a $150 calculator that's guaranteed to sell a tiny number of units.
