Even into the 90s, numerical simulation of naiver-stokes was pretty minimal (I worked with the NASA Numerical Aerodynamic Simulation folks at Ames at the time) and yet up to then the SR-71, X-1B, Valkyrie etc were all designed, built and flown, not to mention some spacecraft too that you might have heard of :-).

There was some computational modeling, but mostly it was old fashioned engineering: slide rules, models, and as someone else remarked, some over-engineering. The same way people built bridges 150 years ago when that was a black art.

There's a wonderful/sad picture at the Hiller museum on 101 in San Carlos. It shows the day the Boeing SST program† was shut down. In front of the mockup (which is also at the museum) are wooden workbenches with vices attached.

It made me realize that not only were they designing everything with slide rules, but people then built all the parts by hand from drawings a draftsman (not draftsperson in those days unfortunately) also produced by hand from hand-sketches and notes by and engineer. Honestly this blew my mind when I realized that the ME-262, SR-71 and LEM were produced in the same way.

And I should have known: at MIT I took Unified the last year (1983) that the practical work was done entirely on old WW II surplus machine tools. In the next room, not yet unpacked, were early CAD systems.

† I still wonder if Boeing was every really serious about passenger SST.