Also, WWII accelerated many discoveries in several areas of mathematics (in America, at least): statistics, optimization, control, numerical analysis. To such an extent, that many of those discoveries are still the main object of study in their respective areas and are being extended to more general versions.

yeah, great, let's start a few more wars to foster progress and inventions ...

/s

Well, how can we get the benefit of wars without the obvious downsides?

Fast feedback loops, obvious win / lose results (a battle takes a few hours, you are dead or not).

Randomly fire members of upper management (equivalent of organisational change by enemy sniper action)?

The research end wasn't really the section getting hit by sniper fire, so that bit probably won't fly.