The study of turbulence, "the most important unsolved problem in classical physics", as Feynman once famously asserted, has seen a long period of expansion and popularity in the era of high-performance computing. This prompted some scientists to state that "turbulence is a solved problem", since simulations could deal with ever-growing Reynolds numbers and complex geometries. However, the imminent arrival of exascale simulations reminds us that there remains the almost philosophical question if being able to simulate a physical phenomenon implies we have also solved the mathematical questions? We shall revisit some of the history of turbulence in the hope that what seems to be the dusk of computational fluid dynamics can become the dawn of a more complete mathematical understanding and modeling of turbulence. The focus of this talk will be on the mathematical and computational science victories as well as current setbacks in embracing the notion that "turbulence is solved".