The plutonium in bombs is essentially "super high grade" reactor fuel. Even degraded after decades in storage it is still far, far better than what is typically used. It just needs to be converted into the MOX (metal oxide) fuel pellets and then used in a reactor, pretty much as-is.
While many supercomputers were funded by stockpile stewardship, the goal was to produce high performance computers capable of a wide range of simulation needs.
One good example would be NERSC at LBL- it's unclassified research only, and their series of supercomputers were never intended to simulate ageing nuclear weapons.
Hard to say exactly what goes on in the classified supercomputers, but they certainly weren't spending much of their time simulating aging nuclear weapons- that was the ostensible reason.
Decades, certainly. All but the first few generations of bombs were designed for long periods of storage.
Notably, many of the TOP500 supercomputers were built with the singular goal of simulating the ageing of nuclear weapons in storage.
If a supercomputer is owned by the DoE or SANDIA, then that's what it is for.
> turn those swords into ploughshares relatively easily.
Yes! Cold-war era warheads from both the Soviet Union and the US have been used as nuclear fuel. A notable one was the Megatons to Megawatts program: https://en.wikipedia.org/wiki/Megatons_to_Megawatts_Program
The plutonium in bombs is essentially "super high grade" reactor fuel. Even degraded after decades in storage it is still far, far better than what is typically used. It just needs to be converted into the MOX (metal oxide) fuel pellets and then used in a reactor, pretty much as-is.