I think “not viable for research” is not a thing, haha. At least, if the ratio of compute to communication is right, we can test our algorithms on old hardware

I think “not viable for research” is not a thing, haha. At least, if the ratio of compute to communication is right, we can test our algorithms on old hardware using smaller problems. Edit: The SuiteSparse collection (formerly UF matrix collection) is really great for this. You can download matrices with their year and some context about their application. The ones smart people were trying to figure out how to solve 20 years ago on clusters. And then you can marvel at how your laptop smashes them using PARDISO or some other direct sparse solver. Or you can try to figure out a clever way to solve them with 10 year old hardware.