I don’t know if this particular funding is reasonable or not, whether it’s cost-effective, but my kneejerk reaction is that I don’t really feel that a large parallel compute system for a university is the worst thing that one could get.
I do kind of wonder whether the university needs to actually own the computer, though. Like, there’s commercial demand for a large parallel computer. This isn’t the Large Hadron Collider or something. Maybe it’d make sense to just do grants for parallel compute time.
EDIT: Plus it sounds like it isn’t just owned by the university, but also co-located:
Edinburgh University had already spent £31m building housing for it.
I mean, why? The university presumably isn’t aiming to advance the system administration or maintenance of parallel compute clusters, but the software that runs on them.