supercomputer

You're assuming the goal is to build the fastest supercomputer in the world.

The problem you were seeing wasn't the supercomputer per se, it was the sheer awfulness of Lustre.

Comparing Bitcoin miners to any supercomputer or server farm is meaningless at best and deceiving at worst.

A lot of that comes from endless days of supercomputer simulation that told them where to place the engine mounts.

This is because the supercomputer community has long ignored the Internet-style of computation (MapReduce etc).

The extrapolation on the top500 supercomputer list [1] estimates the first EFlop computer in 2019.

Perhaps medical research could be used to increase public interest in funding supercomputer research?

You can't write your code in any language you want on a supercomputer.

The current fastest supercomputer is rated at 33,862.7 TFlop/s, or 247 times faster.

As pointed out elsewhere, an exaflop supercomputer will probably come around beforehand.

Maybe for embarrassingly parallel tasks, but if you require nontrivial interprocess communication, a server farm can't compete with the interconnect of a modern supercomputer.

Right now it's a single shared processor emulated with cryptography and massive replication - but that's not an architecture that will get you supercomputer performance.

However,when simulating a nuclear explosion or a weather system (maybe that's what you mean by "real" workloads?), the heavy node-to-node communication makes the supercomputer much, much better suited.

Decades ago, when people worked on supercomputer, hundreds of millions dollars thrown to "parallel compiler", "shared memory machine" which aimed to reduce the complexity of parallel computing for programmers.

It called "embarrassingly parallel" because it is embarrassing for the Cray and other supercomputer vendors sales people when their expensive hardware does not outperform a 5 year old solution that cost 1/2 their fancy gear;) Otherwise data parallel is an other phrase that is used for the same concept.