There are many articles out there on the applications of HPC. This one is different. Hal Hodson of the New Scientist takes a look at HPC making discoveries that “scientists can’t”. At this point I would like to make two observations:
- Scientists can – they are running the supercomputers! The supercomputers however, actually can’t, without the scientists that is. (Somebody needs to be there to turn on the switch and pay the power bill.)
- Scientists couldn’t do basic operations like multiply extremely large matrices together without making a mistake either (that is, without computers) – but with computers they can do that too – and the scientists are still at the helm.
What the two points above bring to the foreground is that HPC helps scientists do what scientists couldn’t do without HPC. Now, that is circular enough, but it is true. However HPC can’t do anything without the scientists.
Back to the article – what this article does present is a few different angles that most “What is HPC used for?” articles don’t have. It’s not a new wind farm, a new way to save logistics companies money on their fuel bills, or making jet engines quieter. It’s more like a computer reading hundreds of thousands of research papers up to a date ten years ago, and with that knowledge, making discoveries that the scientists made the hard way. In addition, the computer also made a few potential discoveries that the scientists might have missed.
What does this say about the future? Put down that test tube Bunsen, and get ready to pony up for a serious electricity bill. It’s faster, still cheaper, and apparently more comprehensive to make discoveries this way. Curious now? Here it is: