Have you thought about what powers complex artificial intelligence technologies such as ChatGPT? The answer is High Performance Computing (HPC)!
How does HPC work? Imagine that you have a factory of people trying to make sense of the information stored in Wikipedia. Let’s say that each factory worker performs all these steps: loading a Wikipedia page, copying the text, reading the text and marking potential useful facts, copying only the useful facts into a central repository, labeling each fact with category names, then notifying the other workers that they are done so that the next page can be processed. If one person is doing all of these things and waiting to hear back from other workers before they can start their assignment, how long do you think it would take to finish processing all Wikipedia pages? (Hint: it would be a rather long time because the number of factory workers limits how much information you are able to get out of Wikipedia).
Now, what if each person could focus on one specific step of that process? For example, the first worker just loads Wikipedia pages; the next just copies text, and so on. This way of processing is much more efficient because context switching takes time (remember the last time you tried to watch TV while doing your taxes?), and each worker is able to quickly process multiple pages at once – advantageous because with the one-worker-does-everything approach, someone calling in sick could prevent a whole page from getting processed. And best of all, every single worker is working at the same time without having to wait for anyone else to be done.
HPC is very similar to the second method of processing information. It allows computational work to be divided and parallelized in such a way that information can be processed or calculations can be performed at speeds we were never able to achieve even 20 years ago. That’s how huge Machine Learning models like ChatGPT exist (by gobbling up as much information as it can find), and how we are able to build supercomputers that can do a quintillion calculations per second.
Praxis Engineering develops and supports HPC in a multitude of ways that can range from low-level analysis at the hardware level to creation of applications taking advantage of the powers of high performance computing. We also offer internal training and certification opportunities to bring employees up-to-speed. There is still much to learn about HPC and its ever-changing landscape.
Join us on this journey!
To learn more about Praxis Engineering Technologies, please visit www.praxiseng.com.
SPONSORED CONTENT: This content is written on or behalf of our Sponsor.
