Hello All,
I'm Anvita Gupta, a senior at BASIS Scottsdale High School, and in the coming months I will be delving into the exciting world of supercomputing. How, you ask? In the broadest terms, I will be building a supercomputer. Specifically, I will be building a High Performance Computing Cluster (HPCC), which is composed of a number of smaller computers that work together in unison. Each computer is called a node.
You might wonder why we even want to build a supercomputer. The answer can lie in something as simple as, say, a spring. We all learn in physics class that F=-kx, but all our laws apply only for ideal springs. To model the movement of springs in the real world, we must take into account several additional variables. For instance- surrounding temperature, heating up of the metal in the spring, wind conditions. For a theoretical spring manufacturer, then, it becomes important to build accurate simulations of these springs. Where will he run these simulations? After all, they will require an enormous amount of computing power. This is where supercomputers come in.
This very rudimentary example (which I might have come up with so that my title would be witty) illustrates one very important concept. The amount of data we have is growing exponentially, and along with it grows our need for infrastructure to handle data processing, and software to find patterns in the data. "Big Data" has become a buzzword not just in the offices of Silicon Valley giants, but even in the vocabulary of us mortals.
My goal is to build a HPCC and optimize it to handle clinical data, for personalized medicine and better disease diagnosis. My goal is efficiency; both of money and of time. I will follow a couple of main stages: design of the computing cluster, actual building of the HPCC, installing/modifying medical software to run on it, and testing my HPCC on medical data.
Within each stage there are a number of factors to consider, such as how to get the nodes to work together, how to keep the HPCC from overheating and catching on fire, (more importantly) how to make the cluster look aesthetically appealing. And in general, how to fit in the greatest amount of computing in the lowest cost and area. I will be interning at AMBA Solutions, a local IT and consulting company.
Looking forward to embarking on this journey!
Anvita
I'm Anvita Gupta, a senior at BASIS Scottsdale High School, and in the coming months I will be delving into the exciting world of supercomputing. How, you ask? In the broadest terms, I will be building a supercomputer. Specifically, I will be building a High Performance Computing Cluster (HPCC), which is composed of a number of smaller computers that work together in unison. Each computer is called a node.
You might wonder why we even want to build a supercomputer. The answer can lie in something as simple as, say, a spring. We all learn in physics class that F=-kx, but all our laws apply only for ideal springs. To model the movement of springs in the real world, we must take into account several additional variables. For instance- surrounding temperature, heating up of the metal in the spring, wind conditions. For a theoretical spring manufacturer, then, it becomes important to build accurate simulations of these springs. Where will he run these simulations? After all, they will require an enormous amount of computing power. This is where supercomputers come in.
This very rudimentary example (which I might have come up with so that my title would be witty) illustrates one very important concept. The amount of data we have is growing exponentially, and along with it grows our need for infrastructure to handle data processing, and software to find patterns in the data. "Big Data" has become a buzzword not just in the offices of Silicon Valley giants, but even in the vocabulary of us mortals.
My goal is to build a HPCC and optimize it to handle clinical data, for personalized medicine and better disease diagnosis. My goal is efficiency; both of money and of time. I will follow a couple of main stages: design of the computing cluster, actual building of the HPCC, installing/modifying medical software to run on it, and testing my HPCC on medical data.
Within each stage there are a number of factors to consider, such as how to get the nodes to work together, how to keep the HPCC from overheating and catching on fire, (more importantly) how to make the cluster look aesthetically appealing. And in general, how to fit in the greatest amount of computing in the lowest cost and area. I will be interning at AMBA Solutions, a local IT and consulting company.
Looking forward to embarking on this journey!
Anvita
