HPC on Demand promises to change the world as we know it. For the first time in history, High Performance Computing (also known as Supercomputing in many cases) is no longer reserved for large, cash-rich companies or well-funded academic institutions. HPC on Demand allows anyone facing a complex compute-intensive problem to solve it in a pay-per-use, self-service cloud model. This is great news for everyone, because according to the Small Business Administration (SBA), small firms are 16 times more innovative than large firms! The one clear advantage large firms have are capital and human resources, to acquire, assemble, and maintain giant High Performance Computing clusters. With HPC on Demand however, this power is now available to anyone, with no up-front investment required.
According to a recent, sponsored article in Datacenter Dynamics, 2014 could be the year of the supercomputer. As someone who lived through “the year of the Linux desktop”, and later, “the year of the virtual desktop”, on an annual basis over the past 2 decades, I’m cynical about predicting trends or proclaiming victory for emerging technologies from the vendor or analyst side of the market. Ultimately, consumers and small businesses will decide if HPC on Demand explodes like we’d all like it to. So instead of trying to predict the market success (or timing) of HPC on Demand, let’s imagine how society can benefit from its availability…
Energy conservation and efficiency is of critical importance to our survival as a global society. Energy producers invest massively in High Performance Computing to find new sources of oil, natural gas, and other resources. Why shouldn’t those of us closer to consumption of energy have the same advantage? Sure, it’s unlikely that individuals will run simulations on how best to drive their cars or set their thermostats, but it’s pretty common that they’ll consider buying products designed to optimize their energy use. This brings us to small firms designing and building aerodynamic devices for vehicles, for example. How do they know that their design and manufacturing dollars will actually pay off for their customers, and in turn, ensure a healthy return on investment?
One method is to use a wind tunnel and run physical simulations at scale. The problem with testing at scale is that you then have to validate your results with life sized models, as well as reconcile the wind tunnel data with real-world results. This is expensive and time consuming for most applications, and is a barrier to entry for small firms to utilize.
Another method is to just guess. Everyone knows that air generally flows less turbulently over curved surfaces and edges, and teardrop shapes create less drag than others, for example. In this scenario, the firm will design and build parts, sell them to consumers, and hope they work. They won’t be able to provide any credible data to the customer, and if it doesn’t work in practice as well as in the designer’s mind, they will simply not recover their investment. This is bad business, especially since tools and techniques are available to ensure this doesn’t happen. Building products that don’t work or don’t sell is a fatal blow to any business, especially a small one.
Computational Fluid Dynamics delivered via HPC on Demand helps designers visualize the behavior of their work without even building scale models of them. They can adjust variables and run simulations at lightning speed, for pennies on the dollar compared to other methods. Most importantly, they can bring products to market they know will work, and can present credible data to consumers in the process. Solving these types of problems on the Nimbix cloud, for example, can be done for about $100-$200 (USD) per job on average. That’s a far cry from building a model and paying for wind tunnel time, and even farther less than producing a product that doesn’t work.
As individuals we are growing ever more interested in our personal and family genetics, as a powerful health tool. For example, if we know we are predisposed to high blood pressure, we may decide to consume less salt in our diets. This is a very broad field that is gaining traction because as a society we are learning that our health is a personal responsibility. This doesn’t mean we no longer take advice from physicians – rather, it allows us to be proactive with our livelihood rather than wait until the doctor tells us to change our behavior.
Bioinformatics is a very common use of supercomputing power in academia and large biotechnology firms leveraging their own High Performance Computing resources. What about small business looking to offer consumer services? HPC on Demand brings these supercomputing capabilities to these firms so they can offer products and services to help consumers make better health decisions on a daily basis. With a healthier society, we all win.
Other Uses for HPC on Demand
HPC on Demand can help independent film makers produce 3D renderings at a fraction of the cost of large movie studios. It can help students advance their STEM studies without having to wait in line for computing time from their schools or colleges. Basically, it puts the power of supercomputing in the hands of the individual or small firm to crunch their way to the “next big thing”. Just like the personal computer brought technology into the hands of the masses, HPC on Demand is poised to do the same for supercomputing. The results will be game changing.
 Source: http://www.sba.gov/advocacy/7540/28811