NECSTonatNECST

During the Easter Holiday some of my friends came to visit me in Stockholm and, talking about this course i am attending in KTH, it came to my attention that two of them created a start-up two months ago. Its name is NECSTonatNECST.

First of all, i have to say that this start-up was born as a university project. Indeed, Luca and Claudio are attending their third year of the bachelor in computer engineering at Politecnico of Milan and they are working with NECST lab (Novel, Emerging Computing System Technologies Laboratory). For clarity, this lab comprises a spread number of researches on advanced topics in computing systems ranging from architectural characteristics, to hardware-software co-design methodologies, to security and dependability issues of complex system architectures. This is where their project starts from.

Now that you know the context the two guys are working in, let’s talk about their idea.

In this project computers are used for the financial purposes. In fact, with it Luca and Claudio try to solve the pricing problem of financial options. Specifically, options are contracts that grant the right, but not the obligation, to buy or sell an underlying asset at a set price on or before a certain date and nowadays are one of the most used financial derivates. The idea was born since, even if options are so common, the pricing process is not that simple because it requires an accurate and quick pricing tool to be successful and so the “perfect” model doesn’t exist yet.

Getting into details of the project, in order to describe options value oscillations as a function of time, in the most appropriate mathematical way, is necessary to use stochastic volatility models. At the current State-of-Art, the Heston model is the best one to describe the market behavior. It is a stochastic volatility model which assumes that the volatility of the asset is not constant, nor even deterministic, but follows a random process and, since it is really complex, it needs Monte Carlo simulations to be implemented. Monte Carlo methods are a broad class of computational algorithms that rely on repeated random sampling to obtain numerical results. These simulations are generally implemented on a FPGA ( field-programmable gate array ) that, thanks to is parallel nature, can be used to reduce the time consumption to run these simulations. Therefore, the idea consists in developing a framework based on a central CPU which coordinates a cluster of FPGAs in order to combine the flexibility of the CPU and the efficiency of FPGAs. In this way, the CPU computes the Monte Carlo algorithm while FPGAs price different options and manage the computational load and, thus, it is possible to overcome the high running time and the low energy efficiency level of the current State of the Art.

I am glad they talked with me about their idea because i found it interesting and a useful innovation. I really hope that they will have success in the future.

I believe this project will be for you as interesting as it has been for me.

If you want to know more about the project, follow Luca and Claudio on the socials:

https://www.facebook.com/NECSTonatNECST/?fref=ts

https://twitter.com/NECSTonatNECST

And if you want to know more about NECST lab: http://necst.it/about