Some of our Research Projects on the Data Vortex

(page under construction)

Spectral computation of incompressible fluid flow

The spectral method is one of the most accurate available to solve Incompressible Fluid Flows with high Reynolds numbers in periodic domains. Since it is based on Fast Fourier Transforms it is also very communication-expensive, thus Data Vortex is a good choice for this kind of problem. (click image for details)

Direct solution of high dimensional Schrodinger’s equation

Quantum mechanical problems involve the solution of high dimensional partial differential equations. As dimensions increase, the communication cost increases together with the computation cost. Here we show how we solve the initial value problem in 6 and higher dimensions for the Schrodinger equation. (click image for details)

Data Vortex Profiler

The Data Vortex profiler is a graphical interactive tool for the optimization of parallel Data Vortex code. It can dissect a run and display information for each of the API calls with the click of a mouse. (click image)

Multi Asset Option valuation

The computation of the value of multi-asset derivatives requires the numerical solution of high dimensional Black Scholes equations that are communication intensive. Here we explain how Data Vortex can be used to compute exotic European and American options. (click image)

Simulation of biological neural networks

The simulation of brain activity is computationally costly and the communication patterns are irregular and massive. Here we explain how we use DV to integrate the evolution of biological models of hundreds of millions of neurons. (click image for details)