simulation, analysis, and optimization. These methods are becoming more and more popular not only in engineering but also in many branches of science – for instance, in chemistry, in biology, and in medicine.

Problems of engineering analysis may be roughly divided into two groups – problems of analysis of solid constructions (solid mechanics) and problems of analysis of fluids and gases (it is called CFD – computational fluid dynamics).

Both types of problems - solid mechanics and CFD - are treated by solving corresponding partial derivative equations (PDE), and all these classes of problems, after discretization, result in solving systems of linear equations with matrices of a similar structure. The structure of these matrices is very specific - the matrices are very sparse, symmetric and positive defined.

For such matrices, there are well known efficient methods - for instance, conjugate gradients method. Some variation of this method is used in most PDE-solvers. There is a common opinion that it is not possible to outperform conjugate gradients.

But it is possible! Our Marlin Solver is dramatically much more efficient (especially when the hardware is capable of mass parallelization).

We see our mission in providing the technology that opens opportunities far beyond conjugate gradients. We called this technology Marlin Solver.