Iota GPU Library

Iota is a library for implementing high-performance engineering, scientific, statistical, and finanical applications using CUDA on an NVIDIA GPU board.

 Iota  features include:

• Can be used with Maxwell-series boards and above. Best performance is achieved using the newer RTX 3000-series of boards and beyond.
• Multi-GPU support.
• Easy-to-use C++ API. (Interfaces to other languages including Java, C, C#, and Python are in the works.)
• Vector classes corresponding to all common C++ data types.
• Built-in vector processing functions for common operations such as dot-product, inner-product, fill, etc.
• Automatically resize all related vectors when you modify your data model. (Simplifies code maintenance.)
• Available for Windows, and Linux (RHEL, CentOS, and Ubuntu).

In addition to the basic Iota library, we are developing these domain-specific sister libraries:

• IotaSPH. This library adds additional functionality to simplify the creation of Smoothed Particle Hydrodynamics applictions. Included are infrastructure support, fast Nearest Neighbor search, and several Kernel "Weight" functions.  This allows you to quickly create an entirely new, GPU-enabled SPH application in only a few hundered lines of code. Just plug in your specific physics routines and go.
• IotaMath. This library adds a number of additional math functions such as linear equation solvers, eigenvalue solvers, and functions frequently used in statistics and finance.

The Iota libraries are supported, commercial products. They can be used to create new, high-performance applications. Or, they can be used to supercharge your existing applications.

Please contact us for additional information or to discuss your High Performance Computing requirements.