All data that enters the simulation is composed into dynamic objects. These objects make up the common fabric of the simulation, being able to represent everything from a simple social media post to a complex actor such as a sports player and everything in between.
Dynamic objects are autonomous, and are capable of navigating through the simulation to satisfy their own set of goals. These goals can vary from one class of object to the next, and even between two object instances of the same type. The Helios Goal system is not a simple framework to construct business rules, orchestrations, or system behaviors, but a completely new way describing what we call "Wisdom". It is this guiding wisdom that objects use to govern their behavior.
Objects possess collections, or families, of events (or actions) that they can perform. The Helios platform determines which event or series of events should be invoked in order to achieve an object's Goal. These events are not static, but can be extended and even abstracted into a more general purpose state through Helios' proprietary AI approach, resulting in an object being able to adapt to changes in its environment.
Developers can implement using standard languages (such as C++ or any of the .NET Languages), as well as, Helios' proprietary AI Language. The Helios AI Language takes inspiration from first order predicate calculus languages is empowered with the parallel processing power of the Helios Platform and native advanced AI functionality. Developers can freely mix and match languages within the platform as needed.
Code developed in the Helios AI Language has the additional benefit of being able to be modified by the Helios AI Manager to dynamically adapt to new discoveries. These modifications can result in adapting to new knowledge, expanding or abstracting code into a more generalized form, as well as allowing for the transference of knowledge.
The Helios platform provides SDKs for all the major OS environments, as well as, a web services layer allowing for connectivity to Helios for virtually any environment. Applications can call on Helios as needed, or establish a constant socket/websocket connection and receive data in real-time. With the Helios Client SDK you can construct dashboards, charts, and reports that constantly update in real-time as data in the simulation changes, as well as tap into the prediction power of Helios to show real-time forecasting and trending.
Free from the constraints of having to support a Legacy platform or architecture, we have been able to design the Helios Platform from the ground up to take advantage of today's modern hardware architectures - specifically the massive parallel processing power of Graphical Processor Units (GPUs) found in NVidia's latest super computing Tesla cards.
Starting at the heart of the Helios platform is the ability to manipulate vast amounts of data. And to that end, the Helios Data Performance Engine encapsulates many of the necessary data manipulation tasks that Helios leverages to effectively run massive simulations, as well as, its advanced Artificial Intelligence and Wisdom systems.
The engineers at Simantex have developed revolutionary new approaches to handle the demand of a constantly simulated environment at massive scales - from high performance querying, to building on-the-fly indexes and data sorting, to exporting data to third parties or persistent storage, and the importing of data into the platform, the Helios Data Performance Engine can handle it with ease.
All of the metrics shown here illustrate the tremendous percent increases the Helios platform can achieve over today's standard environments. However, keep in mind that as impressive as this is, it pales in comparison to the true performance of Helios when it is running on its default Super Computing Node - as defined below.
These impressive numbers shown here are using just one GPU card running at 10 Teraflops. A standard Helios compute node is equipped with 20 GPUs each with 50% more general computing power than the one used here, and a staggering 2.4 Petaflops of additional AI specific computational horsepower!
Simantex's Helios platform is designed to be ultra-configurable to maximum the scaling capabilities of each facet of the architecture. Whether the need arises to scale all features of the simulation, or just areas around simulation size and object management, data querying performance, data processing speed or focus on driving more AI insights faster, the Helios Platform can be configured to handle the load. Simantex offers the Helios Platform in the following scalable configurations:
Seamlessly join multiple super computing nodes to scale out the Helios Platform to meet any scalability challenge. Whether managed in a Simantex Private Cloud, or On-Premise, Helios Super Computing Nodes are designed to deliver incredible performance, maximizing an architecture that has been finely tuned with the hardware solution. The following represents the immense performance and capacity capabilities of a Helios Super Computing Node: