The MÄK RTI enables High Level Architecture (HLA) federations to rapidly and efficiently communicate. Strong performance in an RTI increases a simulation system’s capacity for spatial updates, providing higher fidelity to a simulation. With this in mind, we’ve made significant performance increases for MAK RTI 4.4.2.
Spatial updates are the most common type of update for most exercises. They occur when a single entity changes position, velocity, acceleration, or orientation. When a simulation contains thousands of entities, updates are constantly pouring in. Therefore, when we measure send and receive performance, we generally test updates that are of spatial update size, or around 100 bytes. In our last release, we more than doubled the number of updates that the RTI can handle. In this release, we now had added another 30% on top of that, now more than tripling the RTI's 2014 performance.
When an object first appears in a real-time exercise (usually at the start), it must be registered by the RTI. If a large simulation contains hundreds of thousands of entities, it can take time for everything to be found which can become an issue when processing messages in real-time. MÄK RTI 4.4.2 has increased its object registration performance to over 60x it's previous benchmark, making it an ideal tool for large and complex simulations.
Time-managed exercises require even faster RTIs in order to process data at a higher speed than real-time. This is particularly helpful when running Monte Carlo simulations, which require multiple simulations of similar scenarios. For our customers requiring time managed performance, the .MAK RTI 4.4.2 sends messages 9x faster than its predecessor.
For customers who prefer to use Java, we've spent considerable engineering resources to bring our Java support up to the quality of our C++ and C# support. The new Java bindings are an order of magnitude faster than before, now on par with the native version. Now you can chose your preferred programming language without sacrificing performance.