What is Tercen's computation engine?
One of Tercen's core architectural features is its high-performance computational relational engine. The engine is behind every visual and computation performed on Tercen. Each computation performed (on Tercen) produces a relationship between the input and output data. In a Tercen workflow, a sequence of computations results in a sequence of computational relationships. This architecture's benefit is a very flexible and powerful visualization and computational system and a very formal definition of all computations performed in Tercen. This is essential for reproducible science.
An illustrative example:
Suppose a Tercen user loads up a dataset. The user performs two successive computational steps on the raw values, say an "average" calculation first, followed by a "log" calculation. For each computation, Tercen keeps a relational linkage across the steps, i.e. the raw value, the average value, and the log value. A user can now rapidly create a flexible view (or computation) using the raw values and condition it (e.g. using color or filter) by the average or log values. This ability to produce flexible high-performance visuals with connected values, without programming skills, is unique.
A performance roadmap for each version of Tercen's Computation Engine. This indicates the number of data points a user can upload and the number which can be visually interacted with:
Year | Version Comp Engine | Instantaneous interaction | Processing |
2018 | 0.1.0 | 500,000 | 1,000,000 |
2020 | 0.2.0 | 1,000,000 | 30,000,000 |
2021 | 0.3.0 | 5,000,000 (estimate) | 100,000,000 (estimate) |
An instantaneous interaction is when the user does not wait when manipulating the data via the Tercen visual interface.
We are excited by the current and future possibilities this technology brings to data analysis, and we think users are truly empowered to go further with their data.