Kinetica (software)
| File:Gpudb.png | |
| Developer(s) | GIS Federal |
|---|---|
| Stable release | 3.2.5
|
| Platform | Cross-platform |
| Type | In-Memory |
| License | Proprietary |
| Website | GPUdb Home Page |
GPUdb is an in-memory database that uses a distributed rendering pipeline of many core devices (mainly graphics processing units (GPUs)), to process large amounts of data. GPUdb has SQL-style query capability that enable users to calculate analytics on high velocity, streaming datasets and visualize results with sub-second response time. GPUdb dynamically allocates data between GPU VRAM, system RAM, and flash and disk persistence. Originally designed as a geospatial and temporal computational engine for the United States Army Intelligence and Security Command (INSCOM)[1][2] at Fort Belvoir, GPUdb evolved into a full enterprise database in 2012 and was commercially available to the public in 2014 by GIS Federal.
GPUdb is an object store that allows users to define any number of Sets, and populate these Sets with objects. A Set is analogous to a traditional database table where each column (GPUdb Set Attribute) can be one of the following types: int, long, float, double, or string. GPUdb then allows the user to perform calculations against those Sets. GPUdb also includes a number of GPU accelerated geospatial and temporal functions that can be performed on any Set.[3]
Many core devices
GPUdb leverages many-core devices and supports NVIDIA GPUs and Xeon Phi many-core devices.
Distributed rendering pipeline
GPUdb natively has a built in distributed rendering pipeline that is able to utilize the GPUs to dynamically create on-the-fly images and time referenced videos at query time. Using the distributed rendering pipeline, GPUdb is able to serve as a geospatial web server.
Geospatial web server capabilities
GPUdb has a built in geospatial web server that supports the Open Geospatial Consortium (OGC) compliant Web Mapping Service (WMS) and Keyhole Markup Language (KML) web services. GPUdb supports simple feature, classbreak, and heatmap rendering.
History
GPUdb was incubated by the military in 2009 and went commercially available to the general public in 2014.
Awards
International Data Corporation (IDC) HPC Innovation Excellence award 2014.[4]
References
- ^ Adie, Jeff. "Real time GPU accelerated analysis of Big Data" (PDF). SGI. Retrieved 15 January 2015.
- ^ Woodie, Alex (8 October 2014). "GPU-Powered Terrorist Hunter Eyes Commercial Big Data Role". datanami. Retrieved 15 January 2015.
- ^ "GPUdb". Retrieved 14 January 2015.
- ^ Conway, Steve; Joseph, Earl (18 November 2014). "IDC Announces New Winners of HPC Innovation Excellence Awards". HPC Innovation Excellence Awards. No. 2014. IDC. International Data Corporation. Retrieved 18 November 2014.