MapD opens up its speedy database powered by graphics chips


MapD Technologies Inc., a fast-rising analytics startup backed by the likes of New Enterprise Associates and Nvidia Corp., beefed up its product lineup today in a bid to open new growth opportunities.

The biggest addition is a new open-source version of its flagship MapD Core relational store. What sets the system apart from the alternatives is the fact that it’s designed to run on graphics processing units rather than standard central processing units, which makes it considerably faster at carrying out certain tasks. Chief among them is data visualization.

According to MapD, the system can bring up records at a rate of several terabytes a second and then turn them into an interactive dashboard using an embedded rendering engine. The latter feature saves users the trouble of exporting data to an external graphics tool, which in turn avoids the need for time-consuming data transfers. Combined with MapD’s fast performance, this makes it possible to generate large visualizations with only a few seconds’ delay between changes.

The capabilities that enable the platform to support data visualization projects so well also lend themselves to powering artificial intelligence software, which similarly runs more efficiently on GPUs than CPUs. Releasing MapD Core under an open-source license should enable the startup to more easily court the growing number of organizations that are deploying such workloads. Some of the new users can be expected to become paying customers, while others may contribute improvements and help enhance its feature set.

Falling into the latter category is the University of California at Davis, one of the first adopters to come aboard. Computer engineering professor John Owens said in a statement provided by MapD that his group plans to use the system “as a research platform and to contribute to the code base.” Organizations with more conventional requirements can use a new community edition of MapD Core that is rolling out as well.

The startup’s efforts to make its platform more accessible will no doubt be met by approval at Nvidia. More organizations using GPU-powered software, after all, means higher demand for its silicon, which was likely one of the chip maker’s main motivations for investing in MapD. It contributed to all four of the funding rounds that the startup has secured to date, most recently a $25 million raise in March.

Image: MapD