Follow Datanami:
April 3, 2018

MapD Rolls GPU Cloud

MapD Technologies released a new version of its mainstay database and visualization platform about a year ago with a heavy emphasis on GPU acceleration. It is now migrating its analytics platform to the cloud with what it bills as a software service offering souped-up analytics.

Building on its capabilities in its enterprise analytics platform, the company said this week its MapD Cloud provides “one-click access” to its open source SQL engine and visual analytics platform. The GPU-accelerated cloud was released in connection with last week’s GPU technology forum sponsored by MapD investor Nvidia (NASDAQ: NVDA).

Among the benefits of a GPU-accelerated cloud service is freeing analytics users from having to spin up their own GPU instances. Another is the ability to scale and expand access to faster query and visualization tools.

“Customers tell us their traditional analytics tools are failing to cope with the torrent of massive data sets their organizations are facing,” said Todd Mostak, MapD’s co-founder and CEO. With that in mind, the new cloud service also advances the San Francisco-based company’s strategy of making its GPU-backed tools accessible to more researchers and enterprise users, Mostak added.

Along with GPUs, in-memory computing is increasingly being used to crunch huge amounts of data while in the process adding more interactivity between users and their data sets.

MapD said Tuesday (April 3) it is offering the cloud analytics service in several subscription tiers, with pricing starting at $150 a month for up to 10 million-row structured data sets. A free, two-week trial is also available for up to 100 million rows.

The company is also providing a managed cloud service scaling to tens of billions of records with sub-second query performance.

For Nvidia, which has focused its GPU technology on deep learning and other AI applications, the cloud partnership with MapD strengthens its presence in the high-performance analytics market. The partners portray their collaboration as a step toward “mainstream adoption of GPU-accelerated analytics” by providing what MapD is positioning as the first GPU service in the cloud.

MapD’s cloud service builds on a string of GPU-computing initiatives tied to its SQL-based analytic database that targets real-time visualizations of geospatial and other large data sets. Along with its open-source analytics platform, the startup is a founding member of the GPU Open Analytics Initiative.

The startup was launched in 2013 by researchers at MIT’s Computer Science and Artificial Intelligence Laboratory.

Recent items:

MapD Upgrades GPU-Based Database Performance

Nvidia Riding High as GPU Workloads, Capabilities Soar

GPUs Seen Lifting SQL Analytic Constraints