Magma enables Big Data startup, Map-D, to utilize GPU clusters.
(PR Web Via Acquire Media NewsEdge) San Diego, California (PRWEB) June 23, 2014
Magma and Map-D have teamed up to develop an integrated solution for the visualization and analysis of Big Data across a cluster of GPUs. This exciting Big Data solution combines new ultra fast expansion technology from Magma, with Map-D's real-time, end-to-end visualization and analytics platform. Magma's expansion chassis, the ExpressBox 3600, allows Map-D to parse out parallel computations to multiple GPUs without the bottlenecks of traditional super computer architecture. By housing GPUs in one single chassis, Magma allows Map-D users the ability to work on huge data sets with a workstation that can sit at a desk, instead of large and expensive super computers occupying multiple racks of networked computers.
"We are pleased with the computational results Map-D has been experiencing with the adoption of the ExpressBox 3600," said Jim Medeiros, strategic business development manager at Magma. "By teaming with Map-D, we are able to provide a combined solution that unlocks the novel data mining capability of Map-D with linear scalability of up to 9 GPUs per expansion system."
Magma's newest high performance computing chassis, the ExpressBox 3600, is a 4U rack enclosure that can house up to 9 double-wide world class GPUs and connect them over PCIe Gen 3 interconnect. Magma's PCIe Gen 3 backplane is designed to efficiently move the terabytes of data needed for Big Data applications between GPUs and between GPUs and the host compute complex. Dedicated hot swappable power supplies and fans are integrated into the ExpressBox 3600 chassis to provide high reliability and serviceability to the solution.
"Magma's expansion solutions allow Map-D to execute Big Data visualizations using multiple GPUs and adds near-linear scalability for our customers," said Tom Graham, co-founder of Map-D. "We believe we will see amazing applications using the Map-D platform, and are excited to make this much computing power available to our customers at a price and footprint that was inconceivable only a few years ago."
Map-D's exciting new platform allows developers and end users to manipulate and visualize Big Data in real-time, orders of magnitude faster and cheaper than contemporary solutions. This new solution has applications in a broad range of industries, from finance to medicine, defense to advertising, and beyond. To showcase the possibilities of Map-D's platform, they are currently running 2 interactive streaming demos on their website at http://www.map-d.com.
Tweet map demo: http://mapd.csail.mit.edu/tweetmap-desktop/
Campaign map demo: http://mapd.csail.mit.edu/campaign-map/
Magma develops and manufactures expansion computing solutions using Peripheral Component Interconnect (PCI), Peripheral Component Interconnect Express (PCIe), and Thunderbolt Systems for servers, desktops and portable computing platforms. Magma's line of products includes enclosed and non-enclosed solutions ranging from single slot expansion products to fault-tolerant, Reliable Available Serviceable (RAS) systems specifically designed for mission critical environments. Magma products are platform independent and are used in a variety of industries including audio and video production, test and measurement, medical imaging, surveillance, aerospace and defense, telecommunications, data acquisition, and high performance computing. Magma's team of experienced engineers has been developing expansion and connectivity solutions since 1987 and holds more than a dozen technology patents, making them a pioneer and leader in the market space. All Magma products are designed and assembled in the United States. For more information about Magma visit http://www.magma.com.
Map-D is the world's fastest end-to-end big data visualization and analytics platform. Powered by a unique GPU in-memory SQL database, Map-D delivers real-time big data processing at millisecond latency across billions of rows of data. Their revolutionary database architecture executes queries and processes visualizations and analytics entirely from GPU memory, enabling unprecedented interactivity and speed across the Big Data pipeline. Map-D's platform helps researchers and enterprises work with enormous data sets in industries including; finance, security, social media, and politics. Map-D is based in Cambridge, Massachusetts and was founded in 2013, by MIT and Harvard researchers Todd Mostak and Thomas Graham. For more information about Map-D and their dynamic founders, visit http://www.map-d.com.
Read the full story at http://www.prweb.com/releases/2014/06/prweb11965657.htm
(c) 2014 PRWEB.COM Newswire
[ Back To Telecom Signaling's Homepage ]