Analytics platform MapD has partnered with IBM Power Systems to enhance its analytical performance. California-based MapD has optimized its MapD Core database and MapD Immersive visual analytics client to take advantage of IBM Power Systems to target the speed at which SQL queries can be performed.
Running queries on IBM Power Systems S822LC for HPC servers versus x86-based servers offers an SQL querying speed that is up to 1.3 x faster than recent benchmarks. On average, the updated queries show a 65% acceleration over the original, benchmarked SQL queries. Upgrades such as these help users analyze billion-row data sets in real-time.
“We’re bringing new performance capabilities to a broader audience only available on IBM Power Systems,” said Ashish Bambroo, VP of Business Development at MapD Technologies. “IBM Power Systems are optimized for the kind of compute-intensive applications where MapD’s technology excels. By harnessing the advantages of IBM Power Systems, we can provide our enterprise customers with more options for tackling the toughest workloads.”
To increase the speed, MapD is leveraging IBM’s POWER CPUs and NVIDIA Tesla P100 graphics processors. The POWER processor CPU works together with the NVIDIA Tesla P100 GPU, accelerating CPU-to-GPU analytic pipelines up to 2.5X. This acceleration bolsters MapD’s platform that uses GPUs to query and visualize billions of records in real-time. The company’s GPU-based platform performs from 75 to 3,500 times faster than traditional CPU databases.
MapD was founded in 2013. The company’s CEO Todd Mostak most recently demoed the company’s Core and Immerse components at FinovateSpring 2017. Mostak also gave a presentation at FinDEVr New York 2017 titled How GPU-Powered Visual Analytics Are Remaking Financial Services. Last month, MapD ranked among the best database and analytics tools according to InfoWorld.