Accelerating Analytics with GPU Platforms

Due to exponentially growing data stores, organizations today are facing slowdowns and bottlenecks at peak processing times, with queries taking hours or days. Some complex queries simply cannot be executed.

Data often requires tedious and time-consuming preparation before queries can be run. David Leichner, CMO, SQream, demonstrated how the power of GPUs can help conquer these challenges with his presentation, “Accelerating Analytics in a New Era of Data,” during Data Summit 2019.

“Data governance is great but one of the biggest issues with data governance is you can’t get to the data,” Leichner said.

SQL Queries and BI analytics are taking too long, Leichner said. According to Leichner, 90% of BI data is lost because the data can’t be found.

“You don’t even know what you’re missing,” Leichner said.

Companies need a new paradigm to handle the mass amounts of data proliferating systems, Leichner explained.

He recommended keeping it simple and create more suitable architecture to store and analyze data.

SQreamDB provides a solution that is powered by GPUs. It is a massively scalable SQL database that can be extended for AI and machine learning. It has a minimal footprint and is lightening fast, Leichner said.

“We do not replace existing data warehouses,” Leichner said. “We come in and sit alongside your data warehouse. We’re providing the ability to run adhoc queries on your raw data or significantly increase the speed of queries.”

Many Data Summit 2019 presentations are available for review at