The SQL Processing Unit (SPU™)
The SPU is an industry-first processor architecture proven to accelerate high-throughput data analytics workloads.

How It Works
The SPU supports common columnar file formats. When a query is pushed down from the query engine, the SPU directly processes the data files from local storage and outputs the results to the query engine in native query engine layouts.
How to Deploy
SQL Processing Unit (SPU™)NeuroBlade SPU processes data at I/O speed, regardless of query complexity. The SPUs are installed as PCI-e cards that can be deployed in your data center.
HEQS is a rack-mounted server integrating NeuroBlade’s SPUs with high bandwidth storage and networking. Performance scales to your biggest needs with pure linear scalability.
How To Integrate
NeuroBlade provides an open set of APIs that snaps into any query engine using existing integration points.