Sneller SPRINT Query Engine

SQL has been resistant to efficient SIMD acceleration. Sneller SPRINT changes that. AVX-512 vector units have been shipping in Intel processors since 2018. Our engine directly and efficiently converts SQL statements to native AVX-512 assembly. You can now query terabytes of JSON directly, including nested fields, at interactive speeds without requiring specialized acceleration hardware.

Blog: Sneller & Elastic Benchmark Results

Check out the results of our head-to-head benchmark with Elastic for semi-structured event data analytics workloads. (Spoiler alert: we won)

Sneller has built a direct SQL to AVX-512 compiler. AVX-512 units provide 16x the number of execution cores compared to standard query processing.

Powerful AVX-512 instructions such as per lane predication, gather load and scatter store enable efficient execution of SQL statements.

Sneller's AVX-512 routines are very efficient at parsing and filtering JSON data at high throughput rates.

Sneller operates scalable, distributed in-mem compute engines. Data is appropriately distributed within the clusters and cached based on query frequency.

Sneller's AVX-512 code can accelerate SQL statements by > 10x, but to speed up entire queries we eliminate all bottlenecks using vector acceleration from start to end.

Ready to speedup and simplify your event data analytics?

Sneller is also available as an open source project on GitHub.