sirius-db/sirius

Health v2 · maintenance only

Flags: no_releases no_readme

Adoption not part of the score

1059 stars · 121 forks observed · 2026-09-03

What it is AI-extracted, prompt v1, taxonomy v1, 2026-08-30, confidence not recorded

Sirius is a GPU-native SQL analytics engine written in C++ that accelerates query execution by offloading it to GPUs. It integrates with existing databases like DuckDB and Doris via the Substrait query format, requiring no query rewrites, and includes CPU fallback for compatibility.

Use cases

• accelerate SQL analytics queries on GPUs
• speed up DuckDB workloads without rewriting queries
• run OLAP analytics 10x faster at the same hardware cost
• plug GPU acceleration into existing database stacks via Substrait
• deploy GPU-accelerated analytics on cloud or on-prem
• offload heavy analytical processing from CPU engines

When to choose

• you have analytical SQL workloads bottlenecked by CPU execution
• you want GPU speedups without changing your existing database or queries
• you need Substrait-compatible query acceleration with CPU fallback
• you run large-scale OLAP analytics where hardware cost efficiency matters

When to avoid

• you lack GPU hardware or cannot rent GPU instances
• your workload is mostly transactional (OLTP) rather than analytical
• you need a fully mature, battle-tested engine for critical production systems
• your queries are small and latency is dominated by data transfer to GPU

Facets

library · maturity active

database search-engine gpu-computing databases analytics big-data gpu-computing cloud self-hosted gpu-native sql-engine query-acceleration substrait olap duckdb-integration linux gpu

Member repositories

For agents

markdown · JSON · MCP: product_card(name="sirius-db/sirius")