sirius-db
Official@sirius-db
Offers specialized diagnostic and performance benchmarking capabilities for GPU-accelerated database engines and concurrent C++/CUDA query execution environments.
Agent Skills by sirius-db
Showing 16 vetted skills indexed across 1 GitHub repositories.
sirius-dbbenchmark
Runs TPC-H benchmarks comparing Super Sirius GPU and DuckDB CPU query performance.
sirius-dbpre-commit-cleanup
Removes AI artifacts and temporary content from Git changes before committing.
sirius-dbupdate-docs
Synchronizes Super Sirius documentation with merged pull requests since the last update.
sirius-dbmodule-context
Loads relevant cudf, rmm, duckdb, and cucascade API documentation based on task keywords.
sirius-dbmodule-discover
Analyzes dependency libraries and generates module-level API documentation with codebase usage mappings.
sirius-dblog-analyzer
Parses Sirius query execution logs into structured artifacts for performance and correctness diagnosis.
sirius-dbvalidate
Diagnose incorrect Sirius GPU query results against a DuckDB CPU baseline.
sirius-dbdataset-manager
Generate and optimize TPC-H Parquet datasets with cudf or pyarrow.
sirius-dbprofile-analyzer
Analyze NVIDIA Nsight Systems profiles for GPU performance and regressions.
sirius-dboptimization-advisor
Analyze nsys profiles to map GPU hotspots to source functions.
sirius-dbbisect
Automate Git bisect to identify regression-introducing commits with build and test execution.
sirius-dbrace-check
Detect race conditions and deadlocks in concurrent Sirius database code.
sirius-dbconfig-optimizer
Optimize Sirius database configurations for TPC-H workloads across scale factors.
sirius-dbbuild-errors
Parse compiler output and iteratively fix C++/CUDA build errors.
sirius-dbtpcds-benchmark
Execute TPC-DS benchmarks across Legacy Sirius, Super Sirius, and DuckDB CPU engines.
sirius-dbruntime-errors
Diagnose runtime query errors using Sirius logs and debugging tools.
Frequently Asked Questions About sirius-db
FAQPage SchemaWhat specific tasks can engineers perform using sirius-db?▼
Engineers can validate GPU query accuracy against CPU baselines, profile performance bottlenecks using Nsight Systems, detect concurrency deadlocks, and execute TPC-DS benchmarks. These capabilities enable systematic regression identification and configuration tuning for high-performance database engines.
Which technical personas benefit from these capabilities?▼
Database engine developers, GPU performance engineers, and systems reliability specialists benefit from these capabilities. The functionality is specifically designed for those maintaining C++/CUDA-based database kernels who require granular visibility into query execution, build-time error resolution, and concurrent memory safety.
What are the primary dependencies for running these diagnostic tasks?▼
Execution requires a configured environment with NVIDIA Nsight Systems for profiling, a TPC-H or TPC-DS dataset generator, and a compatible C++/CUDA compiler environment. Users must also have access to the Sirius database source code and a DuckDB installation to serve as the reference baseline.