Skip to content

Latest commit

 

History

History
42 lines (32 loc) · 1.79 KB

File metadata and controls

42 lines (32 loc) · 1.79 KB

Benchmarks

The repository contains standalone performance and utility benchmarks for stochastic aggregation and differential privacy. The exact paper environment, experiment matrix, bounds, result columns, and plotting commands are documented in repro.md.

Run commands from the repository root. Build the paper executables with:

cmake --build build/release --target \
  as_tpch_benchmark as_clickbench_benchmark \
  dp_benchmark_runner

Paper Benchmarks

Workload Executable Configuration
TPC-H AS as_tpch_benchmark benchmark/configs/utility/as_tpch_sf30.json
TPC-H variable m as_tpch_benchmark benchmark/configs/utility/as_tpch_sf30_m_sweep.json
ClickBench AS as_clickbench_benchmark benchmark/configs/utility/as_clickbench_graviton.json
TPC-H DP utility dp_benchmark_runner benchmark/configs/utility/dp_tpch_sf30_utility.json
TPC-H Q01 AVG dp_benchmark_runner benchmark/configs/utility/dp_tpch_sf30_q01_avg_*.json

The runners accept --config <path> --dry-run to validate a JSON configuration without opening a database. Generated databases and results are ignored by Git and written below benchmark/results/ unless a config selects another path.

Additional Benchmarks

Workload Guides