ArcticDB Performance for Quantitative Backtesting and Bar Storage

Summary

This community post compares ArcticDB with DolphinDB, SQLite, and DuckDB for quantitative research workloads. It describes ArcticDB as an embedded database and reports a backtest on 1.47 million records completing in under 19 seconds, compared with about 25 seconds for DolphinDB and under 80 seconds for SQLite. DuckDB is reported at roughly 23 seconds for the backtest, while the author says its data writes were slower. The post also reports storing 1.61 million one-minute bars with ArcticDB in about 30 seconds.

These figures suggest that database choice can affect both backtest runtime and data preparation, and that write and read performance should be considered together. The evidence is an informal, single-user comparison; hardware, configuration, data layout, and reproducibility details are not supplied. The author describes the results as preliminary testing, so the timings are a point of reference rather than a general benchmark or guarantee of performance in other trading systems.

Key ideas

• The post compares four databases on a reported backtest workload.
• ArcticDB had the shortest reported backtest time in this particular comparison.
• The author reports strong ArcticDB performance for storing one-minute bars as well.
• DuckDB's data insertion is described as slower despite a competitive backtest time.
• Missing hardware and configuration details limit how broadly the timings can be applied.

Tags

This summary was written by Stratmill's research agent from the original; it is not a copy of the source.