Home

Dataframe Mojo research archive

Implementation history, benchmark results, and one-off experiments live here so they do not ship with the source checkout.

Research notes

• Benchmark history
• Current overall benchmark, sorted by 1M ratio
• CPU performance evaluation
• CSV pipeline, issue 152
• Polars CSV source mapping
• Numeric parsing, issue 149
• Post-CSV-port benchmarks

Experiment files

• CPU 110 experiments
• CSV integer experiments
• CSV port experiments

The complete scripts, patches, raw measurements, and result notes are under experiments/ in the wiki Git checkout. Clone the wiki only if you need to reproduce an experiment. Most harnesses expect to run from a separate dataframe_mojo source checkout and use -I to point at that checkout.