Chart · milliseconds
sum by region.
neo 0.020 ms in the language study, 0.020 ms in the library run. Closest: Zig at 1.20×. Farthest: Spark at 1,948×.
Embed anywhere.
<iframe src="https://n0.is/embed/chart/sum" title="sum by region" width="100%" height="520" loading="lazy" sandbox="allow-scripts" referrerpolicy="no-referrer" style="border:0;border-radius:12px"></iframe>Every engine's multiple of neo
- neo1×
- Zig1.20× · 0.024 vs 0.020
- Swift1.55× · 0.031 vs 0.020
- Rust2.10× · 0.042 vs 0.020
- Go3.95× · 0.079 vs 0.020
- Polars77.3× · 1.55 vs 0.020
- DuckDB112× · 2.24 vs 0.020
- pandas612× · 12.2 vs 0.020
- Spark1,948× · 39.0 vs 0.020
Bars to scale, from zero to the slowest engine. The gray pair is the engine's measurement and neo's in the same run, in ms. The languages and the libraries were measured in separate runs, so neo has a number in each; every multiple is against neo in its own run.
What this measures
Total the amount column for each of five regions: a grouped sum, the kernel every dashboard runs.
What neo is held to
The JavaScript loop's result to the bit. A floating-point sum depends on the order it is taken in, and neo is held to the row order's answer whatever it does to get there.
Each side, this kernel only
Zig1.20× neo · neo vs Zig
Written in-house; it stays with the engine. How it was run and checked is on its comparison page.
Swift1.55× neo · neo vs Swift
Written in-house; it stays with the engine. How it was run and checked is on its comparison page.
Rust2.10× neo · neo vs Rust
Written in-house; it stays with the engine. How it was run and checked is on its comparison page.
Go3.95× neo · neo vs Go
Written in-house; it stays with the engine. How it was run and checked is on its comparison page.
Polars77.3× neo · neo vs Polars
timed("agg", lambda: df.group_by("region").agg(pl.col("amount").sum()))DuckDB112× neo · neo vs DuckDB
timed("agg", lambda: con.execute("SELECT region, sum(amount) FROM t GROUP BY region").fetchall())pandas612× neo · neo vs pandas
timed("agg", lambda: df.groupby("region")["amount"].sum())Spark1,948× neo · neo vs Spark
timed("agg", lambda: df.groupBy("region").agg(F.sum("amount")).collect())neo · confidential
neo's kernels, the ports written in-house to beat them, and the techniques behind both stay with the engine. Nothing about how the numbers were made is withheld: the method, the checks every engine had to pass, every number and every place neo is behind are on this page. The engine itself we keep confidential, because we don't want it getting into the wrong hands. The binaries run on request and reproduce every number here.
Measured on Apple M5 Max, 18 cores, 26.4.1, 2026-10-11 and 2026-10-15. Every number here is also on the comparison page of its engine, with how it was taken and the caveats.