Benchmarks

Looking for the indicator library's numbers?

This page is about Wickra Proof. Wickra's own indicator benchmarks — the comparison against TA-Lib, talipp, pandas-ta and the other Rust TA crates — live at wickra.org.

A proof's cost is dominated by the backtest run itself; canonicalization and the blake3 hash are a thin, near-constant tail on top. The benchmarks here measure prove end-to-end (backtest + canonicalize + hash) and canonicalize in isolation, so you can see how the determinism layer compares to the work it wraps.

What is measured

The proof-bench crate (criterion) covers:

• prove across a matrix of candle count (200, 1 000, 5 000 bars) × indicator count (specs referencing ~2 and ~10 indicators). Throughput is reported in bars per second.
• canonicalize on a small and a large BacktestReport, both taken from genuine proofs so the JSON shapes match production output exactly.

Methodology

Run against fixed, in-process synthetic universes so the numbers are reproducible and contain no I/O variance:

cargo bench -p proof-bench

Results

Measured with cargo bench -p proof-bench (criterion) on a Windows x86-64 laptop. Figures are the median estimate; treat them as orders of magnitude, not guarantees — they vary with CPU, toolchain and the linked wickra-backtest version.

Benchmark	Bars × indicators	Median	Throughput
prove/200bars_2ind	200 × ~2	1.13 ms	~177 K bars/s
prove/200bars_10ind	200 × ~10	1.30 ms	~154 K bars/s
prove/1000bars_2ind	1 000 × ~2	6.02 ms	~166 K bars/s
prove/1000bars_10ind	1 000 × ~10	6.49 ms	~154 K bars/s
prove/5000bars_2ind	5 000 × ~2	30.3 ms	~165 K bars/s
prove/5000bars_10ind	5 000 × ~10	32.2 ms	~155 K bars/s

Benchmark	Report	Median
canonicalize/small_report	200-bar proof	168 µs
canonicalize/large_report	5 000-bar proof	4.75 ms

The takeaway: per-bar throughput stays roughly constant as history grows (~150–177 K bars/s), so prove cost scales linearly with candle count and only mildly with the number of indicators a spec references. Canonicalization of even a large report is a few milliseconds — a small fraction of the prove that produced it — so the determinism guarantee costs little over the backtest it commits to. The nightly bench.yml workflow reruns this on a clean Linux runner for tracking over time.

Caveats

These figures bound the prove / canonicalize work only. End-to-end time in a real run also depends on loading the candle series from disk, which these in-process benchmarks do not capture. verify recomputes prove, so its cost tracks the prove figures above plus a second canonicalize.

The numbers above are the ones in the repository's BENCHMARKS.md, measured with the commands it names.