Engine Performance Benchmarks & Methodology

Benchmarking Methodology & Reference Hardware

All benchmark measurements were conducted using a repeatable protocol with seeded deterministic inputs:

• Hardware: 4 logical cores, 8 GiB RAM, NVMe SSD storage.
• Environment: Node.js v24 LTS and Chromium modern engine.
• Input Datasets: Standard synthetic vectors (0 B, 1 B, 256 KiB, 1 MiB, 8 MiB, 64 MiB, 256 MiB, 512 MiB) comprising zeros, periodic cycles, pseudo-random sequences, and ASCII strings.
• Repetitions: Minimum 5 warmup runs followed by 10 measured runs; medians and p95 percentiles reported.

Measured Latency & Memory Retention

Note: In 512 MiB streaming read mode, retained memory remains bounded under 50 MiB because chunks are processed incrementally without holding the full file in memory.

Production Build Asset Budgets

Monitored on every build via scripts/verify-build.mjs:

• Homepage HTML: 3.2 KiB gzip (Budget: 6.0 KiB).
• Initial Application JS: 83.2 KiB gzip (Budget: 110.0 KiB).
• Application CSS: 5.9 KiB gzip (Budget: 7.0 KiB).
• Shared SEO CSS: 3.1 KiB gzip (Budget: 7.0 KiB).
• Total Initial Transfer: 95.3 KiB gzip (Budget: 150.0 KiB).

Local processing and privacy

The Benchmarks workflow opens the same Bitpeek browser workspace. Selected file bytes, file names, pasted input, searches, edits, hashes, and comparison data are processed in local browser memory and are not sent to a Bitpeek server.

The static guide itself can be read without opening a file or creating an account.

Frequently asked questions